Maintenance Strategies for the TANZAM Highway in Tanzania
Msc Thesis
August 2009 Ally K. Mwinchande
Delft University of Technology
16
14
12
10
8
6
4 Average Roughness (m/km) AverageRoughness
2
2010 2011 2012 2013 2014 2015 Maintenance Strategies for the TANZAM Highway in Tanzania
by
Ally K. Mwinchande
A Thesis Submitted to the Faculty of Civil Engineering and Geosciences of the Delft University of Technology
in partial fulfilment of the requirements of the degree of
MASTER OF SCIENCE
Examination Committee:
Prof.dr.ir. A.A.A. Molenaar (TU Delft) Ir. L.J.M. Houben (TU Delft) Ir. P.B.L. Wiggenraad (TU Delft)
Delft @ 2009
ii
iii Dedication
To my mom the late mama Sonyo who passed away when I was doing this study
“You did so much for my education mama, may almighty God rest your soul in peace”
iv Acknowledgements
I would like first to thank God for giving me the strength, good health and courage to do this study.
Contributions from many people have made this study successful. I would like to acknowledge and appreciate their contributions.
My special thanks go to the members of the examination committee, prof.dr.ir. A.A.A. Molenaar, ir. L.J.M. Houben and ir. P.B.L. Wiggenraard. I would like to mention and acknowledge the technical assistance, advice and guidance by prof. Molenaar who played a key role as a main supervisor and the head of the section Road and Railway Engineering (RRE) that made my trip to Tanzania for the case study very successful. I would like to convey my sincere gratitude and appreciation to my daily supervisor ir. Houben for his technical guidance and assistance. I have learned a lot from his critical comments and teaching. I acknowledge and appreciate also the valuable advices given by ir. Wiggenraard.
I also acknowledge and appreciate the support and assistance given by my colleagues and the entire staff of the section Road and Railway Engineering during this study.
Many thanks go to Eng. Efata Mlavi and his entire team of RMMS at Tanroads. The support and cooperation given by the team eased my data collection for HDM-4 model application.
I would like also to thank Eng. Innocent Macha from COWI Consulting Engineers (T) Ltd for his important contribution and information about the pavement structures on the case study area.
My special thanks also go to my cousins Ally H. Diwani and Mbwana Ally who assisted me when doing the road condition survey.
I would like acknowledge and appreciate the support, courage, comfort and love given to me by my dear friends from both Tanzania and The Netherlands.
Finally, I would like to thank my dad, my sister and my relatives who were always giving me the courage and support at the time when I needed most.
v Abstract
Road infrastructure is an asset that contributes significant to the economic growth of the nation and reduction in poverty. However, like many other assets, roads do deteriorate with time, and thus need to be maintained. Roads deteriorate due to the effect of environment and traffic loading.
Road maintenance requires a huge amount of resources and thus imposes a major challenge to many low income developing countries where there is always a competing demand for very limited resources available. An economic evaluation of different road investment strategies becomes therefore necessary in order to have scientific criteria for the most economic utilization of the little resources available.
In this research, a section of the TANZAM Highway in Tanzania (Dar es Salaam – Chalinze – Morogoro, 184 km) was selected for a case study. Relevant information about road maintenance in Tanzania was gathered. A road condition survey was done on the case road to identify the existing typical pavement defects and their probable causes. The HDM-4 model was used to predict the life cycle pavement deterioration and to investigate the optimum maintenance strategies from an economic point of view. The study also used the South African mechanistic pavement design analysis method to estimate the pavement life and the viable maintenance strategy in the poor section.
The study revealed that road maintenance financing is one of the biggest challenge facing the government. The visual road condition survey revealed that many of the typical defects observed were probably caused by poor construction quality and vehicle overloading. The section Mlandizi – Chalinze was observed to be in comparative poor condition. Reasonable lifecycle pavement deterioration patterns were obtained by the use of HDM-4 model. However the model has not yet calibrated and adapted for the local conditions thus, the output results must be treated with great care.
Use of the mechanistic pavement analysis method revealed that 40 mm asphalt overlay and pavement reconstruction are viable options for the section Mlandizi – Chalinze.
Key words: road maintenance, maintenance strategies, pavement deterioration, mechanistic pavement analysis,
vi Table of Contents
Dedication ...... iv Acknowledgements ...... v Abstract ...... vi Table of Contents ...... vii List of Figures ...... x List of Tables ...... xiii List of Abbreviations and Symbols ...... xv
1 INTRODUCTION ...... 1 1.1 Background Information ...... 1 1.2 Problem statement ...... 1 1.3 Objectives of the Study ...... 3 1.3.1 Main Objective ...... 3 1.3.2 Specific Objectives ...... 3 1.3.1 Limitation of the Study ...... 3 1.4 Structure of the Report ...... 3
2 LITERATURE REVIEW ...... 4 2.1 Introduction ...... 4 2.2 Asphalt Pavement Distresses ...... 4 2.2.1 Raveling ...... 4 2.2.2 Flushing ...... 5 2.2.3 Rutting ...... 6 2.2.4 Shoving and Surface distortions ...... 7 2.2.5 Transverse Cracking ...... 8 2.2.6 Longitudinal Cracking ...... 8 2.2.7 Alligator Cracking ...... 9 2.2.8 Potholes ...... 10 2.2.9 Road Roughness ...... 11 2.3 Maintenance Management of Asphalt Pavements ...... 12 2.3.1 Management Functions ...... 12 2.3.2 Management Cycle ...... 13 2.3.3 Classification of Maintenance Activities ...... 14 2.3.4 Life Cycle Costing ...... 14 2.3.5 Cost-Benefit Analysis ...... 15 2.3.5.1 Comparison of Alternatives ...... 16 2.4 Road Development and Management Models ...... 17 2.4.1 Past Development on Road Models ...... 18 2.4.2 Road Network Evaluation Tools (RONET) ...... 19 2.4.3 The Highway Development and Management Model (HDM-4) ...... 19 2.4.3.1 HDM-4 Analytical Framework ...... 20
vii 2.4.3.2 HDM-4 Applications ...... 21 2.4.3.3 HDM-4 Modules ...... 22
3 CASE STUDY IN TANZANIA ...... 24 3.1 Introduction ...... 24 3.2 General Information ...... 24 3.3 Tanzania Road Sector ...... 25 3.3.1 Road Network ...... 25 3.3.2 Design and Maintenance Standards ...... 29 3.3.2.1 Traffic Loading ...... 29 3.3.2.2 Climatic Zones ...... 31 3.3.2.3 Pavement Materials ...... 32 3.3.2.4 Subgrade ...... 36 3.3.2.5 Flexible Pavement Design ...... 37 3.4 Visual Road Condition Survey ...... 38 3.4.1 Study Area ...... 38 3.4.2 Ubungo – Mlandizi Section ...... 39 3.4.3 Mlandizi – Chalinze Section ...... 41 3.4.4 Chalinze - Morogoro Section ...... 44 3.5 Conclusion ...... 46
4 APPLICATION OF HDM-4 MODEL ...... 48 4.1 Introduction ...... 48 4.2 Data collection ...... 48 4.2.1 Vehicle Fleet Data ...... 48 4.2.2 Road Network Data ...... 51 4.2.3 Road Works Data ...... 59 4.3 Maintenance Strategies/Project Alternatives: ...... 61 4.4 Running life cycle project and strategy analyses ...... 64 4.5 Results output analysis and discussion ...... 65 4.5.1 Effect of maintenance strategy on deterioration of pavement ...... 65 4.5.2 Life cycle economic analysis ...... 72 4.5.3 Sensitivity analysis for traffic loadings ...... 77 4.5.4 Sensitivity analysis for the choice of discount rate ...... 78 4.6 Conclusions ...... 80
5 MECHANISTIC PAVEMENT ANALYSIS ...... 82 5.1 Introduction ...... 82 5.2 Pavement Structural Analysis ...... 84 5.3 Design Criteria ...... 103 5.3.1 Design criterion for asphalt Layers ...... 104 5.3.2 Design criterion for cement and hydraulically bound (sub) base ayers ...... 105 5.3.3 Design criterion for unbound (sub) base layers ...... 106
viii 5.3.4 Design criterion for sub-grade ...... 107 5.4 Estimation of the pavement life ...... 108 5.5 Optional Maintenance strategies ...... 117 5.6 Conclusion ...... 121
6 GENERAL CONCLUSIONS AND RECOMMENDATIONS ...... 123 6.1 Introduction ...... 123 6.2 Conclusions ...... 123 6.2 Recommendations ...... 126
REFERENCES: ...... 129
APPENDICES ...... 133
Appendix 1:...... 134 Materials Requirements as per Tanzanian Pavement and Materials Design Manual ...... 134
Appendix 2:...... 140 Pavement Catalogues as per Tanzanian Pavement and Materials Design Manual ...... 140
Appendix 3:...... 146 Summary of Road Condition Data ...... 146
Appendix 4:...... 147 HDM-4 Input and Output Results ...... 147
ix List of Figures
Figure 2.1: Typical raveling defect ...... 5 Figure 2.2: Typical flushing with associated rutting ...... 6 Figure 2.3: Typical plastic flow rutting at Chalinze ...... 6 Figure 2.4: Rutting along the wheel paths (left) and simple rut depth measurement (right) ...... 7 Figure 2.5: Typical cracking on asphalt pavement ...... 9 Figure 2.6: Typical alligator cracking on asphalt pavement ...... 10 Figure 2.7: Pothole on asphalt pavement ...... 10 Figure 2.8: Pavement deterioration concept ...... 11 Figure 2.9: Road management cycle ...... 13 Figure 2.10: Transport cost components ...... 15 Figure 2.11: History of the HDM Model ...... 19 Figure 2.12: The concept of the life-cycle analysis in HDM-4 ...... 20 Figure 2.13: HDM-4 System architecture ...... 22 Figure 3.1: Paved and unpaved road network ...... 26 Figure 3.2: Tanzania Trunk road network (Source: Tanroads, 2009) ...... 27 Figure 3.3: Annual Road Fund Revenues (Source: RFB, 2008) ...... 28 Figure 3.4: Tanzania climatic zones ...... 31 Figure 3.5: Average monthly air temperature for Dar es Salaam city ...... 32 Figure 3.6: Average monthly rainfall for Dar es Salaam city ...... 32 Figure 3.7: Design of improved subgrade layers ...... 37 Figure 3.8: Location of the case study road section ...... 39 Figure 3.9: Pavement structure before (left) and after (right) the reconstruction along Ubungo - Mlandizi section ...... 40 Figure 3.10: New cracks emerging from sealed cracks (left) and the sealed l ongitudinal and transverse cracks (right) ...... 41 Figure 3.11: Rutting and longitudinal crack along wheel tracks (left) and new cracks developing besides the sealed cracks (right) ...... 41 Figure 3.12: Pavement structure along Mlandizi - Chalinze section...... 42 Figure 3.13: Alligator cracks initiating potholes (left) and revelling along the wheel tracks (right) ...... 43 Figure 3.14: Shear failure (left) and edge defects and severe rutting due to plastic deformation of the asphalt (right) ...... 44 Figure 3.15: Typical heavy trucks at Chalinze ...... 44 Figure 3.16: Typical pavement structure before (left) and after (right) reconstruction ...... 45 Figure 3.17: Typical rutting at Mikese weighbridge station ...... 46 Figure 3.18: Simple measurement of rut depth (left) and trucks queuing before axle load measurement at Mikese weighbridge station (right) ...... 46 Figure 4.1: ESAL per vehicle data …………………………………………………………...50 Figure 4.2: Road condition data……………………………………………………………….52 Figure 4.3: Road condition (roughness)……………………………………………….…..53
x Figure 4.4: Alligator cracking and shear failure from failed spot…………………..54 Figure 4.5: Typical failed patching and patch work………………………………..…..54 Figure 4.6: Typical edge break and rutting……………………………………………....54 Figure 4.7: Typical rutting and edge defect and patching……………………………55 Figure 4.8: Typical cracks and surface condition section Chalinze - Morogoro ..………………………………………………………………………………………….56 Figure 4.9: Vehicle composition per section……………………………………………….58 Figure 4.10: Number of vehicles per section………………………………………….……58 Figure 4.11a: Pavement deterioration pattern for section Ubungo - Kimara…....68 Figure 4.11b: Pavement deterioration pattern for section Kimara-Dar/Coast.....68 Figure 4.11c: Pavement deterioration pattern for section Dar/Coast-Mlandizi....69 Figure 4.11d: Pavement deterioration pattern for section Mlandizi-Chalinze…....69 Figure 4.11e: Pavement deterioration pattern for section Chalinze-Morogoro....70 Figure 4.12: Summary of undiscounted road agency economic costs per km….73 Figure 4.13: Summary of undiscounted VOC per km…………………………………….74 Figure 4.14: Discounted Net Benefit for each section……………………………….….74 Figure 4.15: Annual transport costs for optimum strategies………………………….76 Figure 4.16: Annual road agency costs for each section……………………………….76 Figure 4.17: Pavement deterioration pattern for section Mlandizi-Chalinze (mean ESAL 10)…...... 78 Figure 4.18: Pavement deterioration pattern for section Mlandizi-Chalinze (mean ESAL 5)…...... 78 Figure 5.1: Typical used truck (left) and trailer (right) that are found in Tanzania road network ...... 83 Figure 5.2: Stresses and Strains in Pavement structure 1 ...... 84 Figure 5.3: Variation of main stresses with loading time for pavement structure 1 - wheel load 40 kN ...... 86 Figure 5.4: Variation of tensile strains with loading time for pavement structure 1 - wheel load 40 kN ...... 86 Figure 5.5: Variation of tensile stresses with loading time for pavement structure 1 - wheel load 40 kN ...... 87 Figure 5.6: Variation of compressive strains with loading time for pavement structure 1 - wheel load 40 kN ...... 87 Figure 5.7: Variation of main stresses with loading time for pavement structure 1 - wheel load 100 kN ...... 89 Figure 5.8: Variation of tensile strains with loading time for pavement structure 1 - wheel load 100 kN ...... 89 Figure 5.9: Variation of tensile stresses with loading time for pavement structure 1- wheel load 100 kN ...... 90 Figure 5.10: Variation of compressive strains with loading time for pavement structure 1 - wheel load 100 kN ...... 90 Figure 5.11: Variation of main stresses with loading time for pavement structure 1 - wheel load 75 kN ...... 92
xi Figure 5.12: Variation of tensile strains with loading time for pavement structure 1 - wheel load 75 kN ...... 92 Figure 5.13: Variation of tensile stresses with loading time for pavement structure 1 - wheel load 75 kN ...... 93 Figure 5.14: Variation of compressive strains with loading time for pavement structure 1 - wheel load 75 kN ...... 93 Figure 5.15: Stresses and strains in pavement structure 2 ...... 94 Figure 5.16: Variation of main stresses with loading time for pavement structure 2 - wheel load 40 kN ...... 96 Figure 5.17: Variation of tensile strains with loading time for pavement structure 2 - wheel load 40 kN ...... 96 Figure 5.18: Variation of tensile stresses with loading time for pavement structure 2 - wheel load 40 kN ...... 97 Figure 5.19: Variation of compressive strains with loading time for pavement structure 2 - wheel load 40 kN ...... 97 Figure 5.20: Variation of main stresses with loading time for pavement structure 2 - wheel load 100 kN ...... 99 Figure 5.21: Variation of tensile strains with loading time for pavement structure 2 - wheel load 100 kN ...... 99 Figure 5.22: Variation of tensile stresses with loading time for pavement structure 2 - wheel load 100 kN ...... 100 Figure 5.23: Variation of compressive strains with loading time for pavement structure 2 - wheel load 100 kN ...... 100 Figure 5.24: Variation of main stresses with loading time for pavement structure 2 - wheel load 75 kN ...... 102 Figure 5.25: Variation of tensile strains with loading time for pavement structure 2 - wheel load 75 kN ...... 102 Figure 5.26: Variation of tensile stresses with loading time for pavement structure 2 - wheel load 75 kN ...... 103 Figure 5.27: Variation of compressive strains with loading time for pavement structure 2 - wheel load 75 kN ...... 103 Figure 5.28: Fatigue crack propagation shift factor for asphalt layers ...... 105 Figure 5.29: Vehicles per year for section Mlandizi – Chalinze ...... 113 Figure 5.30: ESAL per year for section Mlandizi - Chalinze ...... 113 Figure 5.31: Vehicles per year for section Chalinze - Mlandizi...... 113 Figure 5.32: ESAL per year for section Chalinze – Morogoro...... 114 Figure 5.33: Long term behaviour of cemented material for pavement ...... 116 Figure 5.34: Long term asphalt strain for the pavement structure 1 ...... 116 Figure 5.35: Allowable number of load repetitions for pavement structure 1116 Figure 5.36: Pavement structure 1 for asphalt overlay ...... 118 Figure 5.37: Pavement structure 1 for reconstruction ...... 120
xii List of Tables
Table 3.1: Road network condition (Source: Tanroads Quarterly Report for Financial Year 2008/2009 – 31st December 2008) ...... 28 Table 3.2: Heavy vehicle categories ...... 30 Table 3.3: Traffic load classes ...... 30 Table 3.4: Tanzania climatic zones ...... 31 Table 3.5: Natural gravel material classes ...... 33 Table 3.6: Crushed materials - Material classes ...... 34 Table 3.7: Cemented materials - Material classes ...... 34 Table 3.8: Bituminous base course material classes ...... 35 Table 3.9: Mix proportions for Asphalt Concrete ...... 36 Table 3.10: Subgrade strength classification ...... 36 Table 4.1: Tanzania vehicle fleet ...... Fout! Bladwijzer niet gedefinieerd. Table 4.2: Recommended ESAL and operating weights for HDM-4 analysis…50 Table 4.3: Climatic parameters for Tanzania moderate climatic zone………….56 Table 4.4: Details of Homogenous section……………………………………………...57 Table 4.5: Classification of road works in HDM-4 model……………………………59 Table 4.6; Road work costs…………………………………………………………………….60 Table 4.7: Summary of road works standards…………………………………………..64 Table 4.8: Road works summary report…………………………………………………..71 Table 4.9: Optimum maintenance strategies…………………………………………….75 Table 4.10: NPV results form different discount rates………………………………….80 Table 5.1: Stresses and strains in pavement structure 1- wheel load 40 kN ... 85 Table 5.2: Stresses and strains in pavement structure 1 - wheel load 100 kN 88 Table 5.3: Stresses and strains in pavement structure 1 - wheel load - 75 kN91 Table 5.4: Stresses and strains in pavement structure 2 – wheel load 40 kN . 95 Table 5.5: Stresses and strain in pavement structure 2 - wheel load 100 kN . 98 Table 5.6: Stresses and strains in pavement structure 2 - wheel load 75 kN 101 Table 5.7: Allowable number of load repetitions for asphalt layer – structure 1 ...... 108 Table 5.8: Allowable number of load repetitions for asphalt base layer - structure 1 ...... 108 Table 5.9: Allowable number of load repetitions for cement bound sub-base l ayer - structure 1 ...... 109 Table 5.10: Allowable number of load repetitions for sub-grade - structure 1 ...... 109 Table 5.11: Summary of critical allowable number of load repetitions – structure 1 ...... 109 Table 5.12:Allowable number of load repetitions for asphalt layer- structure 2 110 Table 5.13: Allowable number of load repetitions for unbound granular base layer - structure 2 ...... 110 Table 5.14: Allowable number of load repetitions for the bound base layer - structure 2 ...... 111
xiii Table 5.15: Allowable number of load repetitions for the bound sub-base layer - structure 2 ...... 111 Table 5.16: Allowable number of load repetitions for sub-grade - structure 2 ...... 111 Table 5.17: Summary of critical allowable number of load repetitions - structure 2 ...... 112 Table 5.18: Traffic data for sections Mlandizi – Chalinze and Chalinze - Morogoro ...... 112 Table 5.19: Comparison of allowable and occurring axle loads ...... 115 Table 5.20: Asphalt fatigue life for pavement structure 1 ...... 117 Table 5.21: Asphalt fatigue damages for different overlay thicknesses ...... 119 Table 5.22: Asphalt fatigue damage for surface layer at different base layer thicknesses ...... 120 Table 5.23: Asphalt fatigue damage for base layer at different base layer thicknesses ...... 120
xiv List of Abbreviations and Symbols AADT Annual Average Daily Traffic (vehicle per day) AC Asphalt Concrete Surfacing AMAB Asphalt Mix on Asphalt Base AMGB Asphalt Mix on Granular Base AMSB Asphalt Mix on Stabilized Base CBR California Bearing Ratio CDB Construction Defects Indicator for the Base CDS Construction Defects Indicator for the Bituminous Surfacing CML Central Materials Laboratory CRR Crushed Fresh Rock CRS Crushed Stones and Oversize C1 Cemented Material Stabilized with Lime or Cement, UCS ≥ 1 MPa DBM Dense Bitumen Macadam E80 Equivalent Standard Axle (8160 kg) GM Grading Modulus G80 Natural Gravel with CBR ≥ 80% G60 Natural Gravel with CBR ≥ 60% G45 Natural Gravel with CBR ≥ 45% G25 Natural Gravel with CBR ≥ 25% G15 Natural Gravel with CBR ≥ 15% G7 Natural Gravel with CBR ≥ 7% G3 Natural Gravel with CBR ≥ 3% HDM-4 Highway Development and Management Model ICL Initial Consumption of Lime IRI International Roughness Index IRR Internal Rate of Return LAMBS Large Aggregate Mix for Base LL Liquid Limit LS Linear Shrinkage NPV Net Present Value OMC Optimum Moisture Content PCSE Passenger Car Space Equivalent PI Plastic Index PL Plastic Limit PM Penetration Macadam RIav Annual Average Roughness of the Pavement (m/km IRI) S15 Improved Subgrade Layer with CBR ≥ 15% S7 Improved Subgrade Layer with CBR ≥ 7% S3 Improved Subgrade Layer Natural Gravel with CBR ≥ 80 TFV Aggregate Strength (10% fines value) UCS Unconfined Compression Strength VOC Vehicle Operating Costs
xv 1 INTRODUCTION
1.1 Background Information
An efficient road transport system is seen by most countries as an essential pre-condition for general economic development, (Robinson and Thagesen, 2004). Investment in roads is closely related to reduction in poverty, increases in trade, and generally speaking economic growth. Most roads in developing countries are in pathetic state. Poor roads condition is one of the main causes of Africa’s low competitiveness. Africa’s high transport costs are a major burden on competitiveness and growth. According to Limao and Venables (1999), weak infrastructure accounts for most of Africa’s poor trade performance.
It is evident that Africa needs good roads and already pays for good roads. If Africa doesn’t spend the required maintenance costs for good roads, it is then paying for them through higher vehicle operating costs, travel time, accident costs as well as through opportunity costs for lower economic growth (Kruger et al, 2003). Road maintenance reduces the rate of deterioration, it lowers the cost of operating vehicles on the road by improving the running surface, and it keeps the road open on continuous basis (World Bank, 1988).
In low income developing countries, according to Watanatada (1987), there is a competing demand for very limited resources available. The low and diminishing level of funds available for road maintenance in most developing countries has resulted in an alarming rate of deterioration of the road network as a whole. With the explained benefits obtained for having good roads, it is obvious that the road sector shall receive the priority attention that it deserves and thus, timely maintenance interventions and appropriate pavement rehabilitation is necessary.
It is known that every one-dollar (USD 1.0) of essential maintenance postponed in Sub-Saharan Africa, increases the cost of operating a vehicle in the current period (2003) by more than three dollars (USD 3.0). It is also known that the cost of resealing a road (USD 5000/lane/km) is only 1/4 of the cost of rehabilitating a road (USD 20,000/lane/km) and 1/16 of the cost of reconstructing a road (USD 80,000/lane/km) (Kruger et al 2003).
The case for doing maintenance, appropriately and timely, is thus above any possible debate.
1.2 Problem statement
A pavement structure deteriorates due to combined influences of traffic and environmental loads. This means that at a given moment maintenance activities
1 should be scheduled in order to restore the level of service the pavement should give to the road user. It will be obvious that careful consideration should be given to the planning and selection of the maintenance activities. The right strategy should be applied on the right spot at the right time.
Tanzania operates five (5) modes of transport systems consisting of road, rail, maritime, air and oil pipeline. The road sub-sector accounts for 70% of the national freight movements and over 90% of the passenger movements (Tanroads, 2007). In addition to supporting national economic development, the road transport system acts as a vital transit network for neighbouring landlocked countries of Malawi, Zambia, Uganda, Rwanda, Burundi and the Democratic Republic of Congo (DRC). Tanzania mainland has a road network of about 85,525 km, of which only 5.034 km (5.9%) are paved with a road density of 96.5 km/1000 km2. The road density is above the Africa’s average of 50 km/1000 km2. It is obvious that the road network has an asset value that represents a significant proportion of national wealth, and thus, should make an important contribution to the Gross Domestic Product (GDP). In the year 2003 the GDP growth of Tanzania was 5.1%. However the loss to the economy due to bad roads was estimated to be 9% of the GDP (Haule, 2003). This means that the GDP growth would be much higher if the losses are eliminated through better maintenance and funding.
