Bus and coach secondary safety: Final report

Prepared for Vehicle Standards and Engineering Division, Department for Transport, Local Government and the Regions

G J L Lawrence

TRL Report TRL520 First Published 2001 ISSN 0968-4107 Copyright TRL Limited 2001.

This report has been produced by TRL Limited, under/as part of a contract placed by the Department for Transport, Local Government and the Regions. Any views expressed in it are not necessarily those of the Department.

TRL is committed to optimising energy efficiency, reducing waste and promoting recycling and re-use. In support of these environmental goals, this report has been printed on recycled paper, comprising 100% post-consumer waste, manufactured using a TCF (totally chlorine free) process. CONTENTS

Page

Executive Summary 1

1 Introduction 3 1.1 Retrofitted seatbelts 3

2 Analysis of accidents to aid selection of a crash pulse 3 2.1 Method (accident analysis) 3 2.2 Results (accident analysis) 3 2.2.1 UK National Accident 3 2.2.2 Results of Analyses of UK Police Fatal Accident 4

3 Crash pulse for sled testing of minibuses 5 3.1 Crash test data 5 3.2 Selection of a crash pulse 5 3.2.1 Discussion (crash pulse) 5

4 Selection of representative minibus shells for sled testing 6

5 Testing of current and improved minibus shells 6 5.1 Method 6

6 Description of the standard minibus shells tested 7 6.1 The line built or factory made minibus 7 6.2 The van conversion minibuses 7 6.2.1 The first van conversion 7 6.2.2 The second van conversion 7 6.2.3 The third van conversion 7 6.2.4 The fourth van conversion 8 6.3 The coach built minibuses 8 6.3.1 The first coach-built minibus 8 6.3.2 The second coach built minibus 8

7 Results of tests to standard minibus shells 8 7.1 Results and discussion of the tests to the standard minibus shells 8 7.1.1 The standard line built minibus 8 7.1.2 The first standard van conversion 9 7.1.3 The second standard van conversion 9 7.1.4 The third standard van conversion 9 7.1.5 The fourth standard van conversion 9 7.1.6 The first standard coach built minibus 9 7.1.7 The second standard coach built minibus 10

iii Page

8 Improvements of minibus shells and seats 10 8.1 Method 10 8.2 Modifications to the seats, seat legs and seat side mounts 10 8.3 Results and discussion of