In many countries, it has proved to be very difficult for maintenance of roads to be carried out effectively, resulting in rapid deterioration of many roads (Robinson and Thagesen, 2004). According to Mikkelsen (1996), in the study of the road network in Tanzania, it was found out that more than 60% of the trunk roads were in unsatisfactorily poor state of repair. One of the challenges identified by the Tanzania Road Fund Board (RFB) is the availability of adequate maintenance funds (Haule, 2003). Analysis of road maintenance funding in Tanzania shows that the funding gap has increased from USD 43.8 million in the year 2000/2001 to USD 75.7 million in 2005/2006 (AfDB, 2007).
With such a poor condition of the road network, with limited and inadequate funding and bearing in mind the importance of having a good road network for the economic growth of the country, an economic evaluation of different road maintenance investment strategies becomes a necessity in order to have scientific criteria for the most economic utilization of the little national resources available.
2 1.3 Objectives of the Study
1.3.1 Main Objective
The main objective of the study is to identify the most economical life cycle maintenance strategies that will form a basis for the maintenance budget programmes on the selected road links.
1.3.2 Specific Objectives
Analyze the effect of different maintenance strategies on the deterioration of the asphalt road pavement
Examine and identify the life cycle optimal maintenance strategy on the study area.
Undertaking sensitivity analysis to analyse the impact of some of the input parameters on the output results of the model analysis.
Estimate the pavement life and to analyse the favourable maintenance options using the mechanistic pavement design analysis method.
1.3.1 Limitation of the Study
This study is based on the case study on the selected road link in Tanzania. The study used the available data at the Tanzania national road authority and additional data surveyed on site.
1.4 Structure of the Report
This report consists of six chapters. Chapter 1 is mainly an introduction to the study where background information, problem statement and objectives of the study are discussed. Chapter 2 discusses the literature survey on typical asphalt pavement distresses, maintenance management and overview of road development and management models commonly applied in planning and management of road maintenance. Chapter 3 discusses the case study in Tanzania. The information on the road sector and maintenance practice in Tanzania is briefly discussed. Observations made during a visual condition survey along the case road are also presented. Chapter 4 discusses the application of the HDM-4 model to analyse the data collected on a study road. Chapter 5 discusses the application of the mechanistic pavement design analysis method to estimate the pavement life and to analyse the favourable maintenance options on the selected road links. Conclusions and recommendations of the study are presented in chapter 6.
3 2 LITERATURE REVIEW
2.1 Introduction
This section presents and discusses the literature surveyed from different sources on issues concerning the maintenance of asphalt roads. In 2.2, some commonly asphalt distresses, including their probable causes and repairs are discussed. Road maintenance management is discussed briefly in 2.3. In 2.4, some of the road development and management models and an overview of the HDM-4 model are presented.
2.2 Asphalt Pavement Distresses
In all cases of pavement maintenance, it is best to determine the distresses or defects and their cause. Determining the distress cause assists in making the proper repair and in preventing the distress from reoccurring. Identification of the distress is one of the first steps in a pavement maintenance program.
There are four major categories of asphalt pavement distresses namely:
Surface defects: raveling, flushing, oxidation and polishing Surface deformation: rutting, shoving, settling and heaving Cracking: transverse, longitudinal, reflective, block and alligator Potholes
Road roughness by definition also implies that it is a defect and is a result of a chain of distress mechanisms and combines the effects of various modes of distress.
This sub-section discusses some of common typical distress modes on asphalt pavements.
2.2.1 Raveling
Raveling or fretting is the progressive disintegration of the asphalt mixture from the pavement surface downward caused by the loss of the dislodged aggregate particles. Raveling is caused by moisture or solvent-inducing stripping of the asphalt binder film from aggregate, oxidation of the asphalt binder, and poor or low compaction during construction or insufficient asphalt binder content in the mixture.
Moisture induced damage or stripping is the most common cause of raveling. Asphalt pavements with high in place air voids and low asphalt binder contents are likely candidates to experience moisture induced damage and raveling.
4 Traffic will cause raveling in wheel paths to accelerate severely. Figure 2.1 below shows a typical raveling defect on an asphalt surface.
Figure 2.1: Typical raveling defect (Section Mlandizi – Chalinze, km 055, 21-04-2009)
Low severity of raveling can be corrected through application of pavement sealer, or fog seal. In high severity raveling, the pavement must receive some type of rehabilitation. On localized severe raveling the damaged section is removed and replaced with a patch. On larger sections of severe raveling, a surface treatment such as a seal coat or slurry seal, or even an asphalt inlay or overlay may be required.
2.2.2 Flushing
Flushing is the presence of excess asphalt binder on the pavement surface. Flushing appears as a film of asphalt binder on the pavement surface. Flushing is the result of free or excessive asphalt binder migrating upward to the pavement surface. Excessive asphalt binder content in the mixture or very low air voids content can cause flushing. Soft asphalt binders used in hot climates will also contribute to flushing. Flushing typically occurs in vehicle wheel paths and can be accelerated by traffic and hot weather. Figure 2.2 below show a typical flushing caused by the application of the soft asphalt binder.
Low severity flushing is usually not corrected until it becomes medium severity flushing. High severity flushing can be corrected by the removal of the wearing course. A sand coat can also be used to cover up the flushing, however this is somewhat of a temporary solution.
5
Figure 2.2: Typical flushing with associated rutting (Section Mlandizi – Chalinze, km 099, 21-04-2009)
2.2.3 Rutting
Rutting is surface deformation that is a depression in the vehicle wheel paths. Rutting is sometimes called grooving or channelling. Traffic compaction or displacement of unstable asphalt mixtures causes rutting. Rutting may also be caused by base or subgrade consolidation. Rutting is historically a primary criterion of structural performance in many pavement design methods.
Displacement or plastic flow type rutting is related to the design of the asphalt mixture. Low design air voids, excessive asphalt binder, excessive sand or mineral filler, rounded aggregate particles and low voids filled in mineral aggregate (VMA) can all contribute to displacement rutting. This typical rutting can be observed by the occurrence of the heaves along the rut. Figure 2.3 below shows this typical type of rutting.
Figure 2.3: Typical plastic flow rutting at Chalinze (Section Mlandizi – Chalinze, km 099, 21-04-2009)
6 Consolidation rutting is caused by consolidation of the asphalt pavement including an uncompacted base or subgrade. Consolidation rutting can also occur in the mainline or wearing course, when not enough construction compaction has been completed. Proper compaction of all components of the pavement structure will prevent consolidation rutting.
Rutting can also be caused by excessive surface wear on the wearing course. A typical example is caused by abrasion of tire chains or studded tires. Abrasion rutting will also appear as raveling in the wheel paths.
Mechanical deformation is rutting in the wheel path caused by insufficient structural strength in the pavement. It is caused by overloading, insufficient pavement thickness, or a strength deficient asphalt mixture. Mechanical rutting is usually accompanied by longitudinal or alligator cracking. Figure 2.4 below shows a very typical rutting occurring on the asphalt pavement, probably caused by insufficient strength of the sub-grade or a thinner pavement layer.
Figure 2.4: Rutting along the wheel paths (left) and simple rut depth measurement (right) (Section Chalinze – Morogoro, at Mikese weighbridge station, 2009)
Low severity rutting is usually not addressed until it becomes medium or severe rutting. Medium and severe rutting can be corrected by milling off the existing pavement, to a depth below the rut, and then replacing the pavement. The base or sub-grade may need to be reconstructed in cases of severe consolidation rutting. Coring or transverse full depth trenching of the pavement can usually reveal which layer the rutting is occurring in and to what extent.
2.2.4 Shoving and Surface distortions
Shoving is a surface deformation that involves the wearing course to be displaced transversely across the pavement. Shoving is permanent displacement caused by traffic loading. Shoving typically occurs at intersections or entrance/
7 exit of highways. The braking action of trucks and automobiles is the cause of shoving. Shoving, corrugations, and distortions are forms of plastic deformations within the asphalt mixture. The mixture lacks stability for its intended application. The low stability can be due to low air voids content, excessive asphalt binder content, excessive sand or mineral filler, round aggregates, or an asphalt binder that is too soft for the application.
The repair of shoving or corrugations requires milling of the corrugations and replacing the material with a patch or overlay.
2.2.5 Transverse Cracking
An asphalt pavement will crack when temperature or traffic generated stresses and strains exceed the fatigue or tensile strength of the compacted asphalt mixture. Transverse cracking is a crack that extends across the pavement perpendicular to the pavement centreline. Pavement movement due to temperature changes and aging related shrinkage of the asphalt binder causes transverse cracks. Loading or traffic does not cause temperature induced transverse cracks. Non load induced transverse cracking can also be called temperature cracking. During periods of cooling, the asphalt pavement tries to contract and the pavement for most practical purposes is laterally restrained. Thermal stresses develop and exceed the tensile strength of the asphalt mixture, resulting in a transverse crack. Transverse cracks typically appear in asphalt pavements 2 to 7 years after construction. In case of cement bound base that exhibits transverse cracking due to shrinkage and temperature decrease, reflective cracking through asphalt layer may occur.
Selecting the proper type and amount of the asphalt binder for the climate the pavement is located will reduce the occurrence of temperature induced transverse cracks. High severity cracks are routed and filled, but it is usually more feasible and economical to remove the damaged section of pavement and patch it. Medium severity cracks are sealed with crack filler. Low severity transverse cracks are usually not filled.
2.2.6 Longitudinal Cracking
Longitudinal cracking is cracking that appears parallel to the centreline of the pavement. Longitudinal cracks are either load induced or non load induced. Load induced longitudinal cracks occur in the wheel path or loading area of the pavement. Non loading induced longitudinal cracks can occur anywhere throughout the pavement, but are typically in the centre or at the edge of the pavement. Load induced cracks are a form of fatigue cracking which eventually evolve into Alligator cracking. Non load induced cracks usually occur where any longitudinal construction joint is present. A crack can also occur in a longitudinal
8 direction at patches and other pavement repair. A longitudinal crack underneath the pavement can also reflect up to the surface of the pavement. Longitudinal cracking can occur also at locations in pavement that exhibit segregation of the asphalt mixture. Figure 2.5 below shows some typical cracking that can be observed on the asphalt pavements.
Figure 2.5: Typical cracking on asphalt pavement (Section Ubungo – Mlandizi, km 004 left and km 014 right, 2009)
Repairing of longitudinal cracking is similar to that of transverse cracking explained above. Significant longitudinal cracking occurring at the pavement edge require the investigation of the surrounding support and the adequacy of the drainage.
2.2.7 Alligator Cracking
Alligator cracks are interconnecting cracks that form such pieces ranging in size from 25 to 150 mm. The cracks have the same appearance as the pattern on the skin of an alligator. Alligator cracking only occurs from loading or fatigue. The cracks are caused by the complete failure of the pavement due to traffic loading or inadequate base or sub-grade support. When the cracks are caused by traffic loading, they only occur in areas of repeated loading such as wheel paths. Alligator cracking due to inadequate base or sub-grade support have more random appearance throughout the pavement that receives loading. Alligator cracking is a symptom of insufficient structural strength in the pavement, weak sub-grade, or overloading of the pavement. Alligator cracking will progress into pavement disintegration, usually in the form of potholes. Figure 2.6 shows some examples of typical alligator cracking on the asphalt pavement.
9
Figure 2.6: Typical alligator cracking on asphalt pavement (Section Mlandizi – Chalinze, km 056, 21-04-2009)
A permanent solution would require removing the damaged section, repairing subgrade if needed and the installation of a full depth patch. If the pavement overloading is the cause of the alligator cracking, the solution requires the pavement to receive an overlay or to be reconstructed to proper thickness for the applicable loading.
2.2.8 Potholes
A pothole is a cavity in the road surface which is 150 mm or more in average diameter and 25 mm or more in depth (Paterson, 1987). The dimensions are the minimum that affect the motion of a car wheel and measured roughness significantly. Shallower depths are typical for ravelling. Potholes are the most visible and severe form of pavement distress. Potholes are caused by weaknesses in the pavement surface resulting from failure of the base or subgrade, poor drainage or a structurally deficient pavement structure. High severity alligator cracking will continue to become a pothole or a series of potholes. A typical pothole on the asphalt pavement is shown in figure 2.7 below.
Figure 2.7: Pothole on asphalt pavement (Section Mlandizi – Chalinze, km 068, 21-04-2009)
10 Potholes repair is completed by either filling the pothole with a patching or pothole filling mixture or removing the pothole and the immediate surrounding distressed pavement and replacing with a full depth patch.
2.2.9 Road Roughness
Road roughness is the deviations of a travelled surface from a true planar surface with characteristic dimensions that affect vehicle dynamics, ride quality, dynamic loads and drainage (American Society for Testing and Materials (ASTM) Specification E867-82A). Pavement roughness is the result of a chain of distress mechanisms and combines the effects of various modes of distress. Figure 2.8 below shows the mechanisms and interactions of distress in paved roads. Road roughness emerges as a key property of road condition to be considered in any economic evaluation of design and maintenance standards for pavements. Roughness affects the dynamics of moving vehicles, increasing the wear on vehicle parts and the handling of a vehicle. It also increases the dynamic loadings imposed by vehicles on the surface, accelerating the deterioration of the pavement structure.
Methods of measuring road roughness are categorized into four groups’ namely Absolute profile, Moving datum profile instruments, Vehicle-motion instruments (road meters) and Dynamic profile instruments. Roadmeters are the most common instruments.
Figure 2.8: Pavement deterioration concept (Source: Archondo, 2008)
11 The International Roughness Index (IRI) is adopted as a standard measure of roughness. IRI is a mathematically defined summary statistic of the longitudinal profile in the wheel path of a travelled road surface. The index is representative of the vertical motions induced in moving vehicles for the frequency bandwidth which affect both the response of the vehicle and the comfort perceived by occupants. IRI describes a scale of roughness which is zero for a true planar surface, increasing to 6 for moderate rough paved roads, 12 for extremely rough paved roads with potholing and patching, and up to about 20 for extremely rough unpaved roads. The units are dimensionless, but it has been scaled by a factor of 1000 so that it represents i.e. m/km, mm/m e.t.c. The standard presentation is thus e.g. 3.1 m/km IRI, generally reported to one decimal place.
2.3 Maintenance Management of Asphalt Pavements
Maintenance reduces the rate of pavement deterioration, it lowers the cost of operating vehicles on the road by improving the running surface, and it keeps the road open on a continuous basis (World Bank, 1988). In Overseas Road Note 15, road management is defined as the process of maintaining and improving the existing road network to enable its continued use by traffic efficiently and safely, normally in a manner that is effective and environmentally sensitive; a process that is attempting to optimize the overall performance of the road network over time.
2.3.1 Management Functions
The functions involved in maintenance management can be described under four headings (Robinson et al. 1998) explained as follows:
Strategic Planning: These are long term decisions affecting the whole of the road network, undertaken primarily for the benefit of senior managers and policy makers.
Programming: Determining those parts of the road network where work can be undertaken with available resources in the next budget period.
Preparations: Design of works for individual road sections, issuing of contracts or works orders for works that have a budget commitment.
Operations: Managing and supervising on-going works in individual subsections of road.
12 Maintenance management simply aims to get the right resources, to the right place on the road network, to carry out the right maintenance or renewal work, at the right time.
2.3.2 Management Cycle
Traditionally, in many road organizations, budgets and programmes for road works have been prepared on a historical basis, in which each year’s budget is based upon that for the year before, with an adjustment for inflation. Under such a regime, there is no way of telling whether funding levels, or the detailed allocations, are either adequate or fair. There is a requirement for an objective needs–based approach, using relevant knowledge, structure and condition of the roads being managed. Primary functions of planning, programming, preparation and operations provide a suitable framework within which a need-based approach can operate.
Within each of the four primary management functions, a common system can be used. An appropriate approach is to use the concept known as the management cycle illustrated in Figure 2.9 below.
Figure 2.9: Road management cycle (Source: Robinson et al.1998).
A cycle provides a series of well – defined steps that take the management process through the decision making activities. The process typically completes
13 the cycle once in each period covered by the management function. The cycle also indicates the importance of the management information and the linkage of each step in the cycle to the management information database.
2.3.3 Classification of Maintenance Activities
Maintenance activities may be classified in terms of their operational frequency into the following:
Routine Maintenance: Covers activities that must be carried out frequently, that is, once or more per year. They are typically small scale, or simple and often widely dispersed. Two types can be differentiated, reactive (those whose frequency depends on the volume and intensity of traffic and/or pavement condition) and cyclic (those whose need is independent of traffic, and whose frequency relates to rainfall, topography and other local environmental aspects)
Periodic Maintenance: Describes activities that are needed occasionally, that is, after a period of some years. They are usually large scale and require more equipment and skilled labourers than routine maintenance activities.
Urgent Maintenance: Comprises emergency repair
Maintenance activities for asphalt pavements consist of sanding, local sealing, crack sealing, filling depressions, surface patching, and base patching for the routine activities and surface dressing, fog spray and slurry seal, asphalt overlays and pavement reconstruction for the periodic activities.
2.3.4 Life Cycle Costing
Money spent in maintenance should be treated as an investment in the same way as for that spent on new construction. Cost–benefit analysis is an appropriate tool for making decisions about maintenance expenditures. It is important for maintenance activities to consider the impact on the life of the works and the resulting future cost streams. Thus, the application of cost- benefit principles to decisions about maintenance investments implies consideration of the concepts of life cycle costing.
The adoption of higher engineering or maintenance standards normally leads to higher investment costs, but may result in lower costs to the road agency in terms of future costs of maintenance and renewal and will certainly result in lower road user costs.
14 Transport cost components to be considered in a life cycle costing analysis include the road agency costs and the road user costs which are explained below:
Road agency costs: Include the costs of management, operations, labour, equipments, materials, maintenance and rehabilitation.
Road user costs: Include vehicle operating costs, travel time costs and road accident costs.
Some of the above costs are fixed costs while others are variable, and they depend on many factors including the standard of the road concerned. Figure 2.10 below shows the typical transport costs components related to the road standard.
Figure 2.10: Transport cost components (Source: Kerali, 2008)
2.3.5 Cost-Benefit Analysis
The purpose of carrying out cost-benefit analyses is primarily to ensure that an adequate return in terms of benefits results from committing expenditure. An additional purpose is to ensure that the investment option adopted gives the highest return in relation to the standards adopted, and the timing of the investment. For the economic appraisal, the assessment is made in terms of the net contribution that the investment will make to the country’s economy as a whole. The essential approach for the economic appraisal is to use the
15 opportunity cost of the investment as a measure of resource rather than market prices.
Costs and benefits are compared to determine whether the investment is worthwhile, and to identify which is the best of the alternatives being considered. The alternative in which minimum investment takes place is sometimes known as the baseline alternative and is compared against other investment alternatives.
In order to carry out an economic analysis, it is necessary to make adjustments to costs and prices to ensure that they are all measured in the same units. A first step is usually to remove the effect of inflation to enable values to be compared on the same basis over time. It is also necessary to factor costs and benefits to take account of the difference times during the analysis period.
All future costs and benefits are discounted to convert them to present values of costs and benefits using the following formula:
Cn PVC n ……..………………………………………..……………………….. 2.1 1 r 100
Where: PVC = Present Value of Costs and Benefits
Cn = costs or benefits incurred in year n r = discount rate expressed as percentage n = year of analysis, for base year, n = 0.
2.3.5.1 Comparison of Alternatives
The most commonly used criteria for selecting the best investment option are the Net Present Value (NPV), the Internal Rate of Return (IRR) and the Benefit Cost Ratio (BCR) which are explained as follows:
Net Present Value (NPV)
NPV is simply the difference between the discounted benefits and costs over the analysis period. NPV is a measure of the economic worth of an investment.
t1 b C NPV n n …………………………………………………….…..………. 2.2 n n0 1 r 100
Where: t = the analysis period in years n = current year, with n = 0 in the base year
bn = the sum of all benefits in year n
16 Cn = the sum of all costs in year n r = the planning discount rate expressed as a percentage.
A positive NPV indicates that the investment is justified economically at the given discount rate and the higher the NPV, the greater the benefits or the lower will be the costs.
Internal Rate of Return (IRR)
The IRR is the discount rate at which the present values of costs and benefits are equal; in other words, the NPV = 0. IRR is determined by solving the following equation for r : t1 b C n n 0 …………………………………………………………………...... 2.3 n n0 1 r 100
With all the notations as explained above.
IRR gives no indication of the size of the costs or benefits of an investment, but acts as a guide to its profitability. If IRR is higher than the discount rate, then the investment is economically justified.
Benefit Cost Ratio (NPV/Cost ratio)
This ratio represents the magnitude of the return to be expected per unit of investment and is therefore a measure of the efficiency of an investment. Given a constrained budget situation, the most efficient investment is that with the largest NPV/Cost ratio.
2.4 Road Development and Management Models
Road development and management models can be used to assist with the economic appraisal, the preparation of maintenance and investment programmes, and with strategic planning. Typically, these models estimate road construction, maintenance and user costs for a specified analysis period to enable the life cycle costs and benefits to be determined for different sets of assumptions. Models simulate the interaction between pavement construction standards, maintenance standards and the effects of the environment and traffic loading in order to predict the annual trend in road condition. Over the years, increasing efforts have been made to develop and implement improved road management and planning tools (Paterson (1987).
17 2.4.1 Past Development on Road Models
The first move towards producing a road project appraisal model was in 1968 by the World Bank (WB). The first model was produced in response to terms of reference for a highway design study produced by the WB in conjunction with the Transport and Road Research Laboratory (TRRL) and the Laboratoire Central des Ponts et Chaussees (LCPC). Thereafter, the Highway Cost Model (HCM) was developed by the Massachusetts Institute of Technology (MIT) in 1971.
TRRL, in collaboration with the WB, undertook a major field study in Kenya between 1971 and 1975 to investigate the deterioration of paved and unpaved roads as well as factors affecting the vehicle-operating costs in developing countries environment. The results of this were used by TRRL to produce the first prototype version of the Road Transport Investment Model (RTIM) for developing countries (Abaynayaka, 1977). In 1976, the first version of the Highway Design and Maintenance Standards Model (HDM) was produced by MIT (Harral, 1979).
Further work was undertaken in a number of countries to extend the geographic scope of RTIM and HDM models which include the Caribbean study by TRRL, the India study by the Central Road Research Institute (CRRI) and the Brazil study funded by the United Nations Development Program (UNDP).
The results of the TRRL studies were used to develop the RTIM2 model (Parsley and Robinson, 1982), whilst a more comprehensive model, HDM-III incorporating the findings from all previous studies, was developed by the WB (Watanatada et al., 1987). Both models were originally designed to operate on mainframe computers. As computer technology advanced, the University of Birmingham (Kerali et al., 1985) produced a microcomputer version of RTIM2 for TRRL. WB later produced HDM-PC, a microcomputer version of HDM-III (Archondo-Callao and Purohit, 1989).
TRRL produced RTIM3 in 1993 and in 1994 the WB produced two further versions of HDM which are HDM-Q incorporating the effects of traffic congestion into the HDM-III program (Hoban, 1987) and HDM manager, providing a menu- driven front end to HDM-III (Archondo-Callao, 1994). In 2000, the Highway Development and Management model (HDM-4) was produced by the WB.
Figure 2.11 shows the historical development of the road models.
18
Figure 2.11: History of the HDM Model (Source: Kerali, 2008)
2.4.2 Road Network Evaluation Tools (RONET)
The Sub-Saharan Africa Transport Policy Program (SSATP), which is a collaborative framework set up to improve transport policies and strengthen the institutional capacity in the African region, has developed the model Road Network Evaluation Tools (RONET). The latest version 2 of the model was release on January 2009.
RONET is a model which can be used to appreciate the current state of the road network, its relative importance to the economy and to compute a set of monitoring indicators to asses the performance of the road network under different road maintenance standards. RONET is developed from the same principle overlying HDM-4, adopting simplified road user costs relationships and simplified road deterioration equations derived from the HDM-4 research.
The characteristic of RONET is the use of simplified road deterioration and road user costs relationships, the restricted way of defining the standards, the in- ability to evaluate improvement standards, and the lack of a budget constraints optimization module.
2.4.3 The Highway Development and Management Model (HDM-4)
Since this research will make use of the application of the HDM-4 model, this sub-section is introduced to give a brief overview of the model. As discussed in the preceding sections, various versions of different models have been widely used in a number of countries, and have been instrumental in justifying increased road maintenance and rehabilitation budgets. However, it was recognized that there was a need for a fundamental redevelopment of the various models to incorporate a wide range of pavements and conditions of use.
19 There was a need to incorporate the results of the extensive research undertaken around the world. In the case of the vehicle operation costs (VOC), it was recognized that vehicle technology has improved dramatically since 1980 with the result that typical VOCs could be significantly less than those predicted by the RTIM3 and the HDM-III models. It is against this background that the development of HDM-4 was undertaken.
2.4.3.1 HDM-4 Analytical Framework The HDM-4 analytical framework is based on the concept of pavement life cycle analysis. This applied to predict the following over the life cycle of a road pavement, which is typically 15 to 40 years:
Road deterioration Road work effects Road user effects Social-economic and environmental effects
Once constructed, road pavements deteriorate as a consequence of several factors, most notably are traffic loading, environmental weathering and the effect of inadequate drainage systems. The rate of pavement deterioration is directly affected by the standards of maintenance applied to repair defects.
The overall long-term condition of road pavements directly depends on the maintenance or improvement standards applied to the road. Figure 2.12 illustrates the predicted trend in pavement performance represented by the riding quality that is often measured in terms of the International Roughness Index (IRI). When the maintenance standard is defined, it imposes a limit to the level of deterioration that a pavement is permitted to attain. Consequently, in addition to the capital costs of road construction, the total costs that are incurred by road agencies will depend on the standards of maintenance and improvement applied to the road network.
Figure 2.12: The concept of the life-cycle analysis in HDM-4 (Source: Kerali, 2008)
20 The impacts of the road condition, as well as the road design standards, on road users are measured in terms of road user costs, and other social-environmental effects.
Economic benefits from road investments are then determined by comparing the total cost streams for various road works and construction alternatives against a base case (without project or do minimum) alternative, usually representing the minimum standard of routine maintenance.
2.4.3.2 HDM-4 Applications
Strategy Analysis
The concept of strategic planning of medium to long term road expenditures requires that a road organization should consider the requirements of its entire road network asset. Thus, strategy analysis deals with the entire network or sub- networks managed by one road organization.
Strategy analysis may be used to analyze a chosen network as a whole, to prepare medium to long term planning estimates of expenditure needs for road development and conservation under different budget scenarios.
Programme Analysis
This deals primarily with the prioritization of a defined long list of candidate road projects into a one-year or multi-year work programme under defined budget constraints. The candidate road projects are selected as discrete segments of a road network. When all the candidate projects have been identified, the HDM-4 programme analysis application can then be used to compare the life cycle costs predicted under the existing regime of pavement management (without project case) against the life cycle costs predicted for the periodic maintenance, road improvement or development alternative (with project case).
Project Analysis
Project analysis concerns the evaluation of one or more road projects or investment options. The application analyses a road link or section with user- selected treatments, with associated costs and benefits, projected annually over the analysis period. Economic indicators are determined for different investment options.
Project analysis may be used to estimate the economic or engineering viability of road investment projects.
21 2.4.3.3 HDM-4 Modules
The overall structure of HDM-4 is illustrated in Figure 2.13 below. The three analysis tools (Strategy, Programme and Project) operate on data defined in one of the four data managers:
Road Network: Defines the physical characteristics of road sections in the network or sub-network to be analyzed
Vehicle Fleet: Defines the characteristics of the vehicle fleet that operates on the road network to be analysed
Road Works: Defines maintenance and improvement standards, together with their unit costs, which will be applied to the different road sections to be analyzed.
HDM Configuration: Defines the default data to be used in the applications. A set of default data is provided when HDM-4 is first installed, whereby users modify them to reflect local environments and circumstances.
Figure 2.13: HDM-4 System architecture (Source: Kerali, 1999)
Technical analysis within the HDM-4 is undertaken using four sets of models discussed below:
RD (Road Deterioration): Predicts pavement deterioration for bituminous, concrete and unpaved roads.
22 WE (Works Effects): Simulates the effects of road works on pavement condition and determines the corresponding costs.
RUE (Road User Effects): Determines costs of vehicle operation, road accidents and travel time.
SEE (Social and Environment Effects): Determines the effects of vehicle emissions and energy consumption.
The model simulates, for each road section, year by year, the road condition and resources used for maintenance under each strategy, as well as the vehicle speeds and the physical resources consumed by vehicle operation. After physical quantities involved in construction, road works and vehicle operation are estimated, user-specified prices and unit costs are applied to determine the financial and economic costs. Relative benefits are then calculated for different alternatives, followed by the calculation of the present value and the rate of return of consumptions.
23 3 CASE STUDY IN TANZANIA
3.1 Introduction
It has been discussed in chapter 1 that this study is based on the case study that has been done in Tanzania. In this chapter the discussion on the road sector and the maintenance practices will be briefly highlighted.
In 3.2 the general important information about Tanzania is presented. The location, area, population, the climate and topographical conditions are briefly discussed. The Tanzania road sector is discussed in 3.3. The discussion includes the road network, network condition as well as the revenues for the road maintenance. The design and maintenance standards in Tanzania are also presented in 3.3. This includes a brief discussion on the Pavement and Materials Design Manual of 1999 prepared by the then Ministry of Works.
The visual road condition survey on the case road is discussed in 3.4. The observed typical pavement distresses along the road are presented and discussed. A brief conclusion of this chapter is given in 3.5.
3.2 General Information
The United Republic of Tanzania is located in Eastern Africa between longitude 29˚ and 41˚ east, latitude 1˚ and 12˚ south. It was formed out of the union of two sovereign states namely Tanganyika (nowadays referred to as Tanzania Mainland) and Zanzibar. Tanzania has an estimated total area of 945,087 square kilometres and a population of about 41.5 million people (UN, 2008).
The Tanzania climate varies from tropical along the coastline to temperate in the highlands.
Tanzania has four main topographic types which are lowlands (mainly the coastal plain below 200 m above sea level with isolated hills up to 300 m in height), broad nearly flat areas of inland drainage, plateau and highlands. The highlands include mountains with an altitude generally between 1500 m and 3000 m. The plateau rises from the coastal plains to an altitude between 1000 m and 1500 m, adjoining the highlands. The highest point is 5,895 m above sea level, which is the peak of Africa’s highest mountain, Kilimanjaro.
Except for the mountainous and steep rolling terrain of the highlands, the terrain is generally flat to gently rolling in the plateau and lowlands.
24 3.3 Tanzania Road Sector
Tanzania’s vision 2025 with respect to the road sector envisages a plan to undertake an economic transformation that will enable it to move from the category of least developed countries to a medium income country. In the process the transport sector is expected to have an extensive road network that is well maintained, serving all parts of the country as well as neighbouring countries.
Since 1998 the Tanzania road sector has been undergoing far-reaching reforms. The establishment of the Road Fund Board (RFB) in 1998, a dedicated Road Fund to finance road maintenance and a semi autonomous agency, Tanzania National Roads Agency (Tanroads) in 2000 to manage the road network is part of the ongoing road sector reform process.
The planning and programming of the road network in Tanzania is a shared responsibility of the Ministry of Infrastructure Development (MoID) and the Prime Minister’s Office-Regional Administration and Local Government (PMO-RALG). The overall financing is coordinated by the Ministry of Finance. Development and maintenance of infrastructure in Tanzania is financed through budgetary allocation and assistance from the development partners.
3.3.1 Road Network
Tanzania mainland classifies the public road network into two categories namely National roads and District roads.
The National roads include the following:
Trunk road which is primarily a National route that links two or more regional headquarters, or an international through route that links regional headquarters and another major or important city or town or major port outside Tanzania.
Regional road which is a secondary road that connects a trunk road and a district or regional headquarters, or connects a regional headquarter and district headquarters.
The District roads include the following:
Collector road which is a road linking a district headquarter and a division centre or linking a division centre with any other division centre. Also, a route linking a division centre with a ward centre or a road within urban area carrying through traffic which predominantly originates from
25 and destined out of the town and links with either regional or a trunk road.
A feeder road which shall be a road within urban area that links a collector road and another minor road within the vicinity and collects or distributes traffic between residential, industrial and principle business centres of the town. Also a village access road linking wards to other wards centres.
Community road within a village or a road which links a village to another village.
All Trunk and Regional roads are managed by the Ministry of Infrastructure Development (MoID) through Tanroads. Urban, District and Feeder roads are managed by the Prime Minister’s Office for Regional Administration and Local Government (PMO-RALG).
By 2007, Tanzania had a road network of about 85,525 km, 28,900 km of which consisted of trunk and regional roads. The asset value of the road network is estimated to be US dollars 2.6 billions (RFB, 2007). The overall road density is 96.5 km/1000 km2. The total length of the paved road network is 5034 km which is about 5.9% of the entire road network. Figure 3.1 shows the paved and unpaved road network. The map showing the trunk road network is shown in Figure 3.2.
Road Network
40000 30000 20000 10000
Length [km] Length 0 Trunk Regional District Feeder Urban Paved 3917 327 0 0 790 Unpaved 6027 18629 29537 21191 5107 Total 9944 18956 29537 21191 5897 Road Category
Paved Unpaved Total
Figure 3.1: Paved and unpaved road network (Source: Road Fund Board, 2007)
26
Figure 3.2: Tanzania Trunk road network (Source: Tanroads, 2009)
In December 2008, the overall road network condition assessment for the Trunk and Regional roads indicated that 57% of the network was in a good condition, 34% in fair condition and 9% in poor condition (Tanroads, 2009). Table 3.1 shows the road network condition as per the assessment of December 2008.
27 Table 3.1: Road network condition (Source: Tanroads Quarterly Report for Financial Year 2008/2009 – 31st December 2008)
Road Category Good (km) Fair (km) Poor (km) Trunk roads Paved 2,832.44 (72%) 901.02 (23%) 164.98 (4%) Unpaved 3,069.80 (51%) 2,485.17 (41%) 476.73 (8%) Total 5,902.24 (59%) 3,386.19 (34%) 641.71 (6%) Regional roads Paved 293.2 (90%) 31.93 (9%) 2.36 (1%) Unpaved 10,402.67 (56%) 6,320.10 (34%) 1,934.14 (10%) Total 10,701.87 (56%) 6,352.03 (34%) 1,936.49 (10%)
Total roads 16,604.11 (57%) 9,738.22 (34%) 2,578.20 (9%)
RFB revenues come from different sources that include the following: Fuel levies on diesel and petrol Transit fees Vehicle overloading fees Money from any other source at the rate to be determined by Parliament from time to time.
The RFB revenues have increased from Tanzanian shillings, Tshs. 47.252 billions (USD 36.35 million) in the financial year, FY 2000/2001 to Tshs. 217.5 billion (USD 167.31 million) in the FY 2007/2008. Figure 3.3 shows the annual road fund revenues collected. The strong increase of the revenues from the FY 2006/2007 to the FY 2007/2008 is explained by the increase of fuel levy from Tshs. 100 per litre of petrol and diesel to Tshs. 200 from July 2007.
Road Fund Revenues
300.00
200.00
100.00 Revenues 0.00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 USD [million] 36.35 40.68 45.68 51.80 56.31 56.22 77.86 167.31 Tshs. [billion] 47.25 52.88 59.39 67.34 73.20 73.08 101.22 217.50 Financial Year
USD [million] Tshs. [billion]
Figure 3.3: Annual Road Fund Revenues (Source: RFB, 2008)
28 Despite the increase in road fund revenues and the support from donor communities, it has been reported that in the FY 2007/2008 there was a budget deficit of about Ths. 3,343 million (USD 2.6 million) for the road maintenance of Trunk and Regional roads (Tanroads, 2007). It has to be noted also that the presented revenues have not been corrected for the inflation.
3.3.2 Design and Maintenance Standards
Tanzania through the then Ministry of Works (MoW) has prepared the Pavement and Materials Design Manual -1999 for design of new roads and rehabilitation of existing roads in order to standardize design practices in the country. The manual has particular reference to the prevailing conditions in Tanzania.
In addition to the Pavement and Materials Design Manual, the Ministry has also prepared the Standard Specifications for road works and Laboratory Testing Manual in 2000 to be used for the road works and laboratory testing procedures respectively. The test procedures are in essence based on British Standard (BS) for testing of soils, aggregates and concrete while American Society for Testing and Materials (ASTM) has been adopted for asphalt testing.
The pavement design is based on minimum sub-grade strength of California Bearing Ratio (CBR) 15% that shall be achieved by improved sub-grade layers where necessary. The design is based on traffic loading, defined into seven traffic load classes. The consequences of heavy axle loads (above 13 tonnes) are shown in the procedure, which puts restrictions on certain material types in the base course under such conditions.
Pavement rehabilitation is based on measured properties of materials and thickness of layers in the existing pavement or, alternatively, on criteria for maximum surface deflection. Based on either of the two criteria mentioned above, a design for overlays is provided, alternatively partial or full reconstruction is employed depending on the condition of the existing pavement.
It has been however emphasized that under no circumstances shall the Manual waive professional judgement in applied engineering.
3.3.2.1 Traffic Loading
Only heavy vehicles are considered for the purpose of structural pavement design. Heavy vehicles are defined as those having a registered un-laden weight of 3 tonnes or more. Large buses having a seating capacity of 40 or more are included as heavy vehicles. Heavy vehicles are grouped into categories given in Table 3.2.
29 Table 3.2: Heavy vehicle categories
Heavy Vehicle Category Definition - 2 axles, including steering axle, and Medium Goods Vehicle (MGV) - 3 tonnes empty weight, or more - 3 axles, including steering axle and Heavy Goods Vehicle (HGV) - 3 tonnes empty weight, or more - 4 or more axles, including steering axle and Very Heavy Goods Vehicles (VHGV) - 3 tonnes empty weight, or more
Buses Seating capacity of 40, or more
The Equivalent Standard Axle Load (ESAL) for every vehicle in the axle load survey is determined and an average value is subsequently calculated for each heavy vehicle category and for each lane separately. The average ESAL values for the heavy vehicles are shown on Table 4.2. This ESAL is then applied to the results from traffic counts to give the cumulative equivalent standard axles (E80) traffic loading the pavement is subjected to over a given period.
The proportion of the design traffic loading (E80) as a result of axles loaded to above 13 tonnes is calculated from axle load data. If the proportion is 50% or higher then the design traffic loading is defined as Heavy, denoted by the index H to the Traffic Load Class as input to the pavement design.
After determining the design traffic loading, E80, and the heavy axle’s proportion of E80, the values are placed into their correct class in accordance to Table 3.3.
Table 3.3: Traffic load classes
Design Traffic Loading [E80 X 106] Traffic Load Class [TLC]
< 0.2 TLC 02 0.2 to 0.5 TLC 05 0.5 to 1 TLC 1 1 to 3 TLC 3 3 to 10 TLC 10 10 to 20 TLC 20 20 to 50 TLC 50
Where the heavy (>13 tonnes) axles’ proportion of E80 is 50% or higher the TLC is given an index as follows:
TLC 05-H TLC 1-H TLC 3-H TLC 10-H TLC 20-H TLC 50-H
30 3.3.2.2 Climatic Zones
For the purpose of pavement design, Tanzania is divided into three climatic zones:
A dry zone in the interior A large moderate zone Several wet zones, mainly at high altitudes
The three climatic zones are shown in Figure 3.4. The climatic zones are demarcated on the basis of the number of months in a year with surplus of rainfall over potential evaporation as shown in Table 3.4.
Table 3.4: Tanzania climatic zones
Climatic Zone Number of months per year with higher rainfall than evaporation Dry Less than 1 month Moderate 1 to 3 months Wet More than 3 months
Figure 3.4: Tanzania climatic zones
31 The Figures 3.5 and 3.6 show the average monthly air temperatures and rainfall for the Dar es Salaam city which is taken as a representative location for the study area.
Average Monthly Temperatures for Dar es Salaam
40
30
20
10
0
AverageTemperature ['C] Jan Febr. March April May June July August Sept. Oct. Nov. Dec. High 30 31 30 29 28 28 27 27 28 30 30 30 Low 25 24 23 23 22 20 19 19 19 21 22 24 Month
Figure 3.5: Average monthly air temperature for Dar es Salaam city (Source: www.weatherintanzania.com – May, 2009)
Average Monthly Rainfall for Dar es Salaam
300 250 200 150 100
50 Average[mm] Rainfal 0 Jan Febr. March April May June July Augus Sept. Oct. Nov. Dec. Series1 70 60 120 260 180 30 20 20 20 40 80 90 Month
Figure 3.6: Average monthly rainfall for Dar es Salaam city (Source: www.weatherintanzania.com – May, 2009)
3.3.2.3 Pavement Materials
Materials used in structural layers of pavement are selected according to criteria of availability, economic factors and experience. In the Pavement Materials and
32 Design Manual all materials are indicated by means of codes such as G80, C2, CM etc. which refer to materials with certain defined properties. Material types commonly used in Tanzania are categorized into Unbound materials that include natural gravel/soils and processed or crushed materials and Cemented materials that include all natural crushed materials where a stabilizer of cement or lime has been admixed.
Unbound Materials
As explained earlier unbound materials include natural gravel/soils and processed or crushed materials.
Natural gravel category includes granular materials without any admixture of stabilizers and having the composition of 100% natural gravel or natural gravel with such small proportions of crushed particles that the material properties are almost identical to uncrushed proportion. Pavement materials falling under this category are shown on Table 3.5
Table 3.5: Natural gravel material classes
Material Class Characteristics
G80 - CBR min 80% - the class includes crushed materials where less than 50% by mass of particles retained on the 5 mm sieve has a crushed face
G60 CBR min 60%
G45 CBR min 45%
G25 CBR min 25%
The natural granular pavement materials shall comply with the requirements shown in Appendix 1 of this report.
Crushed materials category includes crushed granular materials without any admixture of stabilizers where the full range of particle sizes from fines up to the maximum nominal size are included. The classes of pavement materials falling into this category are shown in Table 3.6.
33 Table 3.6: Crushed materials - Material classes
Material Class Characteristics
CRR - fresh, crushed rock or large, crushed boulders, with diameter > 0.3 m - requirements are restrictive - compaction requirements are restrictive
-the class includes crushed oversize from gravel sources, crushed all-in CRS sources of boulders and crushed coral rocks of selected qualities -minimum 50% by mass of particles retained on the 5 mm sieve shall have at least one crushed face
The crushed granular materials for pavement layers shall comply with the requirements shown in Appendix 1 of this report.
Cemented Materials
Cemented materials described in the Pavement Materials and Design Manual includes all natural or crushed materials where a stabilizer of cement or lime has been admixed. The classes of cemented materials are shown in Table 3.7.
Table 3.7: Cemented materials - Material classes
Material Class Characteristics
C4 - UCS min. 4 MPa –used as subbase in concrete pavements - made from source materials of quality nominally as CRS with modified requirements
- UCS min. 2 MPa C2 - made from source materials of quality nominally as G45 with modified requirements
C1 - UCS min. 1 MPa - made from source materials of quality nominally as G25 with modified requirements
CM - UCS min. 0.5 MPa, modified material - made from source materials of quality nominally as G7 with modified requirements
Cemented materials for pavement layers shall comply with the requirements shown in Appendix 1 of the report.
34 Bituminous Base Course Materials
Bitumen penetrated macadam and bituminous mixes whether mixed in plant or mixed on the road are also used as base course materials. Penetration macadam is particularly well suited on roads with low traffic speed such as urban roads. Table 3.8 shows classes of bituminous base course materials.
Table 3.8: Bituminous base course material classes
Characteristics Material Class Name Process Mixing Method Dense bitumen macadam Hot DBM Large aggregate mix for Hot Mixing plant LAMBS bases PM Penetration macadam Cold On the road, sprayed FBMIX Foamed bitumen mix Cold Mixing plant or on BEMIX Bituminous emulsion mix Cold the road
In this section only the requirements for the DBM, LAMBS and PM are presented in brief.
Dense bitumen macadam for base course layers shall comply with the requirements shown on Appendix 1 of the report.
LAMBS is a hot mixed bituminous material for base course on heavily trafficked roads and areas of extreme loading. LAMBS shall comply with the requirements given in Appendix 1 of this report. Penetration macadam base course materials shall comply with the requirements given also in Appendix 1 of the report.
Asphalt Concrete (AC) Surfacing
Under this heading a summary of requirements for continuously graded hot premixed asphalt concrete (AC) surfacing is discussed. Mix requirements for severely loaded areas of which, the case road is amongst is described in brief.
The severely loaded areas include the following:
All climbing lanes with gradient 6% or steeper Climbing lanes with gradient 4% or steeper, sustained for 1 km or longer Approaches to major junctions All major town roads Areas where traffic is channelled or slow moving for other reasons
35 An AC mix of high stability shall be used in areas that are severely loaded. The air voids of the mix shall be minimum 3% and remaining at least 3% after traffic loading throughout the design period. The mix type AC 20 (the number refers to maximum grain size) shall be considered for wearing course. 40/50 penetration grade bitumen shall be used in severely loaded areas, with alternatively modified binders which have documented good performance under similar conditions may be used. The required properties for AC are shown in Appendix 1 of the report.
Typical mix proportions for asphalt concrete are presented in Table 3.9. However, the given nominal mix proportions are only indicative as exact proportions are determined after Marshall design procedures.
Table 3.9: Mix proportions for Asphalt Concrete
Nominal Mix Asphalt Concrete Proportions AC 20 AC 14 AC 10 Aggregate [%] 95 94.5 94 Binder [%] 5 5.5 6 Normal loading conditions: 60/70 or 40/50 penetration grade Type of Bitumen Severely loaded areas: 40/50 penetration grade or modified binders
3.3.2.4 Subgrade
In the Pavement Materials and Design Manual the subgrade is classified according to its CBR strength as shown in Table 3.10. The CBRdesign shown in Table 3.10 is the CBR value of a homogenous section, for which the subgrade is classified into S15, S7 and S3 for the purpose of design.
As explained in section 3.3.2, all subgrade shall be brought to a strength of minimum CBR = 15% by constructing one or more improved subgrade layers where necessary. Figure 3.7 shows the design of improved subgrade layers depending on the subgrade CBRdesign determined.
Table 3.10: Subgrade strength classification
CBRdesign [%] Subgrade Wet or Moderate Dry Climatic Zones Density for Class Climatic Zones (both requirements to be met) determination of 4 days soaked Tested at OMC 4 days soaked CBRdesign [% MDD] value value S15 Min 15 Min 15 Min 7 95 BS Heavy S7 7 – 14 7 – 14 3 - 14 93 BS Heavy S3 3 – 6 3 - 6 2 - 6 100 BS Light
36
Figure 3.7: Design of improved subgrade layers
3.3.2.5 Flexible Pavement Design
A commonly used pavement structure in Tanzania is the flexible pavement type whereby a surface treatment or a bituminous mix is placed over a base course made of granular materials. Another flexible pavement type being used consists of base courses made of bituminous mixes or penetration macadam.
Pavement design catalogues are provided for the following pavement types and site conditions:
Granular base course, dry or moderate climatic zones Granular base course, wet climatic zones Cemented base course Bituminous mix base course Penetration macadam base course
The input data for the pavement design catalogues are climate, traffic, subgrade and pavement materials as discussed in previous sections.
Substitutes of subbase materials can be made as listed hereunder. The layer thickness shall not be changed when doing the substitutes.
C1 can be replaced by C2 CM can be replaced by C1 or C2 G45 can be replaced by CM, C1, G60, G80 or CRS
37 G25 can be replaced by CM, C1, G45, G60, G80 or CRS
The design catalogues specify surfacing types for rural – and interurban roads. On all major urban roads where the base course is not of a bituminous type, consideration is given to the use of asphalt concrete surfacing irrespective of traffic loading due to common excess of slow moving and turning traffic.
The pavement design catalogues are presented in Appendix 2 of the report.
3.4 Visual Road Condition Survey
As part of the study, a visual road condition survey was conducted on the case study road section Dar es Salaam – Chalinze – Morogoro. The goal of the condition survey was to identify the typical pavement distresses along the road section and to gather the relevant observations. Some measurements of pavement defects were done during this survey using simple tools. A brief overview of the study area is given in 3.4.1. To discuss the observations of the survey, the road was divided into three (3) sections namely Ubungo – Mlandizi, Mlandizi – Chalinze and Chalinze – Morogoro. The observations for the respective sections are discussed under 3.4.2, 3.4.3 and 3.4.4.
3.4.1 Study Area
The study is based on a road section Dar es Salaam (Ubungo) – Chalinze – Morogoro (183.8 km) which is part of the TANZAM highway (so named because it links Tanzania and Zambia). The TANZAM highway is approximately 921 km long and is designated as Trunk Road T1 in the Tanzania Road Network. This major highway links the port of Dar es Salaam to the land locked countries of Zambia, Malawi, Democratic Republic of Congo, Rwanda, Burundi and Uganda. It also links many of the country’s regions to the business capital city of Dar es Salaam. The study section is the most heavily loaded and most highly trafficked road section in the country. Figure 3.8 shows the location of the case study road section.
38
Figure 3.8: Location of the case study road section
3.4.2 Ubungo – Mlandizi Section
This section (54.4 km) starts at the city of Dar es Salaam and ends at Mlandizi town in the Coast region. The first 5.3 km is a dual carriageway. This section of the road was last reconstructed in the year 2001. The pavement structure is consisting of asphalt concrete on natural sand stabilized with gypsum activated granulated blast furnace slag (GBFS) imported from France. The existing wearing course material was milled, recycled and reused in the reconstruction of the new asphalt concrete binder course. Figure 3.9 shows the details of the pavement structures before and after the reconstruction.
39
Figure 3.9: Pavement structure before (left) and after (right) the reconstruction along Ubungo - Mlandizi section
The typical defects observed on this section are longitudinal cracks, transverse cracks and rutting. The longitudinal and transverse cracks are the dominating distress type observed. The cracks have been observed on the entire road section and are concentrated on both the carriageway and on the shoulders. These cracks occurred only two years after the completion of the reconstruction. The observed cracks have been sealed and very few new cracks have been observed. In some cases it has been observed that new cracks have developed on the already sealed cracks.
Rutting was observed on a few locations and mainly on the up-hills along the wheel paths.
The concentrated longitudinal and transverse cracks observed on the surface layer in this section are probably due to the reflection of the cracks in the pozzolan stabilized base. The use of pozzolanic binders in stabilization as it is and other cementitious stabilizing materials in road pavements normally results in cracks forming in the layer that, depending on the thickness of the overlying asphalt courses, eventually reflect to the surface.
Other longitudinal cracks observed along the wheel tracks are probably caused by the heavy loads imposed on the pavement by the heavy trucks. It is evident that vehicle overloading contributes significant to the destruction of the road pavements thus lowering the pavement service life.
40 The typical rutting observed is probably caused by mechanical deformation of the thin asphalt pavement layer when heavily loaded trucks impose high loads while driving on up-hills with low speed.
Despite the observed distresses, the road in this section is comparatively in good condition.
The typical observed defects along this road section are shown in the figures 3.10 and 3.11.
Figure 3.10: New cracks emerging from sealed cracks (left) and the sealed longitudinal and transverse cracks (right)
Figure 3.11: Rutting and longitudinal crack along wheel tracks (left) and new cracks developing besides the sealed cracks (right)
3.4.3 Mlandizi – Chalinze Section
This section of the road (44.3 km) starts at Mlandizi town in the Coast region and ends up at Chalinze town in the Coast region. The last major rehabilitation work
41 on this section was completed in 1994, and has since then received routine maintenances. The existing pavement structure consists of an asphalt concrete wearing course on the asphalt concrete base course. Figure 3.12 below shows the existing asphalt pavement structure.
Figure 3.12: Pavement structure along Mlandizi - Chalinze section
This section was observed to be comparatively in poor condition. Typical defects observed in this section are rutting, potholes, alligator cracks, edge break and raveling. Excessive rutting and edge break defects have been observed at Chalinze Township where heavily loaded trucks are normally slowing down and parking. Excessive rutting and edge break have also been observed at Ruvu and Vigwaza area. In many locations excessive rutting due to plastic deformation of the surfacing was observed. A maximum rut depth of 150 mm was measured in this section. In some locations potholes resulted from the development of alligator cracks and plastic deformation of the surface was observed. In other locations the failure of patch works was observed.
Longitudinal unevenness of the road surface was observed in some locations which results into uncomfortable drive.
The poor construction quality and vehicle overloading are probable the main causes of the observed defects.
The information extracted from the African Development Bank Group that financed the rehabilitation of this section indicated that during the rehabilitation, construction work showed premature failure that led to consequently suspension of the works for two years to allow for independent investigations. The outcome of the investigations showed that inadequate construction materials were used. With this fact it is believed that poor construction quality and non adherence to
42 the design and specifications contribute significantly to lower the pavement service life.
A comprehensive axle load survey conducted by Central Materials Laboratory (CML) in co-operation with the Road Safety Unit of the then Ministry of Works (MoW) in 1998 and 1999 revealed that the proportion of overloaded axles and the magnitude of the overloading in the country was between 20% to 30% above the legal limit of 10 tons, and a proportion as high as about 3% above 14 tons is alarming. The study revealed that the section Morogoro – Mikumi along the TANZAM highway was the most heavily loaded section whereby out of 3521 axles weighed for all vehicles categories, the percentages overloading (axle load or group of axles exceeding the legal limit set in the regulations, i.e. single axle non steering > 10 tons) in the direction to Mikumi and Morogoro are 34.8% and 25.7% respectively. The study suggested that overloading is apparently caused by traffic moving goods from and to the neighbouring countries, especially the landlocked countries using the port of Dar es Salaam.
The observed condition supports the idea for a need of carrying out the major rehabilitation on this section in a near future.
The typical defects observed in some locations along this section are shown in the figures 3.13 and 3.14 and typical heavy trucks in figure 3.15.
Figure 3.13: Alligator cracks initiating potholes (left) and revelling along the wheel tracks (right)
43
Figure 3.14: Shear failure (left) and edge defects and severe rutting due to plastic deformation of the asphalt (right)
Figure 3.15: Typical heavy trucks at Chalinze
3.4.4 Chalinze - Morogoro Section
This section (85.1 km) was recently reconstructed in 2004 and is comparatively in very good condition. The pavement structure consists of a virgin asphalt concrete wearing course on top of an asphalt concrete binder course with some percentage of recycled asphalt milled from the existing wearing course. The surface layer is built on the crushed aggregates; CRR base course layer. The details of the asphalt pavement structure before and after the reconstruction are depicted in the figure 3.16.
Minor defects especially rutting have been observed along the wheel paths on up-hills locations. The typical rutting observed is probably caused by mechanical deformation of the thin asphalt pavement layer when heavily loaded trucks impose high loads while driving on up-hills with low speed.
44 The only location where excessive rutting has been observed is at Mikese weighbridge station where heavy vehicles are slowing down and queuing on the asphalt pavement. But this location is not in the road carriageway. As no heaves are present aside the wheel tracks the observed excessive rutting may be due to consolidation of the under-laying pavement layers due to improper compaction or mechanical deformation due to insufficient structural strength in the pavement structure. Vehicles overloading and poor construction quality could have been played a significant role on the observed defects.
The observed defects at Mikese weighbridge station are shown in figure 3.17 while a simple measurement of the rut depth and heavy vehicles queuing at the weighbridge station are shown in figure 3.18.
Figure 3.16: Typical pavement structure before (left) and after (right) reconstruction
45
Figure 3.17: Typical rutting at Mikese weighbridge station
Figure 3.18: Simple measurement of rut depth (left) and trucks queuing before axle load measurement at Mikese weighbridge station (right)
3.5 Conclusion
It has been shown in this chapter that Tanzania road sector has been undergoing far reaching reforms since 1998. Two important organizations, Road Fund Board (RFB) and Tanroads have been established since then. This is a positive move towards the better road maintenance management in the country.
Only 5.9% of the entire road network of around 85,525 km is paved. This implies that much of the work is still to be done. As explained in the background information of this report, poor road condition is a major cause of Africa’s low competitiveness. In order for the country to achieve general economic development, it is a must to invest in road transport system.
46 It has been shown in this chapter also that the RFB revenues have comparatively increased over the last two years. The increase has been explained by the decision by the government to increase the fuel levy by 100%. Although this results in increase in revenues for the road maintenance, however, this has an impact on the increase in the transport costs. High transport costs have been identified by several researchers to be a major burden on competitiveness and growth.
The observations from the visual road condition survey revealed that the section Mlandizi – Chalinze is comparatively in poor condition. The typical defects observed in this section are probably caused by poor construction quality and vehicle overloading. The major cracking defects observed along the section Ubungo – Mlandizi are believed to be caused by the reflection cracks from the stabilized base cause layer.
47 4 APPLICATION OF HDM-4 MODEL
4.1 Introduction
The interaction between pavement construction standards, maintenance standards, geometric standards of the road and the effect of the environment and traffic loading has to be evaluated in order to predict the annual trend in road condition. The complexity of these interactions means the use of computer based model is essential. In order to carryout this study and to achieve some of the objectives mentioned in chapter 1, the Highway and Management Model (HDM-4) was used.
In 4.2, the data collected for input in HDM-4 model are discussed. The discussion includes the sources of the data and the presentation of some of the data. Maintenance strategies/ project alternatives applied in this study are discussed in 4.3. In 4.4, the project and strategy analyses run of HDM-4 model are discussed.
HDM-4 model results output analysis and discussion is presented in 4.5. The analysis and discussion includes the effect of the applied maintenance strategies on the deterioration of asphalt pavement, lifecycle economic analysis to obtain the optimum maintenance strategy for each homogenous section and the sensitivity analyses for traffic loading and choice of the discount rate. In 4.6, the concluding remark of this chapter is given.
4.2 Data collection
Much of the input data used in this study were collected from the Tanroads database. Tanroads has developed HDM-4 compatible survey procedures for both paved and unpaved roads. The data include road network data, road works data, vehicle fleet data and configuration data. Tanroads collects and updates these typical data from time to time. Other relevant data and information were collected from the Consulting firms involved in the design and construction of the road sections, relevant reports, design and material manuals as well as the visual condition survey discussed in the previous sections. Some of these data are discussed and presented in the following sections while many other input data are presented in the Appendices of this report.
4.2.1 Vehicle Fleet Data
These data identify and define the vehicle fleet that operates on the road network to be analyzed. It includes basic data as well as economic data. The data for this study were extracted from the Tanroads Vehicle Operating Costs (VOC) Study of 2004. In this study, VOC and Value of Travel Time (VOTT) in the Tanzania context were studied and presented. It has been reported in the Study
48 that based on the general traffic characteristics of the Regional and Trunk roads networks, it is likely that for most analyses Non Motorized Transport (NMT) data would not be required as an input. In this particular Study, the NMT is excluded from the economic analyses. The vehicle fleet is categorized into 7 groups as shown in Table 4.1.
Table 4.1: Tanzania vehicle fleet (Source: Tanroads, 2004)
Class Base Type Category Cars Passenger Cars Car Small Motorized Large Buses Buses Bus Heavy Motorized Light Lorries Trucks Truck Light Motorized Medium Lorries Trucks Truck Medium Motorized Pick-up or Vans Passenger Cars Car Medium Motorized Small Buses Buses Bus Light Motorized Heavy/ Very Heavy Trucks Truck Articulated Motorized Lorries
The Study derived the basic vehicle characteristics and economic unit costs for each of the vehicle categories. The HDM-4 default values are used for Passenger Car Space Equivalent (PCSE), radial tyres and number of retread per tyre.
The Equivalent Standard Axle Loads (ESAL) unit is used by the HDM-4 model to estimate the cumulative damage caused to pavements by an expected traffic stream. Extensive survey data were provided by the Tanroads Central Materials Laboratory (CML) consisting of about 20,000 vehicle-weighing in the period between 1998 and 2002, at 13 sites throughout the country. Figure 4.1 present the ESAL per vehicle calculated using the power 4 obtained from the axle loads survey. A site Morogoro – Mikumi which is along the TANZAM highway can clearly be seen to have the highest ESAL values for the Very Heavy Goods Vehicle (VHGV) category.
The recommended ESAL and operating weights values for HDM-4 analysis are given in the Table 4.2.
The significance of vehicle overloading can be clearly observed from figure 4.1 and table 4.2. Almost all vehicle categories presented in table 4.2 show significant overloading in some road sections. It can be observed that while the proposed mean ESAL to be used for HDM-4 analysis for the VHGV category is 7.10, the ESAL per vehicle data presented in many road sections for this vehicle category is much higher. The ESAL value of 13 was presented in figure 4.1 for this vehicle category along the section Morogoro – Mikumi. It has been explained in section 3.4 of this report that vehicle overloading could play a significant role in the deterioration of the road pavements in the country.
49 ESAL per Vehicle
16 14 12 10 8
ESAL 6 4 2 0
Dar-Kibiti(2001) Himo-Moshi(1998)Himo-Same(1998)Himo-Town(1998) Mbeya-Igawa(1998) Uyole-Ibanda(1998)Uyole-Ibanda(2000) Musoma-Sirari(2001)Tanga-Segera(2000) Arusha-Minjingu(1998)Arusha-Minjingu(2001) Mbeya-Songwe(1998)Morgoro-Mikumi(1999) Chalinze-Segera(2000)Dodoma-Singida(2002) Morogoro-Mikumi(2000) Kobero-Nyakasanza(2001) Site
Bus MGV HGV VHGV
Figure 4.1: ESAL per vehicle data (Source: CML survey data by vehicle 1998 – 2002)
Table 4.2: Recommended ESAL and operating weights for HDM-4 analysis (Source: Tanroads VOC Study, 2004)
Vehicle Type Mean ESAL Operating Weight [kg] Large Bus 3.50 16,700 MGV 1.70 13,800 HGV 3.30 26,000 VHGV 7.10 45,000
The recommended interest rate for VOC economic analysis is estimated at 5 percent per year.
The vehicle service lives, vehicle prices, tyre costs, fuel costs, workshop labour costs, overhead costs and the value of travel time input data were all extracted form the VOC study, 2004. These costs and prices have not been revised nor corrected for inflation. It is obvious that there is a significant difference of the costs and prices for the same items for the effective year of analysis.
50 4.2.2 Road Network Data
The road network data define the physical characteristics of the road sections in the road network to be analyzed. It includes condition data, climate data, pavement characteristics and traffic data. Most of these data were obtained from Tanroads with much of other information obtained from the visual road condition survey conducted, pavement and materials design manual as well as consulting firms involved in the design and construction of the road sections. Based on these data and information, homogenous road sections were derived which are the fundamental units of analysis within HDM-4.
Road Condition Data
Tanroads, being responsible for maintenance and development of trunk and regional roads, collects road condition data frequently for the purpose of planning maintenance activities. Pavement characteristics and condition data used in this study were extracted from the data bank of this road authority. Tanroads has its own data collection manual that describes the methodology for collection of data to be used for input in evaluation tools. The authority has developed HDM-4 compatible survey procedures.
The assessment of the condition of the pavement is conducted using a special condition data collection vehicle. The survey vehicle is fitted with a bump integrator (BI) mounted to the floor of the vehicle. The vehicle is operated by a driver and two recorders. One recorder is responsible for the automatic measurement of roughness and the recording of assessed defects while the second recorder is responsible for recording the visual assessment of the road condition. Figure 4.2 shows a typical data collection vehicle used by Tanroads. The data are recorded using the code stipulated in the data collection manual. Appendix 3 of the report shows the summary of road condition data for the case road.
To provide an overall measure of pavement condition, and to provide a simple means of estimating roughness, the following criteria are adopted for asphalt pavements by Tanroads:
1 – Very Good No visible defects, i.e. the running surface is in the ‘as built’ condition. IRI less than 2.5 m/km
2 – Good Low frequency of defects with low severity. Containing few or no potholes or patches, or areas of visible cracking. IRI less than 3.5 m/km
51 3 – Fair Low frequency of defects with medium severity or medium frequency of defects with low severity. Occasional potholes and surface patches (< 5/km), visible cracking and/or ruts of low severity (<25 mm) affect less than 10% of the length. IRI less than 5 m/km.
4 – Poor Medium frequency of defects with high severity, or high frequency of defects with medium severity. Frequent potholes and/or ruts of high severity affecting up to 20% of the length. IRI less than 7 m/km. (Resealing or structural overlay required)
5 – Very Poor High frequency of defects with high severity. Extensive potholes and patches (>20/km) and/or cracking and severe rutting affecting more than 20% of the length. IRI greater than 7 m/km. (Reconstruction or extensive structural patching are essential)
Figure 4.2: Road condition data collection vehicle
Figure 4.3 shows the condition of the road based on the road roughness data as collected using the bump integrator (BI) mounted on the vehicle. It can be observed from the figure that higher roughness values were recorded between
52 chainage 55 km and 100 km. This is the section Mlandizi – Chalinze which has been observed during the visual road condition survey to be in poor condition. The chainage between 55 km and 70 km is between Mlandizi and Vigwaza area where it has been observed during the visual road condition survey that the area is highly defected. The typical defects observed in this stretch are potholes, severe edge break, alligator cracking, ravelling and patching most of which observed to have been failed.
Much of the typical defects on this area have been explained under section 3.4 of this report when discussing the observations from the visual road condition survey for the section Mlandizi – Chalinze. Together with some figures presented in previous sections explaining the typical defects on this section, Figures 4.4, 4.5 and 4.6 are also included to explain the typical damage along the chainage between 55 km and 70 km. It can be observed from the typical defects that probably the poor construction quality and overloading might have played a significant role on the causes of the defects.
Road Roughness
6
5
4
3
2
1 Roughness [m/km IRI] Roughness
0 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 Chainage [km]
Figure 4.3: Road condition (roughness)
53
Figure 4.4: Alligator cracking at km 56 (left) and shear failure resulting in a failed spot (right)
Figure 4.5: Typical failed patching at km 70 (left) and patch work at km 68 (right)
Figure 4.6: Typical edge break at km 58 (left) and rutting at km 57 (right)
54 The chainage km 90 to km 100 of which higher roughness values have been recorded as shown on figure 4.3 is along Chalinze Township. In this section severe rutting and edge break defects were observed. Figure 4.7 shows some typical defects at Chalinze. It is also believed that poor construction quality and overloading might have played a significant role on the causes of the defects.
The first 20 km which have been observed to have a significant amount of transverse and longitudinal cracks during the visual road condition survey can be seen from figure 4.3 that a maximum of a roughness value of 3.5 m/km IRI was recorded. It has been explained under section 3.4 when discussed the visual road condition survey observations for section Ubungo – Mlandizi that the pavement structure in this section is consisting of a base course with natural sand stabilized with pozzolan. It is therefore believed that the observed cracks are mainly due to reflection of the cracks on the pozzolan stabilized base to the surface.
It can be observed from figure 4.3 that the chainage between km 110 to km 183 is in very good condition. This section is between Chalinze and Morogoro where the pavement reconstruction was completed in the year 2004.
Figure 4.8 shows the typical observed pavement condition along Ubungo – Mlandizi and Chalinze – Morogoro road sections.
Figure 4.7: Typical rutting at km 98 (left) and edge defect and rutting at km 99 (right)
55
Figure 4.8: Typical cracks at km 2.5 (left) and typical surface condition along Chalinze – Morogoro section (right)
Climate
As discussed in sub section 3.3.2.2, Tanzania is divided into three climatic zones which are dry, moderate and wet. As shown in Figure 3.4 in sub section 3.3.2.2, the study road section Dar es Salaam – Chalinze – Morogoro falls under the moderate climatic zone. Table 4.3 shows the climatic parameters for the moderate climatic zone in Tanzania used as input to the HDM-4 model.
Table 4.3: Climatic parameters for Tanzania moderate climatic zone
Moisture Classification Sub-humid Thornthwaite Moisture Index 0 Duration of Dry Season (months) 4.8 Mean Monthly Precipitation (mm) 100 Temperature Classification Tropical Mean Air Temperature (degrees centigrade) 27 Average Temperature Range (degrees centigrade) per year 5 Number of Days with Air Temperature higher than 32˚C (days) 90
Homogenous Sections
Based on the information and data gathered as well as the visual road condition survey, the case study road was divided into five homogenous sections that will be used as the fundamental unit of analysis. The homogenous sections were mainly based on the traffic load, overall road condition, pavement type and other key pavement characteristics. The average of the condition data and traffic load were calculated in each of the homogenous section. Table 4.4 gives a summary of the details of the homogenous sections of the case road. Road condition data, traffic data and other pavement characteristics used to derive the homogeneous sections are attached in Appendices of the report.
56 It can be seen from the table 4.4 that the section Mlandizi – Chalinze is comparatively in poor condition while the section Chalinze – Morogoro which was recently reconstructed is in very good condition.
It can be observed from the same table that some of the observed and discussed typical defects have not been recorded in the table. Edge defect which was one of the major defects along Mlandizi – Chalinze has not been included and thus its omission has an impact on the output results of the analysis. Texture depth and skid resistance data were also not recorded. This could be due to lack of equipments needed for the collection of these typical types of data. The default values were applied for these two typical data. This also has a significant impact on the result output.
With the exception of the road roughness data collected using the bump integrator, the collection of the other condition data is highly based on the subjective assessment and thus highly depending on many factors that include the experience of the personnel.
Table 4.4: Details of Homogeneous sections
Ubungo-Kimara Kimara- Dar/Coast Mlandizi- Chalinze- Section (dual carriageway) Dar/Coast -Mlandizi Chalinze Morogoro Length 5.3 17.9 31.2 44.3 85.1 AADT 25622 13445 5611 4226 3468 AADTyear 2008 2007 2004 2007 2004 Roughness 2.95 2.60 2.39 4.30 2.38 Mean rut depth [mm] 10 10 10 30 5 Edge break [m2/km] - - - - - Pot holes [No/km] 0 0 0 8 0 Wide Cracks [%] 5 10 5 20 2 Ravelled area [%] 5 5 5 10 2 Texture depth [mm] - - - - - Skid resistance - - - - - (SCRIM 50 kph) Drainage good good good good excellent Pavement Type AMSB AMSB AMSB AMAB AMGB Carriage width [m] 12 per carriageway 12.4 7 6.8 7 Overall condition 2 2 2 4 1 Last reconstruction 2001 2001 2001 1994 2004 year
Vehicle Composition per Homogenous Section
Based on the analysis of the Traffic count data on the case road, the vehicle composition on each of the homogenous sections is presented in figure 4.9. It can be observed from the figure 4.9 that the percentage of Cars is higher in section 1 and is decreasing when moving towards section 5. This is due to the
57 fact that sections 1 and 2 are in Dar es Salaam city where there is a high concentration of personal cars. Figure 4.10 shows the Average Annual Daily Traffic (AADT) per homogenous sections.
Vehicle Composition Per Section
35 30 25 20 15 10 5
Percentage [%] Percentage 0 1 2 3 4 5 Section
Cars Pickups/Van Light Lorries Medium Lorries Heavy/Very Heavy Lorries Small Busses Large Buses
Figure 4.9: Vehicle composition per section
AADT per Section
10000 8000 6000
AADT 4000 2000 0 1 2 3 4 5 Section
Cars Pickups/Van Light Lorries Medium Lorries Heavy Lorries Small Busses Large Busses
Figure 4.10: Number of vehicles per section
58 4.2.3 Road Works Data The road works data define the local maintenance and improvement standards for the selected road sections. This includes identifying local distress modes and range of distress values as well as determination of the maintenance unit costs to be applied. Two kinds of work are defined in HDM-4 model for flexible pavements:
Improvement standards comprise pavement reconstruction, lane addition, lane upgrading, partial widening and realignment.
Maintenance standards are applied to meet specific objectives that are related to functional characteristics of the road network system.
Works for crack sealing, fog seal, edge repair, patching, drainage, overlay and flexible pavement reconstruction are included in the maintenance standards. Table 4.5 shows the classification of road works in HDM-4 model.
Table 4.5: Classification of road works in HDM-4 model (Source: Odoki and Kerali, 2000)
Works Works Class Works Type Works Activity/Operation Category Routine pavement Patching, edge-repair, crack sealing, maintenance shoulder repair etc. Routine Drainage Culvert repair, clearing side drains etc. maintenance Routine Vegetation control, line marking, traffic miscellaneous signs etc. Preventive Fog seal, rejuvenation, etc. maintenance Surface dressing, slurry seal, cape seal Preservation Periodic Resurfacing etc. maintenance Thick overlay, mill and replace, inlay Rehabilitation etc. Partial reconstruction, full pavement Reconstruction reconstruction Clearing debris, repairing washout or Emergency subsidence, traffic accident removal, Special etc. Winter Snow removal, salting/ gritting, etc. Widening Partial widening, lane addition Horizontal and vertical geometry Realignment improvements, junction improvement Improvement Shoulder addition, shoulder upgrading, Off-carriageway side drain improvements, etc. Development Upgrading Upgrading by changing the road Construction surface class. New section Dualisation of an existing section, new section/ link.
59 For asphalt pavements, the following are some of the basic types of road works which are normally adopted by Tanroads for maintenance of Trunk and Regional roads:
Crack sealing Patching Edge repair Drainage works 30 mm Overlay 50 mm Overlay 75 mm Overlay Single Bituminous Surface Treatment (SBST) Pavement reconstruction
Some of these road works were used to formulate the project alternatives to be applied in the HDM-4 analyses. The feasibility of application of these road works when considering the observations of the existing pavement condition from the visual condition survey is discussed in section 4.3.
A survey on road work costs data was done. The financial and economic costs of the crack sealing, patching, edge repair, overlays, SBST and pavement reconstruction works are shown in Table 4.6. The financial costs is the sum of the market price of materials, labour, equipment and overheads incurred while the economic costs are the actual costs of the work exclusive of the government taxes.
Table 4.6: Road work costs (Source: Surveyed data, 2009)
Financial Cost, Economic Cost, Work Class Work Type Work Activity [USD] [USD] Crack Sealing 3.08/m 2.00/m Routine Routine Surface Patching 32.69/m2 26.15/m2 Maintenance Maintenance Edge Repair 32.69/m2 26.15/m2 Resurfacing Surface Treatment SBST -10 mm 6.15/m2 4.92/m2 Periodic Resurfacing Asphalt 30 mm Overlay 23.08/m2 18.46/m2 Maintenance Concrete 50 mm Overlay 30.77/m2 24.62/m2 Rehabilitation 75 mm Overlay 38.46/m2 30.77/m2 Pavement Full reconstruction reconstruction 500,000/km 400,000/km
60 4.3 Maintenance Strategies/Project Alternatives:
Optimum works standards can be investigated using the strategic analysis application within HDM-4 by specifying a series of maintenance or improvement alternatives and applying them to representative road sections.
In order to evaluate the impact of different maintenance strategies on the condition of the road and to meet the objectives of this study, four project alternatives have been formulated and used in this case with different maintenance standards and intervention levels. The formulation of these project alternatives is based on the available data and information gathered. Below is the list of the four project alternatives followed by the discussion of its feasibility on the case road:
Project Alternative 1: Routine maintenance and reconstruction at 10 m/km IRI - (Base case)
This alternative involves the routine maintenance to be done at any particular time based on the condition of the road. It also includes a condition-responsive reconstruction, triggered only when the road is in extreme poor condition, e.g. roughness ≥ 10 m/km IRI.
Project Alternative 2: Routine maintenance only
This alternative is included in the evaluation to show the consequences of doing just routine maintenance over the evaluation period. This normally happens when the available budget for maintenance is very limited.
The routine maintenance to be carried out each year based on the condition of the road includes the crack sealing if the area of wide structural cracking is greater than 5% of the carriageway and the number of thermal cracks is greater than 20 per km, patching if the severely damaged area is greater than 10% and edge repair when the damaged area is greater than 5000 m2/km.
Project Alternative 3: Single Bituminous Surface Treatment at 10% Area of All Cracking and routine maintenance
This alternative involves the application of a 10 mm thick Single Bituminous Surface Treatment (SBST) when the percentage of Area of All Cracking (ACA) reaches 10%. The routine maintenance to be carried out each year based on the condition of the road is also included.
Project Alternative 4: Asphalt Concrete Overlay at 7 m/km IRI and routine maintenance
61 This alternative involves the application of 50 mm asphalt concrete overlay when the road roughness reaches 7 m/km IRI. It also includes the routine maintenance to be carried out each year based on the condition of the road.
Many factors have to be considered when formulating the maintenance/project alternatives for a particular road section or on the network as a whole. Some of the factors to be considered include the agency policy, practical construction aspects, minimum user acceptable levels of riding quality and the availability of skills, materials and other resources.
In terms of minimum user acceptable levels of riding quality which is dependent on the road class and traffic category, the following is discussed.
Bearing in mind the importance of the case study road and the observed road pavement condition, the intervention levels in some of the project alternatives can be seen clearly as unrealistic to be applied in this road.
Section 4.2.2 of this report discusses criteria provided by Tanroads data collection manual to rank the overall pavement condition. In the criteria, the pavement is rated to be in poor condition when the roughness value reaches 7 m/km IRI on which resealing or structural overlay are mentioned as the maintenance options. The pavement is rated to be in a very poor condition when the roughness value is greater than 7 m/km on which the pavement reconstruction or extensive structural patching is mentioned to be essential maintenance options. The above formulated intervention levels are in somehow in line with these criteria.
The Pavement and Materials Design Manual rate the pavement condition for the higher traffic classes’ roads in terms of roughness values into the following:
Sound (adequate condition): - roughness < 3 m/km IRI Warning (uncertainty exists about the adequacy of the condition) – roughness 3-6 m/km IRI Severe (inadequate condition) – roughness > 6 m/km
The above condition rating shows that the intervention levels used to formulate the project alternatives for both overlay and pavement reconstruction are beyond the minimum roughness value of 6 m/km IRI (severe condition) and thus not acceptable levels of pavement condition for this important road category. The minimum acceptable level of pavement condition with roughness value of 6 m/km IRI is seen feasible for this road category. It has to be noted that vehicle operating costs (VOC) are dependent of the pavement condition and thus higher roughness values result in higher VOC that implies higher transport costs. Higher
62 transport costs were mentioned in the introduction of this study to be one of the reasons Africa’s low competitiveness and growth.
In terms of practical construction aspects the following is discussed on the application of the asphalt overlay and SBST as maintenance options.
Asphalt overlays are used for the purpose of adding sufficient structural strength for the pavement to carry future traffic in design period and to store the riding quality of the pavement. However, some conditions limit the application of the overlay on the existing pavement. For example, overlay is not proposed to be used on severely cracked pavement where there is a risk of early crack reflection through the new layers. It is also proposed not to apply overlay on pavements with deformation (i.e. shoving) in bituminous layers unless repair or removal of the deformed material is carried out. Overlay is not proposed also to be applied where there is uncertainty about the performance of the overlay due to defects in the existing base course or in patches in the existing pavement.
Based on the road condition survey observations, overlay is probably not the option to be applied on the section Ubungo – Mlandizi due to the risk of early crack reflection through the new surface layers and also on the section Mlandizi – Chalinze due to the observed deformation defects on bituminous layers.
On the other hand, SBST is suitable for maintenance resealing. SBST is normally applied on asphalt concrete that are high in air voids. It is applied for the purpose of minimizing seepage of harmful fluids into the pavement layers and to improve the skid resistance on AC. However, SBST is insufficient to be applied on severely deteriorated pavements. In this case Double Bituminous Surface Treatment (SBST) may be used instead.
The visual road condition survey observations therefore imply that SBST is not recommended as maintenance option for the section Mlandizi – Chalinze due to its poor condition.
The project alternatives formulated do not include the drainage works and miscellaneous maintenance (vegetation control, sign repair, line marking) activities for routine maintenance because the data on these items were not obtained. While drainage and vegetation works have an impact on the pavement structure, the sign repair and line marking costs are normally included in the maintenance budgets. The information and data on the milling and replacing asphalt pavement could not be obtained also. Based on the observation of the road condition in some sections, this typical operation could be included. However, the alternative of total pavement reconstruction is included instead.
63 Table 4.7 shows the summary of road works standard for each of the formulated project alternative.
Table 4.7: Summary of road works standards for each project alternative
Project Road Work Standard Effective Maintenance Works Alternative Year
Alternative 1: Routine and 2010 Reconstruct at 10 m/km IRI, Crack RM + Reconstruction sealing, Patching and Edge repair Reconstruction Alternative 2: RM only Routine maintenance 2010 Crack sealing, Patching and Edge only repair Alternative 3: Routine and SBST 2010 SBST at 10% ACA , Crack sealing, RM + SBST Patching and Edge repair
Alternative 4: Routine and 50 mm AC 2010 50 mm AC overlay at IRI >= 7 RM + 50 mm overlay m/km, Crack sealing, Patching and overlay Edge repair
4.4 Running life cycle project and strategy analyses
Based on the objectives of the study described in section 1.3 of the report, the HDM-4 project and strategy analysis were run to analyze the effect of the different formulated maintenance strategies on the deterioration of the road pavement and to examine and identify the life cycle optimal maintenance strategy on the study road. Two methods for analyzing investment options are provided in HDM-4 that includes analyses by section and by project. While analyses by project consider the whole road network as a single package, the analyses by section treat each road section different from the others. In this particular study, the project analysis by section was adopted so that each representative road section can be discussed separately.
It is possible to assign different maintenance strategies in different road sections based on the observed pavement condition and construction practicability. In this study however, all four maintenance strategies were applied into all five homogeneous road sections for comparison.
For both strategy and project analyses the discount rate (interest rate used in discounting future cash flows) of 12% and the analysis period of 20 years were adopted based on the Tanzanian economic evaluation of road investments. In both cases the effective year is 2010. Accident costs, energy balance emissions and acceleration effects (effect of vehicle acceleration due traffic flow characteristics and vehicle interactions) are not included in the analysis.
64 Estimation of a suitable discount rate is often one of the most difficult and uncertain part of discounted cash flow valuation. Basically no agreement exists on what the correct discount rate is, in many project evaluations. Controversy over discounting lies at the heart of the debate on Cost Benefit Analysis (CBA) is that the choice of discount rate can often determine whether net benefits are found to be positive or negative. In this study the sensitivity analysis was done to examine the effect of the choice of the discount rate on the project alternatives. This was done by varying the discount rate to 5% and 18% from the proposed 12%.
To examine the effect of the overloading of the heavy goods vehicles, the sensitivity analysis was carried out by varying the magnitude of the mean ESAL to 10 and 5 from the proposed mean ESAL of 7.10. This is based on the observations on the collected ESAL data presented in the earlier sections on which an ESAL value of 13 was recorded in one of the sections of the TANZAM highway.
Both project and strategy output results as well as the sensitivity analyses are discussed in the following sections.
4.5 Results output analysis and discussion
This section presents the result output of the HDM-4 project and strategy analyses done for the case road based on the input data discussed in the previous sections. The discussion of the results will consider only the outputs which are relevant to the objectives of the study which have been described in section 1.3 of the report. The HDM-4 deterioration/works effects and cost streams generated reports are analyzed and discussed in this section. The effect of the application of the formulated project alternatives on the pavement condition is discussed. In the end the optimal maintenance strategy for each representative section is determined based on the highest economic return (Net Present Value). The total transport costs and agency costs (funding requirements) for each representative section for the determined optimum maintenance strategy are obtained. The effect of overloading of heavy goods vehicles and effect of choice of discounting rate for economic analysis are also discussed in this section based on HDM-4 sensitivity analyses.
4.5.1 Effect of maintenance strategy on deterioration of pavement
It is known that pavements deteriorate due to the effects of the interaction between pavement construction standards, maintenance standards, geometric standards of the road and the effects of the environment and traffic loading. The complexities of these interactions have been analyzed by means of the HDM-4 model for each of the representative sections of the case road.
65 Figure 4.11 shows the HDM-4 automatically generated graphs and it shows how each of the formulated maintenance strategies effect the deterioration of the road pavement in terms of the progression of the average roughness on each of the representative sections.
The curves in figure 4.11 have been obtained from the modelling equations within the HDM-4 Model. Although the scope of the study does not require the in depth discussion on the modelling of different pavement distresses by the HDM-4 Model, the following equations were extracted from the HDM-4 documentation series and are presented here to show how figure 4.11 was derived. The incremental change in roughness is given in equation 4.1:
5 RIs a0 expmKgm AGE31 SNPK b YE 4 ………………………………………………..4.1
Where
SNPKb MAX SNPa dSNPK ,1.5 …………………………………………………………….4.2 and
MINa1 ACX a HSNEW dSNPK K snpka0 ……………...... 4.3 MAX MINACX a PACX ,a2 ,0HSOLD
The used parameters are defined below:
ΔRIs – incremental change in roughness due to structural deterioration during the analysis year (IRI m/km) dSNPK – reduction in adjusted structural number of pavement due to cracking. SNPKb – adjusted structural number of pavement due to cracking at the end of the analysis year. SNPa - adjusted structural number of pavement due to cracking at the end of the analysis year. ACXa - area of indexed cracking at the start of the analysis year (% of total carriageway area) PACX - area of previous indexed cracking in the old surfacing (% of total carriageway area HSNEW - thickness of the most recent surfacing [mm] HSOLD - total thickness of previous underlying layers AGE3 - pavement age since last overlay (rehabilitation), reconstruction or new construction (years) YE4 - annual number of equivalent standard axles (millions/lane) m - environmental coefficient (dependent on climate zone)
66 Kgm - calibration factor for environmental coefficient Ksnpk - calibration factor for SNPK
It can be observed from Figure 4.11 that in all representative sections the maintenance strategies Routine Maintenance Only (RM) and Routine Maintenance and Single Bituminous Surface Treatment (RM+SBST) do not contribute to the reduction in the road roughness. Instead the SBST has the effect of reducing the rate of increase of road roughness. It has been explained in the previous sections that SBST is mainly applied to improve the skid resistance and minimizing seepage of harmful fluids into the pavement layers and does not contribute on the pavement structural strengthening. In all representative sections it can be observed from Figure 4.11 that the maintenance strategy RM Only results in a higher rate of deterioration of the road pavement, and depending on the section characteristics, the road deterioration increases to the maximum specified roughness of 16 m/km for section Ubungo – Kimara, Kimara – Dar/Coast and Mlandizi – Chalinze.
As observed during the visual road condition survey and on the data collected, the section Mlandizi – Chalinze, which was rated as in poor condition, is observed in Figure 4.11 to require pavement strengthening or total reconstruction in near future after the effective year. The rate of pavement deterioration is very high and results into extremely higher roughness values in early years after the effective year when applying the maintenance strategies RM Only and RM+SBST on this section.
The maintenance strategy RM + Reconstruction at 10 m/km IRI is not triggered in section Chalinze – Morogoro. This is due to the fact that the road section was observed to be in very good condition and has lower traffic levels as compared to other representative road sections. It can be observed from the figure 4.11 that the intervention level at roughness of 10 m/km IRI is too high to be reached within the analysis period.
As discussed in section 4.2.3, the intervention levels used for both AC50 and pavement reconstruction are extremely beyond the feasible acceptable range of roughness for this important road and thus they have been used for comparison purposes only.
67
Figure 4.11a: Pavement deterioration pattern based on maintenance strategy for section Ubungo - Kimara
Figure 4.11b: Pavement deterioration pattern based on maintenance strategy for section Kimara – Dar/Coast
68
16 16 Base Alternative RMBase Only Alternative 14 RM Only 14 RM+AC50 RM+SBSTRM+AC50 12 RM+SBST 12 10 10 8 8 6 6 4
Average Roughness (m/km) AverageRoughness 4 Average Roughness (m/km) AverageRoughness 2
2
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
Year Year
Figure 4.11c: Pavement deterioration pattern based on maintenance strategy for section Dar/Coast - Mlandizi
Figure 4.11e: Pavement deterioration pattern based on maintenance strategy for section Mlandizi - Chalinze
69
16 Base Alternative 16 RMBase Only Alternative 14 RM Only 14 RM+AC50 RM+SBSTRM+AC50 12 RM+SBST 12 10 10 8 8 6 6 4
Average Roughness (m/km) AverageRoughness 4 Average Roughness (m/km) AverageRoughness 2
2
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 Year Year
Figure 4.11e: Pavement deterioration pattern based on maintenance strategy for section Chalinze - Morogor
Table 4.8 shows the summary of the road works report whereas triggering of major maintenance works is shown. It has to be noted that in each alternative a routine maintenance is included. The table 4.8 elaborate the timing of the maintenance works as shown on figure 4.10 thus; it can be read in conjunction with the figure 4.11.
It is observed in table 4.8 that the SBST works are triggered at fixed intervals of four years in all road sections regardless of the pavement condition. As explained in the previous sections SBST is mainly applied to improve the skid resistance and minimizing seepage of harmful fluids into the pavement layers. It does not contribute to the strengthening of the pavement structure and does not reduce the pavement roughness. Due to its poor condition the section Mlandizi – Chalinze requires the application of overlay two times within a 20 years analysis period and at an interval of 9 years.
70
10 Base Alternative 9 RM only RM+AC50 8 RM+SBST
7
6
5
4
Average Roughness (m/km) AverageRoughness 3
2
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
Year Table 4.8: Road works summary report
Year of Maintenance Road Section Alternative maintenance works works RM+Reconstruction 2022 Reconstruction RM Only 2010 - 2029 RM each year 2010 SBST 2014 SBST Ubungo - Kimara RM + SBST 2018 SBST 2022 SBST 2026 SBST RM + AC50 2018 AC50 RM+Reconstruction 2022 Reconstruction RM Only 2010 - 2029 RM each year 2010 SBST 2014 SBST 2018 SBST Kimara – Dar/Coast RM + SBST 2022 SBST 2026 SBST 2029 SBST RM + AC50 2018 AC50 RM + Reconstruction 2024 Reconstruction RM Only 2010 – 2029 RM each year 2010 SBST 2014 SBST 2018 SBST Dar/Coast – Mlandizi RM + SBST 2022 SBST 2026 SBST 2029 SBST RM + AC50 2019 AC50 RM + Reconstruction 2016 Reconstruction RM Only 2010 – 2029 RM each year 2010 SBST 2014 SBST Mlandizi – Chalinze RM + SBST 2018 SBST 2022 SBST 2026 SBST 2029 SBST 2013 AC50 AC50 2022 AC50 RM + Reconstruction - No reconstruction RM Only 2010 - 2029 RM each year 2010 SBST 2015 SBST Mlandizi - Chalinze RM + SBST 2019 SBST 2023 SBST 2027 SBST RM + AC50 2024 AC50
71 4.5.2 Life cycle economic analysis
HDM-4 simulates the total life cycle conditions and the costs for the analysis period under specified scenarios of circumstances. The primary cost set for the life cycle analysis includes the cost of capital investment, maintenance, vehicle operation and travel time costs. The costs of accidents and environmental pollution may be included in the analysis. In this particular study the costs of accidents and environmental pollution are not included in the analysis.
The physical quantities involved in construction, maintenance and vehicle operation are estimated, and the specified prices and unit costs are applied to determine the financial and economic costs. Relative benefits are calculated for different alternatives, followed by present values computations.
Depending on the maintenance policy of the country, the appropriate maintenance standards can be chosen based on either the minimum user acceptable levels of riding quality on the road section, clear definition of maintainable and un-maintainable conditions as well as economic criteria such as maximum benefit to the society (Maximum NPV). In this study the optimum maintenance standards on each of the representative sections were obtained based on the economic criterion that is the maximum NPV.
Figures 4.12 and 4.13 show the undiscounted road agency cost per km length and undiscounted vehicle operating cost (VOC) per km length for different maintenance alternatives and for each section for 20 years analysis period respectively. Comparison from the two figures shows that the VOC are much higher than the road agency costs. Both figures 4.12 and 4.13 show that the costs are function of the pavement section characteristics. For example the section Ubungo – Kimara is observed to have higher agency cost per km as compared to the other section. This can be explained by the fact that this section is within the city of Dar es Salaam in which it has comparative high traffic volume and it has wider pavement width. High frequency of routine maintenance and large quantities of road works is expected that contribute to the high agency costs.
The influence of traffic volume is also observed in figure 4.13. The sections Ubungo – Kimara and Kimara – Dar/Coast which have higher traffic volumes are observed to have higher VOC. From the two figures it can therefore be seen that together with other pavement characteristics the traffic volume has much influence into the agency and VOC costs.
72 The effect of applying RM only alternative can be observed from figure 4.13. It can be observed from figure 4.13 that the application of RM only alternative results into comparatively higher VOC. As observed from figure 4.11 the RM only alternative results into higher values of pavement roughness values. Pavement roughness also contributes significantly into VOC. This is in line with what has been explained by Kruger, (2003) that if Africa doesn’t spend the required maintenance costs for good roads, it is then paying for them through higher vehicles operating costs, travel time, accident costs as well as through opportunity costs for lower economic growth.
Figure 4.14 shows the discounted net benefit (NPV) for each road section and maintenance strategy as compared to the base alternative. It can be observed in the figure that maintenance strategy RM only results into zero or negative discounted net benefit. It can be observed also in the same figure that the maintenance strategies RM only and RM+SBST result in very high negative discounted net benefit for the section Mlandizi – Chalinze which was observed to be in poor condition during the visual road condition survey.
The detail of the economic analysis summary report is presented in Appendix 4 of this report.
Road Agency Cost per km for 20 years
3.0 2.5 2.0 1.5 1.0
[USD millions] 0.5 Cost per km length Cost per length km 0.0 Ubungo-Kimara Kimara-Dar/Coast Dar/Coast-Mlandizi Mlandizi-Chalinze Chalinze-Morogoro (10.8 km) (17.9 km) (31.2 km) (43.3 km) (85.1km) Road Section
RM+Reconstruction RM Only RM+SBST RM+AC50
Figure 4.12: Summary of undiscounted road agency economic cost per km length for 20 years analysis period.
73 Vehicle Operating Cost per km for 20 years
60 50 40 30 20 10 [USD millions] [USD 0 Cost per km length length km per Cost Ubungo-Kimara Kimara- Dar/Coast- Mlandizi-Chalinze Chalinze- (10.8 km) Dar/Coast (17.9 Mlandizi (31.2 (43.3 km) Morogoro km) km) (85.1km) Road Section
RM+Reconstruction RM Only RM+SBST RM+AC50
Figure 4.13: Summary of undiscounted VOC per km length for 20 years analysis period.
Discounted Net Benefit Stream
30 10 -10 -30
Ubungo - Kimara Kimara - Dar/Coast - Mlandizi - Chalinze - Net BenefitNet -50 Dar/Coast Mlandizi Chalinze Morogoro -70 -90 Road Section
RM Only RM+SBST RM+AC50
Figure 4.14: Discounted Net Benefit for each section and maintenance strategy
Since in this study the economic criteria adopted for the economic analysis is the maximum net benefit to the society (NPV), the optimum maintenance strategies can then be obtained from Figure 4.14. Table 4.9 presents a summary of the optimum maintenance strategy for each of the representative road sections. A summary of maintenance road works by section in corresponding years, quantities and costs is shown in Appendix 4 of the report.
74 Table 4.9: Optimum maintenance strategies for representative road sections
Optimum Maintenance Road Section Strategy NPV Ubungo – Kimara RM+SBST 8.19 Kimara – Dar/Coast RM+AC50 7.14 Dar/Coast - Mlandizi RM+AC50 12.95 Mlandizi - Chalinze RM+AC50 7.64 Chalinze - Morogoro RM+SBST 38.07
Figure 4.15 shows the total transport costs (sum of road agency costs and road user costs) for the optimum maintenance strategies for each road section. Road agency costs include the capital costs of investment and recurrent costs while the road user costs include the vehicle operating costs and travel time costs. It can be observed from the figure that there is no decrease in total transport costs for the sections Chalinze – Morogoro and Ubungo – Kimara. This can be explained by the fact that the optimal maintenance strategy in these sections is the RM+SBST which as explained in the previous paragraphs has no contribution to the strengthening of the pavement structure rather than only affecting the rate of increase in road roughness. The road roughness is one of the key factors contributing to the vehicle operating costs. The higher total transport costs observed for the section Chalinze – Morogoro is contributed by the section characteristics as compared to the other sections. This is the longest section with the hilly and bendy geometry. These characteristics also contribute much to the increase in vehicle operating costs. The impact of applying an asphalt overlay on the road pavement is observed on the sections Kimara-Dar/Coast, Dar/Coast- Mlandizi and Mlandizi-Chalinze. This result into a decrease in the total transport costs immediately after application of the asphalt overlay which decreases the road roughness, thus decrease in vehicle operating costs.
Figure 4.16 shows the annual total road agency costs (sum of the capital costs and recurrent costs of maintenance) for each road section at their optimum maintenance strategies. It can be observed from the figure 4.16 that the section Mlandizi – Chalinze requires comparative large amounts of maintenance funding at the time of intervention. This is probably due to the fact that the section is comparatively in poor condition. It can also be observed from the figure 4.16 that the maintenance cost is significantly reduced as soon as the capital maintenance investment is done. This is due to the fact that the pavement requires very minimal maintenance interventions after major arrest of the defects.
75 Yearly Transport Costs
150.000 130.000 110.000 90.000 70.000 50.000
TotalCosts 30.000 [USDmillions] 10.000 -10.000 2009 2011 2013 2015 2017 2019 2021 2023 2025 2027 2029 2031 Year
Chalinze-Morogoro Mlandizi-Chalinze Dar/Coast-Mlandizi Kimara-Dar/Coast Ubungo-Kimara
Figure 4.15: Annual transport costs for the optimum maintenance strategies
Annual Total Road Agency Costs
8.00 7.00 6.00 5.00 4.00 3.00
2.00 AgencyCosts [USDmillions] 1.00 0.00 2009 2011 2013 2015 2017 2019 2021 2023 2025 2027 2029 2031 Year
Chalinze-Morogoro Dar/Coast-Mlandizi Kimara-Dar/Coast Mlandizi-Chalinze Ubungo-Kimara
Figure 4.16: Annual road agency costs for different road sections at respective optimum maintenance strategies
76 4.5.3 Sensitivity analysis for traffic loadings
Overloaded vehicle axles contribute significantly to lower the pavement service life. In this study it has been observed that while a mean ESAL value of 7.10 was recommended for HDM-4 analysis for the heavy/very heavy goods vehicle category, the axle load data presented showed that in some of the sections of the Tanzania road network the value of the ESAL for this vehicle category is much higher. An ESAL value of 13 was observed in one of the road sections. To analyze the consequences of this, a sensitivity analysis using the HDM-4 model is done. Bennett and Paterson (1998) argue that, although there are many approaches that can be used to do the sensitivity analysis, the traditional ‘centris paribus’ method which involves changing a single factor while holding all other factors constant is the most commonly used.
In this study a value of mean ESAL for the heavy/very heavy vehicle category to be applied in the HDM-4 analysis is changed first to 10 and then to 5 while holding all other input factors the same. The effect on the pavement deterioration is then examined by comparing the rate of deterioration in one of the section for the three different mean ESAL input. The section Mlandizi – Chalinze is chosen for the discussion.
Figure 4.17 shows the pavement deterioration pattern for the section Mlandizi – Chalinze for mean ESAL input of 10 for heavy/very heavy vehicle category while Figure 4.18 shows the same for mean ESAL input of 5. The comparison is made to Figure 4.11d presented in section 4.5.1.
It can be observed from the three figures that the rate of pavement deterioration is higher when applying a mean ESAL input of 10 (Figure 4.17) and lower when applying the mean ESAL input of 5 (Figure 4.18). One can observe that while the reconstruction (base alternative) and overlay (RM+AC50 alternative) are triggered in the years 2015 and 2012 respectively when applying mean ESAL of 10 for the heavy/very heavy vehicle category, the same is triggered in 2016 and 2013 respectively, one year later, when applying the proposed mean ESAL input of 7.10. The observations are in line with much of the research findings on the effect of overloaded axles on the road pavements. However, the effect of overloaded axles on the pavement structure is believed to be much higher than what is observed in the HDM-4 results for this case.
77
Figure 4.17: Pavement deterioration pattern based on maintenance strategy for section Mlandizi – Chalinze (mean ESAL -10)
Figure 4.18: Pavement deterioration pattern based on maintenance strategy for section Mlandizi – Chalinze (mean ESAL– 5)
4.5.4 Sensitivity analysis for the choice of discount rate
A discount rate is the percentage by which the value of a cash flow in a discounted cash flow valuation is reduced for each time period by which it is recovered from the present. The discount rate reflects the rate at which money can increase in productive investments (i.e. productivity of capital). It has been
78
16 16 Base Alternative RMBase Only Alternative 14 RM Only 14 RM+AC50 RM+SBSTRM+AC50 12 RM+SBST 12 10 10 8 8 6 6 4
Average Roughness (m/km) AverageRoughness 4 Average Roughness (m/km) AverageRoughness 2
2
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
Year Year reported in many literatures that the estimation of a suitable discount rate is often the most difficult and uncertain part of the discounted cash flow valuation. The controversy over the choice of the discount rate lies at the heart of the debate on Cost Benefit Analysis.
In this study the life cycle economic analysis by the HDM-4 model was done using the discount rate of 12%. This rate has been commonly proposed and used in many of the Tanzanian transport sector economic studies. Again, in this case the traditional ‘centris paribus’ sensitivity analysis method which involves changing a single factor while holding all other factors constant has been applied. The life cycle economic analysis using HDM-4 model was done using the discount rate of 5% and 18% and the NPV results were compared with the NPV obtained when using the 12% discount rate.
Table 4.10 shows the comparison between the NPV obtained for the three discount rates at each road section and maintenance alternative. It can be observed from table 4.10 that although the different discount rates applied resulted into different NPV as expected, in three road sections with exception of the section Ubungo – Kimara and Dar/Coast - Mlandizi, the optimum maintenance strategies remain the same regardless of the choice of the discount rate. The optimum maintenance strategy for the section Ubungo – Kimara changes to RM + AC50 from RM + SBST when the discount rate of 5% is applied while for the section Dar/Coast – Mlandizi the optimum maintenance strategy changes to RM + SBST from RM + AC50 when the discount rate of 18% is applied.
It can be observed from the same table that the optimum project alternatives are more favoured when choosing the low discount rate.
79 Table 4.10: NPV results for different discount rates
NPV NPV at NPV at Section Alternative at 5% 12% 18% Base 0 0 0 RM Only -48.1 -16.8 -7.2 RM + SBST 5.3 8.2 7.6 Ubungo - Kimara RM + AC50 8.5 5.1 3.2
Base 0 0 0 RM Only -197.4 -84.5 -43.7 RM + SBST -147.6 -55 -23.8 Mlandizi - Chalinze RM + AC50 7.8 7.6 6.8
Base 0 0 0 RM Only -86.2 -29.9 -12.8 RM + SBST -19.8 -0.1 4.4 Kimara - Dar/Coast RM + AC50 11.1 7.1 4.6
Base 0 0 0 RM Only -71.1 -23.1 -9.3 RM + SBST 2 9.6 9.4 Dar/Coast - Mlandizi RM + AC50 25.4 13 7.3
Base 0 0 0 RM Only 0 0 0 RM + SBST 80.2 38.1 22.6 Chalinze - Morogoro RM + AC50 54.4 17.6 7.1
4.6 Conclusions
It can be concluded in this chapter that the HDM-4 analyses have shown expected output results based on the input data used. Road condition data collected have shown that the section Mlandizi – Chalinze is in comparatively poor condition while the section Chalinze – Morogoro is in good condition. The same was observed during the visual road condition survey. It has however been observed in the data collected (year 2008) that some of the typical pavement defects i.e. edge defect observed during the visual condition survey were not included. The texture depth and skid resistance data were not included also. Based on the literature survey the omission of these data has an impact on the results of the HDM-4 model analyses.
The data on the vehicle fleet used are all based on the study conducted by the Tanroads in the year 2004. It is believed that the costs and prices have significantly changed. No updated data were found and there were no
80 corrections on the effect of inflation which is one of the major concerns in developing countries.
The intervention levels for overlay and pavement reconstruction used to formulate the project alternatives are beyond the minimum roughness value of 6 m/km IRI which is considered in the Pavement and Materials Design Manual as severe condition. The intervention levels were based on the criteria provided by the Tanroads data collection manual. However, it is believed that it is not feasible to let this particular important road to deteriorate to reach severe condition. The application of these intervention levels is reflected into the HDM-4 economic analyses on which the VOC were observed to be extremely high.
The effects of the maintenance strategies on each of the representative sections have shown that the maintenance strategy routine maintenance only (RM Only) is not the preferred maintenance option due to the high rate of pavement deterioration and the higher total transport costs. The section Mlandizi – Chalinze was observed to require comparative huge capital investment when intervention is done. This can be explained by the fact that the section is in comparative poor condition.
The HDM-4 sensitivity analysis on the effect of heavy vehicle overloading has shown that overloading does affect the road pavement structure by increasing the rate of pavement deterioration. A shift of only one year for the maintenance programme was observed. It is however believed that the overloading does have a higher impact than what has been shown by the HDM-4 results.
The HDM-4 sensitivity analysis has shown also that different discount rates result into different NPV, as expected, and thus do affect the decision making. It has been however discussed in literature that the choice of discount rate is in the centre of debate when doing Cost Benefit Analyses.
The study has used much of the default model parameters due to the fact that the road agency has yet to calibrate the model to adapt the local conditions. It has been observed during the visual road condition survey that much of the defects observed are probably caused by the poor quality construction practice and probably the overloading plays a significant role. Although the HDM-4 model considers the effect of quality of construction of the surface layer and the base layer through the input parameters Construction Defects Indicators, CDS and CDB respectively, it is believed that the estimation of these input parameters is not enough to address the extent of the problem.
It can be concluded therefore that the observed HDM-4 output results are only indicative and might not represent the real situation.
81 5 MECHANISTIC PAVEMENT ANALYSIS
5.1 Introduction
It has been discussed in chapters 3 and 4 that the two sections of the case road, Mlandizi – Chalinze and Chalinze – Morogoro are in poor and good condition respectively. It has been discussed also that heavy vehicle overloading has probably played a significant role in the cause of many of the observed typical pavement defects. HDM-4 sensitivity analysis discussed in chapter 4 has not shown expected results on the effect of overloading on the pavement deterioration. Based on fundamentals of engineering and many research findings it is believed however that overloading has a significant impact on the pavement deterioration.
In this chapter mechanistic pavement analysis of the two sections is done to estimate the pavement life and to determine preferable maintenance option for the section Mlandizi – Chalinze taking into consideration the typical axle loads. Three axle load cases have been considered in the analysis. The axle loads considered include 80 kN (dual tyre, i.e. 2*20 kN wheel load), 200 kN (dual tyre, i.e. 2*50 kN wheel load) and 150 kN (wide base tyre, i.e. 1*75 kN wheel load) each at loading time of 0.01 seconds (vehicle speed 100 km/h), 0.02 seconds (vehicle speed 60 km/h) and 0.04 seconds (30 km/h). The tyre pressures of 700 kPa, 800 kPa and 1000 kPa were assumed for the 20 kN (80 kN axle load), 50 kN (200 kN axle load) and 75 kN (150 kN axle load) tire loads respectively.
The three axle loads cases considered are based on the typical axle loads expected to be found in Tanzania road network. 80 kN axle load is considered as the standard axle load. It has to be noted that over the years now there has been an increase of importation of used trucks to Tanzania and neighbouring countries from Europe and other parts of the world. Figure 5.1 shows an example of the typical truck under consideration. In most cases the driven axle of the typical truck mounted with dual tires is extremely overloaded. In this particular case it is assumed that the driven axle has an axle load of 200 kN (2*50 kN wheel load). The three trailer’s axles of the typical truck are normally mounted with wide base tires. It is assumed in this case that the heavily loaded trailer’s axle has an axle load of 150 kN (i.e. 75 kN wheel load).
Section 5.2 presents the pavement structural analysis. The strains and stresses in the pavement structures were analyzed using the linear – elastic multilayer computer program BISAR. Reasonable assumptions were applied with regard to the materials properties and characteristics due to the fact that very limited information was obtained. The South African material classified as G1 tested at the Laboratory of Road and Railway Engineering of the Delft University of Technology was assumed to be equivalent to the Tanzanian CRR material. The
82 cemented material classified as C1 in Tanzania was assumed to be equivalent to the cemented material classified as C4 in South Africa. In both asphalt surface layer and asphalt base course layer the type of bitumen used is 40/60 pen-grade with asphalt temperature assumed to be 30˚C (assuming that heavy vehicles are travelling from the evening).
The computer program BANDS was used to determine the stiffness of both the bitumen and the asphalt mix.
Section 5.3 discusses the design criteria for the asphalt pavement layers. The South African Mechanistic Design Methods (SAMDM) was adopted in this study. The adoption is based on the fact that the South African conditions are assumed to be more or less similar to those in Tanzania. The Tanzanian pavement and materials design manual also proposes the use of SAMDM when the mechanistic pavement design method is to be applied. It has to be noted however that in reality the South African conditions are much different from Tanzanian conditions with specific example on the experience of the road engineers, quality and technology.
Section 5.4 discusses the estimation of the pavement life based on design criteria, material characteristics and loading conditions. Section 5.5 presents the analysis of the asphalt overlay and pavement reconstruction options for the section Mlandizi – Chalinze. Section 5.6 discusses the conclusion of the chapter.
Figure 5.1: Typical used truck (left) and trailer (right) that are found in Tanzania road network (Source: CRM Nederland BV)
83 5.2 Pavement Structural Analysis
The pavement structural analysis was done using the linear-elastic multilayer computer program BISAR to determine the stresses and strains in different locations of the pavement layers to be used as input in the design criteria for each of the two sections.
Pavement Structure 1: Mlandizi – Chalinze Section
Figure 5.2 shows the pavement structure with the locations where the stresses and strains were calculated based on the loading conditions and the materials characteristics. The modulus of the asphalt material for the surface layer and for the loading times of 0.01 s, 0.02 s and 0.04 s is 2890 MPa, 2110 MPa and 1490 MPa respectively while for the asphalt base layer the modulus is 3150 MPa, 2350 MPa and 1700 MPa respectively.
Figure 5.2: Stresses and Strains in Pavement structure 1
Table 5.1 shows the results output from the BISAR computer program for the 80 kN standard axle load (40 kN wheel load consisting of dual tyres each with 20 kN) at different loading time. In this case positive number means tensile while negative number means compression. Figures 5.3, 5.4, 5.5 and 5.6 show the relationship between the loading time and the stresses and strains in different locations.
It can be observed from figure 5.4 that the tensile strains at the bottom of both asphalt layer and cement – bound layers increases with an increase in loading
84 time. In figure 5.6 the compressive strain at the top of the sub-grade is observed to increase with increase in loading time.
Table 5.1: Stresses and strains in pavement structure 1- wheel load 40 kN
Time of Loading σ1 σ2 σ3 εh εv σv σh Layer Loaction [s] [MPa] [MPa] [MPa] [μm/m] [μm/m] [MPa] [MPa] 0.04 -0.404 -0.415 -0.691 1 mid depth 0.02 -0.425 -0.440 -0.690 0.01 -0.448 -0.467 -0.689
0.04 37 1 bottom 0.02 33 0.01 29
50 mm 0.04 -0.192 -0.090 -0.443 2 from top of 0.02 -0.194 -0.093 -0.435 layer 2 0.01 -0.194 -0.093 -0.435
0.04 27 2 bottom 0.02 30 0.01 32
0.04 -0.193 3 top 0.02 -0.178 0.01 -0.164
0.04 55 0.130 3 bottom 0.02 52 0.123 0.01 49 0.117
0.04 -139 4 top 0.02 -130 0.01 -121
85 Main Stresses - wheel load 40 kN
0.0 0 0.01 0.02 0.03 0.04 0.05 -0.2
-0.4
-0.6
Main Stress[MPa] Main -0.8 Loading Time [s]
Sigma1(x)-asphalt sigma3(z)-asphalt sigma1(x)-asphalt base sigma3(z)-asphalt base
Figure 5.3: Variation of main stresses with loading time for pavement structure 1 - wheel load 40 kN
Tensile Strains - wheel load 40 kN
60.0 50.0 40.0 30.0
20.0 [microstrain] Tensile Strain 10.0 0.0 0 0.01 0.02 0.03 0.04 0.05 Time of Loading [s]
Asphalt surface layer Asphalt base layer Cement stabilized sub-base
Figure 5.4: Variation of tensile strains with loading time for pavement structure 1 - wheel load 40 kN
86 Tensile Stresses - wheel load 40 kN
0.140 0.120 0.100 0.080 0.060 0.040 Stress[MPa] 0.020 0.000 0 0.01 0.02 0.03 0.04 0.05 Time of Loading [s]
Cement stabilized sub-base
Figure 5.5: Variation of tensile stresses with loading time for pavement structure 1 - wheel load 40 kN
Compressive Strain - wheel load 40 kN
0.0 -20.0 0 0.01 0.02 0.03 0.04 0.05 -40.0 -60.0 -80.0 -100.0
-120.0 Strain [microstrain] Strain -140.0 Time of Loading [s]
Subgrade
Figure 5.6: Variation of compressive strains with loading time for pavement structure 1 - wheel load 40 kN
Table 5.2 shows the results output from the BISAR computer program for the 200 kN axle load (100 kN wheel load consisting of dual tyres each with 50 kN) at different loading time. Figures 5.7, 5.8, 5.9 and 5.10 show the variation of the stresses and strains with the loading time in different locations in the pavement.
It can be observed from table 5.2 and figure 5.8 that there exist no tensile strains at the bottom of the asphalt layer in all loading times. In figure 5.10 the
87 compressive strain at the top of the sub-grade is also observed here to increase with increase in loading time.
Table 5.2: Stresses and strains in pavement structure 1 - wheel load 100 kN
Time of Loading σ1 σ2 σ3 εh εv σv σh Layer Loaction [s] [MPa] [MPa] [MPa] [μm/m] [μm/m] [MPa] [MPa] 0.04 -0.602 -0.627 -0.796 1 mid depth 0.02 -0.652 -0.686 -0.795 0.01 -0.703 -0.749 -0.795
0.04 -2 1 bottom 0.02 -16 0.01 -7
50 mm 0.04 -0.385 -0.247 -0.658 2 from top of 0.02 -0.390 -0.254 -0.645 layer 2 0.01 -0.391 -0.257 -0.632
0.04 50 2 bottom 0.02 59 0.01 65
0.04 -0.384 3 top 0.02 -0.358 0.01 -0.332
0.04 131 0.312 3 bottom 0.02 124 0.295 0.01 117 0.280
0.04 -338 4 top 0.02 -315 0.01 -295
88 Main Stresses - wheel load 100 kN
0.0 -0.2 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 -0.4 -0.6
Stress[MPa] -0.8 -1.0 Time of Loading [s]
Sigma 1(x) - asphalt surface Sigma 3(z) -asphalt layer Sigma 1(x) - asphalt base layer Sigma 3(z) - asphalt base layer
Figure 5.7: Variation of main stresses with loading time for pavement structure 1 - wheel load 100 kN
Tensile Strains - wheel load 100 kN
150.0
100.0
50.0
0.0
Strain [microstrain] Strain 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 -50.0 Time of Loading [s]
Asphalt Layer Asphalt base layer Cement stabilized sub-base
Figure 5.8: Variation of tensile strains with loading time for pavement structure 1 - wheel load 100 kN
89 Tensile Stresses - wheel load 100 kN
0.350 0.300 0.250 0.200 0.150 0.100 Stress[MPa] 0.050 0.000 0 0.01 0.02 0.03 0.04 0.05 Time of Loading [s]
Cement stabilized sub-base
Figure 5.9: Variation of tensile stresses with loading time for pavement structure 1- wheel load 100 kN
Compressive Strain - wheel load 100 kN
0.0 0 0.01 0.02 0.03 0.04 0.05 -100.0
-200.0
-300.0 Strain [microstrain] Strain -400.0 Time of Loading [s]
Subgrade
Figure 5.10: Variation of compressive strains with loading time for pavement structure 1 - wheel load 100 kN
Table 5.3 shows the results output from the BISAR computer program for the 150 kN axle load (75 kN wheel load, wide base tyre) at different loading time. Figures 5.11, 5.12, 5.13 and 5.14 show the variation of the stresses and strains with the loading time in different locations in the pavement.
90 It can be observed also from table 5.3 and figure 5.12 that there exist no tensile strains at the bottom of the asphalt layer in all loading times. In figure 5.14 the compressive strain at the top of the sub-grade is also observed in this case to increase with increase in loading time.
Table 5.3: Stresses and strains in pavement structure 1 - wheel load - 75 kN
Time of Loading σ1 σ2 σ3 εh εv σv σh Layer Loaction [s] [MPa] [MPa] [MPa] [μm/m] [μm/m] [MPa] [MPa] 0.04 -0.724 -0.724 1 mid depth 0.02 -0.783 -0.783 0.01 -0.845 -0.845
0.04 -14 1 bottom 0.02 -22 0.01 -25
50 mm 0.04 -0.286 -0.286 2 from top of 0.02 -0.289 -0.289 layer 2 0.01 -0.288 -0.288
0.04 57 2 bottom 0.02 63 0.01 66
0.04 -0.482 3 top 0.02 -0.445 0.01 -0.408
0.04 109 0.270 3 bottom 0.02 102 0.254 0.01 96 0.239
0.04 -302 4 top 0.02 -280 0.01 -260
91 Main Stresses - wheel load 75 kN
0.0 0 0.01 0.02 0.03 0.04 0.05 -0.5
-1.0 Stress[MPa]
-1.5 Time of Loading [s]
Sigma 1(x) -Asphalt Layer Sigma 3(z)-Asphalt Layer Sigma 1(x) - Asphalt base layer Sigma 3(z) - Asphalt base layer
Figure 5.11: Variation of main stresses with loading time for pavement structure 1 - wheel load 75 kN
Tensile Strains - wheel load 75 kN
150.0
100.0
50.0
0.0
Strain [microstarin] Strain 0 0.01 0.02 0.03 0.04 0.05 -50.0 Time of Loading [s]
Aspahlt Layer Asphalt base layer Cement stabilized sub-base layer
Figure 5.12: Variation of tensile strains with loading time for pavement structure 1 - wheel load 75 kN
92 Tensile Stresses - wheel load 75 kN
0.300 0.250 0.200 0.150 0.100
Stress[MPa] 0.050 0.000 0 0.01 0.02 0.03 0.04 0.05 Time of Loading [s]
Cement stabilized sub-base
Figure 5.13: Variation of tensile stresses with loading time for pavement structure 1 - wheel load 75 kN
Compressive Strains - wheel load 75 kN
0.0 -50.0 0 0.01 0.02 0.03 0.04 0.05 -100.0 -150.0 -200.0 -250.0
-300.0 Strain [microstrain] Strain -350.0 Time of Loading
Subgrade
Figure 5.14: Variation of compressive strains with loading time for pavement structure 1 - wheel load 75 kN
Pavement Structure 2: Chalinze - Morogoro Section
Figure 5.15 shows the pavement structure with the locations where the stresses and strains were calculated using BISAR computer program based on the loading conditions and the materials characteristics. The modulus of the asphalt material
93 for the surface layer and for the loading times of 0.01s, 0.02s and 0.04s are 2890 MPa, 2110 MPa and 1490 MPa respectively.
Figure 5.15: Stresses and strains in pavement structure 2
Table 5.4 shows the results output from the BISAR computer program for the 80 kN standard axle load (40 kN wheel load consisting of dual tyres each with 20 kN) at different loading time. Figures 5.16, 5.17, 5.18 and 5.19 show the variation of stresses and strains with the loading time in different locations of pavement layers.
It can be observed from figure 5.17 that the tensile strain at the bottom of both asphalt layer and cement – bound layers increases with an increase in loading time. In figure 5.19 the compressive strain at the top of the sub-grade is observed to increase with increase in loading time.
94 Table 5.4: Stresses and strains in pavement structure 2 – wheel load 40 kN
Time of Loading σ1 σ2 σ3 εh εv σv σh Layer Location [s] [MPa] [MPa] [MPa] [μm/m] [μm/m] [MPa] [MPa] 0.04 -0.202 -0.198 -0.585 1 mid- 0.02 -0.201 -0.197 -0.563 depth 0.01 -0.199 -0.195 -0.544
0.04 176 1 bottom 0.02 159 0.01 142
0.04 -0.103 -0.029 -0.255 2 50 mm 0.02 -0.094 -0.029 -0.235 from top 0.01 -0.086 -0.030 -0.217
0.04 -0.150 3 top 0.02 -0.142 0.01 -0.135
0.04 20 0.030 3 bottom 0.02 20 0.030 0.01 20 0.031
0.04 -0.063 4 top 0.02 -0.061 0.01 -0.058
0.04 44 0.106 4 bottom 0.02 43 0.104 0.01 42 0.102
0.04 -117 5 top 0.02 -114 0.01 -111
95 Main Stresses - wheel load 40 kN
0.0 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 -0.2
-0.4
-0.6 Stress[MPa]
-0.8 Time of Loading [s]
Sigma 1(x)- Asphalt layer Sigma 3(z) - Asphalt layer Sigma 1(x) -Granular base Sigma 3(z) - Granular base
Figure 5.16: Variation of main stresses with loading time for pavement structure 2 - wheel load 40 kN
Tensile Strains - wheel load 40 kN
200.0
150.0
100.0
50.0
Strain [microstrain] Strain 0.0 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 Time of Loading [s]
Asphalt layer Cement bound base layer Cement bound sub-base layer
Figure 5.17: Variation of tensile strains with loading time for pavement structure 2 - wheel load 40 kN
96 Tensile Stresses - wheel load 40 kN
0.120 0.100 0.080 0.060 0.040
Stress[MPa] 0.020 0.000 0 0.01 0.02 0.03 0.04 0.05 Time of Loading [s]
Cement bound base layer Cement bound sub-base layer
Figure 5.18: Variation of tensile stresses with loading time for pavement structure 2 - wheel load 40 kN
Compressive Strains - wheel load 40 kN
0.0 -20.0 0 0.01 0.02 0.03 0.04 0.05 -40.0 -60.0 -80.0 -100.0
-120.0 Strain [micro strain] Strain -140.0 Time of Loading [s]
Subgrade
Figure 5.19: Variation of compressive strains with loading time for pavement structure 2 - wheel load 40 kN
Table 5.5 shows the results output from the BISAR computer program for the 200 kN axle load (100 kN wheel load consisting of dual tyres each with 50 kN) at different loading time. Figures 5.20, 5.21, 5.22 and 5.23 show the variation of the stresses and strains with the loading time in different locations in the pavement.
97 It can be observed from figure 5.21 that the tensile strain at the bottom of both asphalt layer and cement – bound layers increases with an increase in loading time. In figure 5.23 the compressive strain at the top of the sub-grade is observed to increase with increase in loading time.
Table 5.5: Stresses and strain in pavement structure 2 - wheel load 100 kN
Time of Loading σ1 σ2 σ3 εh εv σv σh Layer Location [s] [MPa] [MPa] [MPa] [μm/m] [μm/m] [MPa] [MPa] 0.04 -0.453 -0.373 -0.735 1 mid- 0.02 -0.465 -0.379 -0.714 depth 0.01 -0.474 -0.383 -0.693
0.04 247 1 bottom 0.02 236 0.01 222
0.04 -0.207 -0.094 -0.488 2 50 mm 0.02 -0.192 -0.088 -0.462 from top 0.01 -0.178 -0.085 -0.436
0.04 -0.330 3 top 0.02 -0.316 0.01 -0.302
0.04 48 0.070 3 bottom 0.02 48 0.072 0.01 47 0.073
0.04 -0.149 4 top 0.02 -0.144 0.01 -0.139
0.04 107 0.255 4 bottom 0.02 105 0.251 0.01 103 0.241
0.04 -285 5 top 0.02 -278 0.01 -271
98 Main Stresses - wheel load 100 kN
0.0 0 0.01 0.02 0.03 0.04 0.05 -0.2
-0.4
-0.6 Stress[MPa]
-0.8 Time of Loading [s]
Sigma 1(x) - Asphalt layer Sigma 3(z)-Asphalt layer Sigma 3(z) - Granular base layer Sigma 1(x) - Granular base layer
Figure 5.20: Variation of main stresses with loading time for pavement structure 2 - wheel load 100 kN
Tensile Strains - wheel load 100 kN
300.0
200.0
100.0
0.0 Strain [microstrain] Strain 0 0.01 0.02 0.03 0.04 0.05 Time of Loading [s]
Asphalt layer Cement bound base layer Cement bound sub-base layer
Figure 5.21: Variation of tensile strains with loading time for pavement structure 2 - wheel load 100 kN
99 Tensile Stresses - wheel load 100 kN
0.300 0.250 0.200 0.150 0.100
Stress[MPa] 0.050 0.000 0 0.01 0.02 0.03 0.04 0.05 Time of Loading [s]
Cement bound base layer Cement bound sub-base layer
Figure 5.22: Variation of tensile stresses with loading time for pavement structure 2 - wheel load 100 kN
Compressive Strains - wheel load 100 kN
0.0 -50.0 0 0.01 0.02 0.03 0.04 0.05 -100.0 -150.0 -200.0
-250.0 Strain [microstrain] Strain -300.0 Time of Loading [s]
Subgrade
Figure 5.23: Variation of compressive strains with loading time for pavement structure 2 - wheel load 100 kN
Table 5.6 shows the results output from the BISAR computer program for the 150 kN axle load (75 kN wheel load, wide base tyre) at different loading time. Figures 5.24, 5.25, 5.26 and 5.27 show the variation of the stresses and strains with the loading time in different locations in the pavement.
It can be observed from figure 5.25 that the tensile strain at the bottom of both asphalt layer and cement – bound layers increases with an increase in loading
100 time. In figure 5.27 the compressive strain at the top of the sub-grade is observed to increase with increase in loading time.
Table 5.6: Stresses and strains in pavement structure 2 - wheel load 75 kN
Time of Loading σ1 σ2 σ3 εh εv σv σh Layer Location [s] [MPa] [MPa] [MPa] [μm/m] [μm/m] [MPa] [MPa] 0.04 -0.444 -0.444 -0.931 1 mid- 0.02 -0.442 -0.442 -0.904 depth 0.01 -0.438 -0.438 -0.877
0.04 261 1 bottom 0.02 249 0.01 233
0.04 -0.117 -0.117 -0.637 2 50 mm 0.02 -0.110 -0.110 -0.598 from top 0.01 -0.104 -0.104 -0.560
0.04 -0.403 3 top 0.02 -0.382 0.01 -0.360
0.04 41 0.058 3 bottom 0.02 40 0.059 0.01 40 0.059
0.04 -0.153 4 top 0.02 -0.146 0.01 -0.140
86 0.214 4 bottom 84 0.210 83 0.205
0.04 -225 5 top 0.02 -219 0.01 -213
101 Main Stresses - wheel load 75 kN
0 -0.2 0 0.01 0.02 0.03 0.04 0.05 -0.4 -0.6
Stress[MPa] -0.8 -1 Time of Loading [s]
Sigma 1(x) - Asphalt layer Sigma 3(z) - Asphalt layer Sigma 1(x) - Granual base layer Sigma 3(z) - Granular base layer
Figure 5.24: Variation of main stresses with loading time for pavement structure 2 - wheel load 75 kN
Tensile Strains - wheel load 75 kN
300.0
200.0
100.0
0.0 Strain [microstrain] Strain 0 0.01 0.02 0.03 0.04 0.05 Time of Loading [s]
Asphalt layer Cement bound base layer Cement bound sub-base layer
Figure 5.25: Variation of tensile strains with loading time for pavement structure 2 - wheel load 75 kN
102 Tensile Stresses - wheel load 75 kN
0.25 0.2 0.15 0.1
Stress[MPa] 0.05 0 0 0.01 0.02 0.03 0.04 0.05 Time of Loading [s]
Cement bound base layer Cement bound sub-base layer
Figure 5.26: Variation of tensile stresses with loading time for pavement structure 2 - wheel load 75 kN
Compressive Strains - wheel load 75 kN
0.0 0 0.01 0.02 0.03 0.04 0.05 -50.0
-100.0
-150.0
-200.0 Strain [microstrain] Strain -250.0 Time of Loading [s]
Subgrade
Figure 5.27: Variation of compressive strains with loading time for pavement structure 2 - wheel load 75 kN
5.3 Design Criteria
As explained in section 5.1 the design criteria by the South African Mechanistic Design Method (SAMDM) are adopted in this study. Different transfer functions have been developed in SAMDM for different design criteria and for different asphalt pavement layers and materials characteristics. In this case the design criteria for asphalt layer, cement and hydraulically bound (sub) base layers, unbound (sub) base layers and sub-grade are discussed in the following sections.
103 5.3.1 Design criterion for asphalt Layers
The design criterion is to prevent the asphalt from fatigue cracking (from bottom to top).
With respect to fatigue, in SAMDM a distinction is made between thin asphalt surfacing layers (thickness less than 50 mm) and thick asphalt bases (thickness more than 75 mm).
All fatigue relations are provided by the following expression:
N c. n ...... 5.1
Where: N = allowable number of repetitions of strain = flexural tensile strain at the bottom of the asphalt (µm/m)
The values of c and n are given for different asphalt layers and for different categories of road. In this case for an equivalent road category A (interurban freeways and major interurban roads) the following values for c and n are provided: c = 25.12*1016 and n = 5.12 for continuously graded asphalt surfacing layer c= 1.23*1016 and n = 4.79 for thick asphalt base with E = 2000 MPa c= 0.60*1016 and n = 4.73 for thick asphalt base with E = 3000 MPa
A shift factor (SF) is given to convert the crack initiation life based on the laboratory fatigue relations to the total fatigue life after surface cracks appear on the road surface. Figure 5.28 shows the SF as a function of the asphalt layer thickness.
104
Figure 5.28: Fatigue crack propagation shift factor for asphalt layers
5.3.2 Design criterion for cement and hydraulically bound (sub) base layers
The design criterion taken into account is the structural fatigue at the bottom of the bound base layer and the crushing at the top of the base layer
In the SAMDM the critical parameter with respect to fatigue is the maximum flexural tensile strain at the bottom of the cemented layer. The following effective fatigue transfer function is applied:
log Neff 7.06 0.201u1 /7.86 0.214u b ...... 5.2
Where: N eff = number of repetitions of flexural tensile strains = flexural tensile strain (m/ m)
b = strain at break (m/ m) u = u-value of standard normal distribution (u = 1.645 for category A roads)
A SF is used to allow thicker layers to have an extended effective fatigue life compared to thinner layers subjected to the same strain. The SF is given by the following expression:
SF = 10(0.00285d-0.293) ...... 5.3
Where: d = thickness of the bound layer in mm
105 SF = 1 if d < 102 SF = 8 if d > 419
In SAMDM the critical parameter with respect to crushing is the maximum vertical compressive stress at the top of the cemented layer. Transfer functions are provided for two crushing conditions, namely crush initiation with roughly 2 mm permanent deformation on top of the layer and advanced crushing with 10 mm permanent deformation and extensive breakdown of the cemented material.
The fatigue-type of transfer function with respect to crush initiation is:
log NCi 8.516 0.79u1 v /(1.21 0.085u)UCS ……………………………………..5.4
The fatigue-type of transfer function with respect to advanced crushing is:
log NCa 8.894 0.55u1 v /(1.31 0.085u)UCS ……………………………………..5.5
Where: Nci = allowable number of repetitions of stress v until crush initiation
Nca= allowable number of repetitions of stress until advanced cracking = vertical stress at the top of the base (kPa) UCS = unconfined compressive strength of the base material (kPa) u = u-value of standard normal distribution (u=1.645 for category A roads)
5.3.3 Design criterion for unbound (sub) base layers
In SAMDM a design criterion included aims to limit the permanent deformation in granular (sub) base layers. The occurring shear stresses due to the traffic loadings should be substantially smaller than the shear strength. This is in order to not only prevent shear failure but also to limit the development of permanent deformation in the (sub) base during the service life of the road which would reflect as rutting at the pavement surface.
The safety factor against shear failure is defined using the Mohr-Coulomb failure curve. The safety factor F represents the ratio of the material’s shear strength and the occurring shear stress and is given by the following expression:
1 Ktan2 (45 / 2) 1 2Kc tan(45 / 2) F 3 ...... 5.6 R ( 1 3 )
Where: F = safety factor against shear failure R = ratio of occurring shear stress and shear strength
106 1 = major principle stress (kPa) - (compressive stress is positive)
3 = minor principle stress (kPa) – (compressive stress is positive) c = cohesion (kPa) = angle of internal friction (°) K= constant (0.65 for saturated conditions, 0.80 for moderate moisture conditions, 0.95 for normal moisture conditions)
For static loads the safety factor F must be greater than 1 while for dynamic loads the safety factor F is related to allowable number of load repetitions N for different road categories given by the following expression: log N = 2.60 F – (3.708 + 0.473u) ...... 5.7
Where: N = allowable number of load repetitions F = safety factor against shear u = u-value of standard normal distribution (u=1.645 for category A roads)
5.3.4 Design criterion for sub-grade
The design criterion aiming to limit the permanent deformation in the sub-grade during the service life of the road (which also would reflect at the road surface as rutting) is given in SAMDM by the following transfer functions for the maximum vertical elastic compressive strain at the top of the sub-grade:
For 10 mm terminal rut condition the following expression is used:
log N 26.30 0.255u 10logV ...... 5.8
For 20 mm terminal rut condition the following expression is used:
log N 23.30 0.255u 10logV ...... 5.9
Where: N = allowable number of load repetitions
v = maximum vertical elastic compressive strain at the top of the sub-grade (in m/m) u = u-value of standard normal distribution (u = 1.645 for category A roads)
107 5.4 Estimation of the pavement life
For both two pavement structures the estimation of the pavement life is done by determining the allowable number of load repetitions based on the design criteria discussed in section 5.3 for the stresses and strains occurring by different loading conditions presented in section 5.2. The occurring number of repetitions of heavy goods vehicles per year (AADT) from the collected data on the case road is then compared with the allowable number of load repetitions.
Pavement structure 1
Equation 5.1 is used to determine the allowable number of load repetitions for the asphalt surface layer and asphalt base course layer. The values for c and n are 25.12E16 and 5.12 respectively for the continuously graded asphalt layer and 1.23E16 and 4.79 respectively for the asphalt base course layer.
Table 5.7 shows the allowable number of load repetitions calculated for the asphalt surface layer for different loading cases. The maximum tensile strains at the bottom of the layer are extracted from tables 5.1, 5.2 and 5.3. The maximum value was obtained at a loading time of 0.04 s (v = 30 km/hr).
Table 5.7: Allowable number of load repetitions for asphalt layer – structure 1
Wheel load [kN] εb [μm/m] SF Nallowable 40 37 2.2 5.14e09 100 No tensile strain 2.2 N/A 75 No tensile strain 2.2 N/A
For the asphalt base course layers the maximum tensile strains at the bottom of the layer were extracted from tables 5.1, 5.2 and 5.3. in this case the maximum values were obtained at a loading time of 0.01 s. Table 5.8 shows the allowable number of load repetitions calculated for different loading cases.
Table 5.8: Allowable number of load repetitions for asphalt base layer - structure 1
Wheel load [kN] εb [μm/m] SF Nallowable 40 32 7.1 5.74e09 100 65 7.1 1.85e08 75 66 7.1 1.66e08
For the cement stabilized sub-base layer the equations 5.2, 5.3, 5.4 and 5.5 are used. The cement stabilized material used is assumed to be C1 Tanzanian classified material, which is equivalent to C4 South African classified, cemented material thus UCS of 1000 kPa is assumed. SAMDM proposes a strain at break of
108 145 μm/m for the C4 classified material. From equation 5.3, a SF of 2.6 is calculated for a 250 mm thick layer. Table 5.9 shows the allowable number of load repetitions for both structural fatigue and crushing criteria for each load case. The values of tensile strains and compressive stresses are extracted from tables 5.1, 5.2 and 5.3.
Table 5.9: Allowable number of load repetitions for cement bound sub-base layer - structure 1
Tensile Compressive Wheel load strain stress σv Neff Nci Nca [kN] εb [μm/m] [kPa] 40 55 193 6.3e06 0.82e06 0.22e08 100 131 384 2.18e06 42,364 0.82e06 75 109 482 3.0e06 9,251 150,480
20 mm terminal rut condition is adopted as design criterion for the sub-grade thus equation 5.9 is used. The maximum vertical elastic strains are extracted from tables 5.1, 5.2 and 5.3. Table 5.10 shows the allowable number of load repetitions for different load cases.
Table 5.10: Allowable number of load repetitions for sub-grade - structure 1
Whel load Vertical compressive strain [kN] εv [m/m] Nallowable 40 139.0e-06 7.10e14 100 337.8e-06 1.01e11 75 302.0e-06 3.03e11
Table 5.11 shows a summary of the critical layers with the respective allowable number of load repetitions for each load case.
Table 5.11: Summary of critical allowable number of load repetitions – structure 1
Wheel load Critical layer Nallowable Nallowable [kN] Light Damage Severe Damage 40 Cement bound sub-base 0.82e06 6.30e06 100 Cement bound sub-base 42,364 0.82e06 75 Cement bound sub-base 9,251 150,480
Pavement structure 2
Similar to pavement structure 1, Equation 5.1 is used to determine the allowable number of load repetitions for the asphalt surface layer. The values for c and n
109 are again 25.12E16 and 5.12 respectively for the continuously graded asphalt layer.
Table 5.12 shows the allowable number of load repetitions calculated for the asphalt surface layer for different loading cases. The maximum tensile strains at the bottom of the layer are extracted from tables 5.4, 5.5 and 5.6.
Table 5.12: Allowable number of load repetitions for asphalt layer- structure 2
Wheel load [kN] εb [μm/m] SF Nallowable 40 176 4.4 3.57e06 100 247 4.4 0.62e06 75 261 4.4 0.47e06
For the unbound granular base course layer the CRR material was assumed to be equivalent to the G1 South African granular material tested at the Delft University of Technology laboratories. The cohesion c and angle of internal friction θ for the material were assumed to be 160 kPa and 55˚ respectively. It has to be kept in mind however that these values are dependent of the material gradation, degree of compaction and moisture content. Equation 5.6 is used to calculate the safety factor against shear failure F while equation 5.7 is used to calculate the allowable number of load repetitions for different load cases. A moderate moisture condition is assumed thus the value of constant K adopted is 0.80. The minor and major stresses are extracted from tables 5.4, 5.5 and 5.6. It has to be noted that the values of σ3 in the tables are major stresses σ1 in the equation 5.6 while the values of σ3 in equation 5.6 are extracted from tables as σ1 or σ2, whichever is small.
Table 5.13 shows the allowable number of load repetitions for the different loading cases.
Table 5.13: Allowable number of load repetitions for unbound granular base layer - structure 2
Wheel load [kN] σ1 [kPa] σ3 [kPa] F Nallowable 40 255 29 2.2 17.14e06 100 488 94 2.4 70.99e06 75 637 117 2.2 17.14e06
For the cement bound base layer and the cement stabilized sub-base layer equations 5.2, 5.3, 5.4 and 5.5 are used. Similar to structure 1, the cement stabilized material used is assumed to be C1 Tanzanian classified material which is equal to C4 South African classified cemented material, thus UCS of 1000 kPa is assumed. A strain at break is assumed to be 145 μm/m. From equation 5.3, a SF of 1.36 is calculated for a 150 mm thick layer. Tables 5.14 and 5.15 show the
110 allowable number of load repetitions for both structural fatigue and crushing criteria for each load case for the bound base course and the bound sub-base course respectively. The values of tensile strains and compressive stresses are extracted from tables 5.4, 5.5 and 5.6.
Table 5.14: Allowable number of load repetitions for the bound base layer - structure 2
Tensile Compressive Wheel load strain stress σv Neff Nci Nca [kN] εb [μm/m] [kPa] 40 20 150 5.44e06 1.60e06 4.73e07 100 48 330 3.70e06 97,989.56 2.10e06 75 41 403 4.09e06 20,742.20 0.59e06
Table 5.15: Allowable number of load repetitions for the bound sub-base layer - structure 2
Tensile Compressive Wheel load strain stress σv Neff Nci Nca [kN] εb [μm/m] [kPa] 40 44 63 3.88e06 8.40e06 2.14e08 100 107 149 1.59e06 1.60e06 4.82e07 75 86 153 2.13e06 1.53e06 4.5e07
Similar to the structure 1, 20 mm terminal rut condition is also adopted as design criterion for the sub-grade in this structure, thus equation 5.9 is used. The maximum vertical elastic strains are extracted from tables 5.4, 5.5 and 5.6. Table 5.16 shows the allowable number of load repetitions for different load cases.
Table 5.16: Allowable number of load repetitions for sub-grade - structure 2
Whel load Vertical compressive strain [kN] εv [m/m] Nallowable 40 116.9e-06 4.00e15 100 285.0e-06 5.40e11 75 225.0e-06 5.74e11
Table 5.17 shows a summary of the critical layers with the respective allowable number of load repetitions for each load case.
111 Table 5.17: Summary of critical allowable number of load repetitions - structure 2
Wheel Critical layer Nallowable Nallowable load [kN] Light damage Severe damage 40 Cement bound (light damage)/Asphalt 1.60e06 3.57e06 (severe damage) 100 Cement bound (light damage)/Asphalt 97,990 0.62e06 (severe damage) 75 Cement bound (light damage)/Asphalt 20,743 0.47e06 (severe damage)
To estimate the pavement life the occurring number of load repetitions shall be compared to the derived allowable number of load repetitions. Table 5.18 shows the traffic data for heavy goods vehicles and large buses for the two sections.
The total ESAL per year when assuming heavy goods vehicles and large busses are traveling everyday in a year is then derived from the following expression:
ESAL/year = (AADAThv*ESALhv + AADTLb*ESALLb)*365 ………….………….………5.10
Where: hv = heavy vehicles Lb = Large buses
From equation 5.10 and table 5.18 the total vehicles per design lane [ESAL/2] for heavy vehicles and large buses for the sections Mlandizi – Chalinze reconstructed in 1994 and Chalinze – Morogoro reconstructed in 2004 are 1,346,100 in the year 2007 and 1,010,000 in the year 2004 respectively.
Table 5.18: Traffic data for sections Mlandizi – Chalinze and Chalinze - Morogoro
Mlandizi – Chalinze Chalinze - Morogoro AADT – Heavy vehicles 399 496 AADT – Large buses 1298 575
AADTyear 2007 2004 Mean ESAL – Heavy vehicle 7.1 7.1 Mean ESAL – Large buses 3.5 3.5
Figures 5.29 and 5.30 show the number of vehicles per year and the total ESAL per year respectively for the section Mlandizi – Chalinze while figures 5.31 and 5.32 show the number of vehicles per year and the total ESAL per year respectively for the section Chalinze – Mlandizi. The vehicle growth rate of 5% per year for both cases is adopted in this study.
112 Vehicles per Year for Mlandizi - Chalinze
2,000
1,500
1,000 AADT 500
- 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 Year
Large Buses Heavy Vehicles
Figure 5.29: Vehicles per year for section Mlandizi – Chalinze
ESAL per YEAR - MLANDIZI - CHALINZE
3,500,000 3,000,000 2,500,000 2,000,000 1,500,000
ESAL/Year 1,000,000 500,000 - 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 Year
Figure 5.30: ESAL per year for section Mlandizi - Chalinze
Vehicles per Year - Chalinze - Morogoro
800
600
400 AADT 200
0 2003 2004 2005 2006 2007 2008 2009 2010 Year
Large Buses Heavy Vehicles
Figure 5.31: Vehicles per year for section Chalinze - Mlandizi
113 ESAL per YEAR - Chalinze - Morogoro
3,000,000
2,500,000 2,000,000
1,500,000
ESAL/Year 1,000,000
500,000 - 2003 2004 2005 2006 2007 2008 2009 2010 Year
Figure 5.32: ESAL per year for section Chalinze – Morogoro
The cumulative total ESAL in the year 2009 for the heavy vehicles and the large buses for the sections Mlandizi – Chalinze and Chalinze - Morogoro therefore become 33,776,535 and 13,739,864 respectively. This implies that the vehicles per the design lane become 16,888,268 and 6,869,932 for the sections Mlandizi- Chalinze and Chalinze – Morogoro respectively.
These occurring number of standard axle load repetitions are compared with the calculated allowable number of standard axle load repetitions. No comparison is done for the axle loads 200 kN and 150 kN due to non availability of the data for the occurrence of these axle loads cases. Table 5.19 shows the comparison between the allowable number of standard axle loads repetitions and the occurring cumulative standard axle load repetitions as of 2009 for the two sections.
It can be observed from the table 5.19 that in both sections the occurring number of standard axle load repetitions exceeds the allowable number of standard axle loads.
The comparison on the section Mlandizi – Chalinze shows that the cement bound sub-base which is the decisive layer must have failed completely. No asphalt fatigue damage was calculated. However, the visual condition survey observations as discussed in previous sections revealed that the asphalt surface layer is also severely damaged. Several typical defects that include pavement cracks, rutting and potholes were observed.
For the section Chalinze – Morogoro the comparison shows that the asphalt layer which is the decisive layer in this case must have been degraded. However this is contrary to the visual condition survey observations. The visual condition survey revealed no major defects on the asphalt surface.
114 Table 5.19: Comparison of allowable and occurring axle loads
Section Nallowable Noccurring Mlandizi - Chalinze 6.30e06 16.89e06 Chalinze - Morogoro 3.57e06 6.87e06
It has to be noted that the previous analytical calculations assumed that the material characteristics remained constant throughout the pavement life. However, the material characteristics do change with time. In this study the analytical calculations were repeated for the pavement structure 1 for the section Mlandizi – Chalinze when considering the changes in material behaviour. Figure 5.33 shows the long term behaviour of the cemented material. Three phases namely pre-cracked, effective fatigue life and equivalent granular are distinguished. During the pre-cracked phase the elastic modulus of the layer is assumed to be in the order of 2000 MPa. This modulus value reduces rapidly to the order of 1000 MPa in the effective fatigue life phase. During the equivalent granular phase the elastic modulus is reduced to the order of 500 MPa and the cemented material acts like a granular layer. Although these changes in the behaviour will gradually occur with time, they are normally modelled as stepwise phases for simplicity.
Figure 5.34 shows the derived tensile strains in the asphalt layers for the three discussed phases for the standard axle loads. It has been assumed in the analytical calculations that the asphalt modulus for both surface layer and base course layer are in the order of 100% in the pre-cracked phase, 75% in effective fatigue phase and 50% in the equivalent granular phase of cement bound layer. It can be observed from figure 5.34 that the tensile strains increase as the materials become inferior in these consecutive phases.
Figure 5.35 shows the calculated allowable number of load repetitions for the standard axle loads in the three discussed phases. It can be observed in this figure that the allowable number of load repetitions decreases when the materials become inferior within the three phases.
115 Reduced Modulus values for cement bound layer
2500
2000 Pre-cracked Effective fatigue life phase 1500 phase Equivalent granular phase
1000 E [MPa] E
500
0 0 2,000,000 4,000,000 6,000,000 8,000,000 10,000,000 12,000,000 14,000,000 16,000,000 18,000,000 Cummulative traffic loading
Figure 5.33: Long term behaviour of cemented material for pavement structure 1
Asphalt strain in long term
140 120 100 80 60 40 20
Strain [microstrain] Strain 0 - 2,000,000 4,000,000 6,000,000 8,000,000 10,000,000 12,000,000 14,000,000 16,000,000 18,000,000
Cummulative traffic loading
surface layer base layer
Figure 5.34: Long term asphalt strain for the pavement structure 1
Allowable number of load repetitions in long term
7000 6000 5000 4000 3000 (*10^6) 2000 1000 Allowable repetitions repetitions Allowable 0 0 2,000,000 4,000,000 6,000,000 8,000,000 10,000,000 12,000,000 14,000,000 16,000,000 18,000,000 Cummulative traffic loading
Asphalt surface Asphalt base layer
Figure 5.35: Allowable number of load repetitions for pavement structure 1
116 Table 5.20 shows the comparison between the derived allowable number of load repetitions and the derived occurring standard axle loads for the three different phases. It can be observed from table 5.20 that the asphalt base layer has failed completely during the equivalent granular phase.
Table 5.20: Asphalt fatigue life for pavement structure 1
Occurring Asphalt surface layer Asphalt base layer load Allowable Fatigue Allowable Fatigue repetition, load damage, load damage, Phase n repetition, N ∑n/∑N repetition, N ∑n/∑N
Pre-cracked phase 3.08E+06 5.15E+09 0.0006 5.74E+09 0.0005 Effective fatigue lfie phase 1.02E+07 1.15E+09 0.009 2.17E+08 0.047 Equivalent granular phase 1.69E+07 1.31E+08 0.129 8.00E+06 2.112
5.5 Optional Maintenance strategies
It has been shown by both mechanistic pavement analyses that the pavement structure 1 for the section Mlandizi – Chalinze has failed completely. It has also been observed during the visual road condition survey that this section is in comparative poor condition. The application of the HDM-4 model shows that the economic maintenance strategies preferred in this section are the application of the asphalt overlay of 50 mm thickness at an interval period of 9 yeras and the pavement reconstruction.
The mechanistic pavement analysis was used again in this case to determine the required asphalt overlay thickness for a pavement life extension with 10 years and to determine the required thickness of the asphalt base layer when the pavement is to be reconstructed by milling off the old asphalt surface layer and base layer for the design period of 20 years.
Option 1 – Asphalt Overlay
Figure 5.36 shows the proposed pavement structure for the section Mlandizi – Chalinze. The moduli for the original asphalt layers were assumed to be equal to 50% of the original values. The standard axle load case load at a loading time of 0.04 s was used. The design criterion adopted is to prevent the asphalt from fatigue cracking at the bottom of the asphalt overlay. Equation 5.1 was applied to determine the allowable number of load repetitions. The BISAR computer program was used to calculate the asphalt strains at the bottom of the asphalt
117 overlay. The total cumulative ESAL for the design lane after 10 years (i.e. to the year 2019) becomes 36.49e06 when assuming the vehicle growth rate of 5%.
Figure 5.36: Pavement structure 1 for asphalt overlay
Table 5.21 shows the asphalt fatigue damage for different asphalt overlay thicknesses. It can be observed from table 5.21 that for a range of overlay thicknesses 20 – 120 mm, asphalt overlay with thickness of 40 mm is the only option that can be applied in the section. It has to be noted that the results are based on the standard axle load case. It can be observed from the table that this proposed overlay will not be able to sustain the heavily overloaded axles (i.e. 150 kN and 200 kN analyzed before). 50 mm asphalt overlay with lifetime of 9 years was an optimum maintenance option resulted from the HDM-4 analyses discussed in chapter four. However, the intervention time for these two options is different. While this analysis assumed the overlay to be applied in the year 2009, the HDM-4 analyses indicated the 50 mm overlay to be applied in the year 2013.
118 Table 5.21: Asphalt fatigue damages for different overlay thicknesses h Strain at bottom of asphalt [mm] overlay, εb [μm/m] SF Nallowable ∑n/∑N 20 96 1.0 1.76E+07 2.076 30 96 1.1 1.97E+07 1.849 40 94 2.0 4.03E+07 0.907 50 121 2.2 1.22E+07 2.991 80 151 4.0 6.99E+06 5.220 100 145 4.4 9.49E+06 3.845 120 133 5.8 1.99E+07 1.838
Option 2 – Pavement reconstruction
The pavement reconstruction option in this case refers to the removal of the material from the existing pavement (mill off) from the asphalt base course layer to the surface layer and replaces them with new materials. The new materials were assumed to have the same characteristics as the original materials applied in the previous sections. The cemented sub-base layer is assumed to be in the equivalent granular phase. Figure 5.37 shows the proposed pavement structure for reconstruction. The cumulative ESAL after 20 years design period when assuming the vehicle growth rate of 5% is 68.41e06.
The design criterion adopted is to prevent asphalt from fatigue cracking at the bottom of both asphalt surface layer and asphalt base course layer. The BISAR computer program was used to calculate the tensile strains in these locations. Table 5.22 and 5.23 show fatigue damage for asphalt surface layer and the asphalt base layer at different base layer thicknesses. It can be observed from the two tables that for the given pavement structure the asphalt base course layer thickness of 210 mm is required for the pavement to sustain the standard axle loads for the given 20 years design period. It can be observed also in this case that the proposed pavement structure can not sustain the heavily overloaded axles (i.e. 150 kN and 200 kN analyzed before) for the given design period. The new structure is similar to what is proposed in the design catalogue provided by the Tanzanian Pavement and Materials Design Manual when applying bituminous mix base course and considering higher traffic loads.
119
Figure 5.37: Pavement structure 1 for reconstruction
Table 5.22: Asphalt fatigue damage for surface layer at different base layer thicknesses
Asphalt base Strain at bottom of thickness, h asphalt surface layer,εb [mm] [μm/m] SF Nallowable ∑n/∑N 100 54 2.2 7.37E+08 0.093 150 55 2.2 6.87E+08 0.100 200 50 2.2 1.10E+09 0.062 210 49 2.2 1.23E+09 0.056 250 45 2.2 1.98E+09 0.035
Table 5.23: Asphalt fatigue damage for base layer at different base layer thicknesses
Asphalt base Strain at bottom of thickness, h asphalt base layer,εb [mm] [μm/m] SF Nallowable ∑n/∑N 100 145 4.4 2.37E+06 28.867 150 107 7.1 1.65E+07 4.136 200 87 9.5 5.86E+07 1.168 210 84 10.0 7.51E+07 0.911 250 71 12.0 1.97E+08 0.347
120 5.6 Conclusion
It has been discussed in the introduction part of this chapter that the mechanistic pavement design analysis of the two sections of the case road was done to estimate the pavement life as well as to determine the favourable maintenance strategy for the section Mlandizi – Chalinze that was observed to be in poor condition. The analysis was done with consideration of the typical axle load cases that are believed to be found in Tanzania.
It has been shown from the analytical calculations that the cement bound layer used as a sub-base for the section Mlandizi – Chalinze must have been failed completely. No asphalt fatigue damage was calculated. However, it has been observed during the visual condition survey along the section that the asphalt layer was seriously damaged, with pavement cracks being among the defects observed. The mismatch can be explained by the fact that the transfer functions used in this study have been derived to be applied for the South African conditions. Theyse et al (1996) explained that the SAMDM has been continuously modified since developed for the first time in 1974 to suit the changes occurring in South African condition. He also explained that the design analysis method was calibrated extensively against the experience of the road engineers from different road authorities in South Africa. Although this method is proposed by the Tanzanian Pavement and Materials Design Manual, it has to be noted that the conditions in Tanzania are much different from those in South Africa, and the results obtained using these transfer functions in Tanzania environment must be treated with great care. The difference between the characteristics of the materials used during the reconstruction and the ones assumed during the analytical calculations probably also contributed into the observed mismatch.
The analytical calculations results show that the asphalt layer for the section Chalinze – Morogoro must have been degraded. This is contrary to what has been observed during the visual road condition survey. The asphalt surface layer in this section was observed to be in a good condition. As explained in the above paragraph, use of SAMDM for Tanzanian conditions and the assumptions made for the materials characteristics may have resulted into this contrary.
The repeated analytical calculations for the reduced modulus values of the asphalt material and cement bound sub-base layer for the section Mlandizi – Chalinze have shown that the asphalt base layer must have been failed completely when the cement bound layer reached the equivalent granular phase.
The results on the two maintenance strategies have shown that 40 mm asphalt overlay can sustain the standard axle loads for the design period of 10 years. The 40 mm overlay however can not sustain the heavily overloaded axles (i.e. 150 kN and 200 kN analyzed) for the design period.
121 The pavement reconstruction by milling off both the existing asphalt base and surface layers and applying new asphalt base and surface layers with thicknesses of 210 mm and 50 mm respectively was also the preferable option for the pavement to sustain the standard axle loads for the 20 years design period. The proposed structure however can not sustain the heavily overloaded axle loads for the same design period.
122 6 GENERAL CONCLUSIONS AND RECOMMENDATIONS
6.1 Introduction
In chapter 1 of the report the importance of road infrastructure in economic development of a country has been discussed. It was highlighted in the same chapter that poor roads condition is the main cause of Africa’s low competitiveness. A case of doing road maintenance appropriately and timely is un-debatable. Four specific objectives of the study were identified in chapter 1. Chapters 3, 4 and 5 were specifically included to achieve the intended objectives. In each of the chapters, specific conclusions were given.
In this chapter more general conclusions and recommendations are given to the study. Section 6.2 presents the general conclusions of the study. The general recommendations of the study based on the conclusions are presented in 6.3.
6.2 Conclusions
The conclusions given under this section include the conclusion on the road condition and maintenance practice in Tanzania supported by the visual road condition survey observations. It includes also the conclusion on the application of the HDM-4 model to predict the effect of maintenance strategies on the deterioration of the asphalt pavements, determination of optimum maintenance strategies and the sensitivity analyses. The conclusion on the application of the mechanistic pavement design analysis method in determining the favourable maintenance strategy in the selected section is also presented.
Road condition and maintenance practice in Tanzania
It has been explained in this study that only 5.9% of the entire road network of about 85,525 km is paved. It has been explained in this study also that the maintenance funding is still a big challenge faced by road authorities regarding road maintenance management. However, the initiative to reform the road sector has been taken by the government on which two important institutions, Road Fund Board (RFB) and Tanroads, have been established to overcome the challenges.
The road condition survey done on the section Dar es Salaam (Ubungo) – Chalinze – Morogoro revealed that most of the typical defects observed in the section are probably caused by the poor construction quality and vehicle overloading. While the maintenance funding has been identified to be a challenge in road maintenance management, vehicle overloading and poor construction quality also imposes challenges to the road authorities.
123 The road condition survey revealed that the section Mlandizi – Chalinze is in poor condition. Postponement of essential maintenance in this section will result in an increased cost of operating vehicles and thus higher transport cost.
Application of HDM-4 model for prediction of effect of maintenance strategies on the deterioration of asphalt pavement in Tanzania.
In this study, based on the data and information gathered, the results of the HDM-4 model analyses have shown that the model is useful in the long term prediction of the effect of formulated maintenance strategies on the deterioration of road pavements. The model is also suitable for long term maintenance budget planning.
The HDM-4 model analyses on the case road have shown that the section Malandizi – Chalinze is in comparative poor condition while that of Chalinze – Morogoro is in good condition. The optimum maintenance strategies obtained from the model analysis were SBST, AC50, AC50, AC50 and SBST for the sections Ubungo – Kimara, Kimara – Dar/Coast, Dar/Coast – Mlandizi, Mlandizi – Chalinze and Chalinze – Morogoro respectively. Although HDM-4 economic analyses resulted into these optimum strategies on the case road, some of the strategies can practically not be applied in some of the sections. For example the optimum maintenance strategy AC50 might practically not a feasible option for the sections Kimara – Dar/Coast, Dar/Coast - Mlandizi and Mlandizi – Chalinze. While there is a risk of early crack reflection to the new layers when the option is applied on sections Kimara – Dar/Coast and Dar/Coast - Mlandizi due to the observed reflection cracks, it is not practical to apply the option on the section Mlandizi – Chalinze due to the observed typical defects on the bituminous layer.
It has been shown on this study that the vehicle operatic costs are much higher as compared to the road agency costs. This is explained by the fact that the formulated maintenance strategies’ intervention levels especially for the asphalt overlay and pavement reconstruction were set at higher roughness values (i.e. when the pavement reaches severe condition). It is however believed that for an important road like this, it is not acceptable to allow the pavement to deteriorate to such severe conditions. The road agency costs formulated from the optimum maintenance strategies can be used for long term maintenance budget planning.
The sensitivity analysis has shown as expected that different discount rates result into different NPV thus influencing the decision making in road investment. The sensitivity analysis on the impact of heavy goods vehicle overloading showed un-expected little difference in pavement deterioration. A shift of only one year for proposed maintenance strategy was observed when the ESAL value for heavy goods vehicle was altered from the proposed value of 7.1 to a value of 10. It is
124 believed however based on the principles of engineering and researches that the effect of overloading is more than what is shown by the model results.
It has been observed during this study however that the model requires extensive amount of input data. To get reliable results regular strategic data collection is required. Calibration and adaptation of the model to suit the local conditions must also be done.
The study has revealed that although the road authorities apply the HDM-4 model for planning and management of maintenance activities, the calibration and adaptation of the model to suit the local conditions is yet to be done. The HDM-4 output results must therefore be treated with great care as they do not present the actual real situation. The data collected in this study have shown that some of the important input data required for the prediction of the pavement deterioration were not collected. These data include but are not limited to skid resistance data, texture depth and edge break. The omission of these data is believed to have an impact on the output results of the model analysis.
It has been revealed during this study that the road authorities are in process of bridging some of the mentioned shortfalls. A study on vehicle operating costs has been done together with preparation of the HDM-4 compatible data collection manual.
Application of the mechanistic pavement design analysis method
The South African mechanistic pavement design analysis method was applied to estimate the pavement life for the sections Mlandizi – Chalinze and Chalinze – Morogoro, which were observed to be in poor and good conditions respectively by both visual road condition survey and HDM-4 model analyses. The mechanistic analysis was also used to propose the maintenance strategy for the section Mlandizi - Chalinze.
The analysis results have shown that the cement bound subbase course layer must have been failed completely in the section Mlandizi – Chalinze. The analysis showed that the failure must have occurred when the bound base reached the equivalent granular base phase. The analysis showed no asphalt fatigue problem. However, the visual road condition survey observations revealed several asphalt damages that include fatigue cracks. The results have shown also that the asphalt layer for the section Chalinze – Morogoro must have been degraded. This is also in contrary to what has been observed during the road condition survey. No fatigue damages have been observed during the condition survey. These mismatches are believed to be caused by the application of the method not developed for the Tanzanian conditions. The SAMDM transfer functions have
125 been developed for use in South Africa where the conditions are quite different from those in Tanzania. The analytical calculations have also assumed materials characteristics which are believed to be different from the characteristics of the materials applied, thus causing this contrary.
The mechanistic analyses have shown that 40 mm asphalt overlay can sustain the standard axle loads for the design period of 10 years. A pavement reconstruction by milling off the existing asphalt base layer and surface layer and replacing them with new asphalt layers of thicknesses 210 mm and 50 mm respectively was also found a favourable maintenance option for another 20 years pavement life on this section. In both cases however the analyses did not consider the heavily overloaded axles due to lack on data with respect to occurring numbers of these loads. It is believed however that the proposed structures can not sustain these heavily axle loads with same design lives.
6.2 Recommendations
Based on the conclusions made on section 6.1 above, the following are the recommendations:
Road Maintenance Financing
It is obvious that road infrastructure is a key to the economic growth and poverty reduction. To deal with the challenges regarding the prevailing poor road condition and the prevailing road maintenance financing gap, it is recommended that the government shall continue to prioritize the road sector and continue to invest in road maintenance to close the gap. Apart from improving the government budgetary allocation which depend heavily on fuel levy and solicit funds from donor agencies, the possibility of involving the private sector in road maintenance through the Public Private Partnership (PPP) arrangements shall be taken into consideration. However, research on the suitable PPP model for Tanzanian conditions has to be done before.
Poor construction quality
This is another big challenge ahead of the government and road authorities. Use of scarce maintenance funds will not be efficient if poor construction quality prevail. It is recommended therefore that:
The government shall continue to invest in capacity building training programmes for local contractors and construction industry personnel through relevant institutions. Low contractors capacity and low supervision capacity have large contribution to the poor construction quality.
126 Inadequacy of essential plants and equipments for road works also play a role on poor construction quality. The government shall make conducive environment for investors to invest in plants and equipments for road works.
Awarding contracts based on the lowest priced tender and not the lowest evaluated tender contributes also to the poor quality of works. In many cases the lowest priced tenderer will temper with the quality of the works to avoid the loss from the project. Adoption of long-term performance based road management and maintenance contracts is preferred option. However, the typical contracts are relevant only when there is a capacity within the contractors and the policy and legal framework are put in place.
The government shall continue to conduct technical auditing of construction projects and implement the auditing recommendations.
Vehicle overloading
Vehicle overloading accelerates the destruction of the road infrastructure. This was identified to be also one of the biggest challenges the road authorities have. It is recommended in this study that:
Axle load control program shall be enhanced and routine axle loads measurements in the road network shall be done. Emphasize shall also be given to the standard quality of vehicles operating in the road network.
Because the rate of vehicle overloading has not significantly slow down despite the amendments of several regulations. A study has to be done to clearly identify what exactly is behind this problem. If the problem is identified i.e. difficulties in enforcements, finance or corruption, then the relevant authorities shall play their role to combat the problem.
Application of HDM-4 model
It is recommended that Tanroads shall invest in the calibration and adaptation of the HDM-4 model for the effective use of the model in planning and management of the road network. Once calibrated, HDM-4 model can reliably be used to predict the pavement condition over time when applying different maintenance strategies. It can also be used in estimating road maintenance and improvement budget allocations. The authority must also keep updates in data collection. The VOC used in this study are based on the 2004 VOC study. It is believed that there is a significant difference between the prices and costs between then and now (2009). It is also recommended that all data required for HDM-4 input are collected as their omission has an impact on the model results.
127 Application of the mechanistic pavement design analysis method
Although the South African Mechanistic Design Method (SAMDM) is proposed by the Tanzanian Pavement and Materials Design Manual, it has to be noted that it is too optimistic to make assumptions that there exist similar conditions between the two countries. It has been explained in this study that the transfer functions in SAMDM have been derived based on purely South African conditions that include extensive calibration of the analysis method against the experience of the road engineers from different road authorities in South Africa. For the time being this study therefore recommends the use of the existing empirically derived design method provided by the Tanzanian Pavement and Materials Design Manual. The manual has a particular reference to the prevailing conditions in Tanzania and is based on the more than 30 years experience on the road activities by the then Ministry of Works (MoW). However, as part of reform in road sector, the government shall invest in research that can lead to the development of the mechanistic design methods based on the Tanzanian conditions.
128 REFERENCES:
1. African Development Bank (AfDB). Upgrading of Singida – Babati – Minjingu Road Appraisal Report, Dar es Salaam, Tanzania. May 2007.
2. African Development Bank Group (2004). TANZAM Highway Rehabilitation: Project Performance Evaluation Report, March 2004.
3. Christopher R. Bennett and William D.O.Paterson (2000). A Guide to Calibration and Adaptation. HDM-4 Series volume 5. The World Road Association (PIARC), Paris, France.
4. David C. Wightman, Eric E. Stannard and Jo M. Dakin. Software User Guide. HDM-4 Series volume 3. The World Road Association (PIARC), Paris, France.
5. Haule, J.O: Experience of the Road Fund Board in Financing the Road Network in Tanzania. Annual Road Convention (ARC) 2003, Dar es Salaam, Tanzania.
6. International Road Maintenance Handbook. Practical Guidelines for Rural Road Maintenance, Volume III of IV – Road Works. Transport Research Laboratory, 1994.
7. Kasuwi, A.O and Mamiro, J.G. Management of Public Funded Construction Projects (2001-2005). A Case Study presented at Construction Industry Forum, Dar es Salaam, 2005.
8. Kerali, H.G.R. Application Guide. HDM-4 Series volume 2. The World Road Association (PIARC), Paris, France.
9. Kerali, H.G.R. Overview of HDM-4. HDM-4 Series volume 2. The World Road Association (PIARC), Paris, France.
10. Kozack, J. Considerations in the Choice of the Appropriate Discount Rate for Evaluating Sovereign Debt Restructuring. IMF Policy Discussion Paper, December, 2005.
11. Kruger, P, Wepener, D.A and WJ Botha. An Analysis of Financing of Road Maintenance Based on a Cross Section of African Countries. Annual Road Convention (ARC) 2003, Dar es Salaam, Tanzania.
129 12. Lavin, P.G. Asphalt Pavements: A Guide to Design, Production and Maintenance for Engineers and Architects. First Publications (2003), Spon Press, London.
13. Marmo, F et al. Application of HDM Technology and International good Practice in Improving Long Term Maintenance Planning in Tanzania. Annual Road Convention (2002), Dar es Salaam, Tanzania. November 2002.
14. Meskellie, S.A. Evaluation of Structural Performance, Deterioration and Optimum Maintenance Strategies for Asphalt Roads. MsC Thesis, Delft University of Technology, 2004.
15. Ministry of Works (1999). Pavement and Materials Design Manual.
16. Molenaar, A.A.A. Flexible Pavements Design. Lecture Notes for CT4860, Delft University of Technology, February 2008.
17. Molenaar, A.A.A. Structural Evaluation and Strengthening of Flexible Pavements using Deflection Measurements and Visual Condition Surveys. Lecture Notes for CT4860 Structural Design of Pavements, Delft University of Technology, 2006.
18. Molenveld, J. Prioritization of Maintenance and Rehabilitation Works in Namibia. A Practical Application of HDM-4. Annual Road Convention (2002), Dar es Salaam, November, 2002.
19. National Construction Council (NCC). Lecture Notes on Pavement Distress Identification, Morogoro, December, 2002.
20. Nordengen, P.A. The Problem of Heavy Vehicle Overloading: Can We Save Our Roads? Annual Roads Convention (2002), Dar es Salaam, November, 2002.
21. Odoki, J.B. and Kerali, H.G.R. Analytical Framework and Model Descriptions. HDM-4 Series volume 4. The World Road Association (PIARC), Paris, France.
22. Overby, C. and M. Besta: Overloaded Vehicles, Is that a Problem for the Road Pavements in Tanzania? Annual Road Convention (ARC) 1999, Dar es Salaam, Tanzania.
130 23. Olekambainei, A.E. The Influence of Moulding Moisture Content on the Engineering Properties of Aggregate – Lime – Natural Pozzolan Mixes. Master of Engineering Thesis, University of Pretoria, 2005.
24. Overseas Road Note 1. Road Maintenance Management for District Engineers. Transport Research Laboratory, Third Edition, 2003.
25. Overseas Road Note 18. A Guide to the Pavement Evaluation and Maintenance of Bitumen-Surfaced Roads in Tropical and Sub-Tropical Countries. Transport Research Laboratory, First Edition, 1999.
26. Road Fund Board (RFB). Annual Report for Financial Year 2008/2009
27. Robinson, R and Thagesen, B. Road Engineering for Development – Second Edition (2004), Spon Press, London.
28. Robinson, R, Danielson, U and Snaith, M. Road Maintenance Management: Concepts and Systems (1998), Macmillan Press Ltd, London.
29. Rodrigo Archondo Callao. Applying the HDM-4 Model to Strategic Planning of Road Works. The World Bank Group Transport Papers, September, 2008.
30. Rodrigo Archondo Callao. HDM-4 version 2.05, World Bank Case Studies Version 1 Documentation. The World Bank Group Transport Papers, September, 2008.
31. Rwebangira, T and Mchomvu Y.E. The Assessment of Application of the Highway Development and Management Tools (HDM-4 Model) for the Highway Investment Appraisal Projects in Tanzania. Annual Road Convention (ARC) 2003, Dar es Salaam, Tanzania.
32. Tanroads (2004). Data Collection Manual Road Mentor version 5. June, 2005.
33. Tanroads (2008). Quarterly Report for Financial Year 2008/2009 (31st December, 2008), Dar es Salaam, Tanzania.
34. Tanroads (2005). Reforms in Financing Roads Maintenance: Achievements, Challenges and Strategies. 5th Years Anniversary, Dar es Salaam, October, 2005.
35. Tanroads (2001). Manual for Weighbridge Operations, Final Version. May, 2001.
131
36. Tanroads (2004). Vehicle Operating Costs Study. Final Report. July, 2004.
37. Theyse, H.L, M de Beer and F.C. Rust. Overview of the South African Mechanistic Pavement Design Analysis Method. 75th Annual Transport Research Board Meeting, January, 1996, Washington, DC.
38. Verhaeghe,R.J, de Boer, E and Sanders, F. Infrastructure Planning. Lecture Notes for CT4701, Delft University of Technology, 2006.
39. Verhaeghe, R.J. Plan and Project Evaluation. Lecture Notes for CT4740, Delft University of Technology, 2007.
40. William D.O. Paterson. Road Deterioration and Maintenance Effects – Models for Planning and Management. A world Bank Publication. The John Hopkins University Press, Baltimore and London, December, 1987.
132
APPENDICES
133 Appendix 1:
Materials Requirements as per Tanzanian Pavement and Materials Design Manual
134 Material Requirements for G80 and G60 material classes
Material Requirements for G45 and G25 material classes
135 Material Requirements for CRR and CRS material classes
Materials Requirements for C1, C2 and CM material classes
136 Material requirements for Dense Bitumen Macadam
Material requirements for LAMBS material class
*) The target grading curve is derived from the formula given below.
137 P = (100 – F)*(dn – 0.075n) / (Dn – 0.075n) + F
Where: P = percentage passing sieve size d [mm] D = maximum particle size [dmax] F = Filler content n = a parameter describing the shape of the grading curve
Material requirements for Penetration macadam
138 Mix Requirements for Asphalt Concrete (AC)
139 Appendix 2:
Pavement Catalogues as per Tanzanian Pavement and Materials Design Manual
140 Pavements with granular base course – dry or moderate climatic zones
Base Course Type: Climate Zones: Granular Dry/Moderate
141 Pavements with granular base course – wet climatic zones
Base Course Type: Climate Zones: Granular Wet
142 Pavements with cemented base course for all climatic zones
Base Course Type: Climate Zones: Cemented All
143 Pavements with a bituminous mix in the base course for all climatic zones
Base Course Type: Climate Zones: Bituminous Mix All
144 Pavements with penetration macadam the base course for all climatic zones
Base Course Type: Climate Zones: Penetration Macadam All
145 Appendix 3:
Summary of Road Condition Data
146 Appendix 4:
HDM-4 Input and Output Results
147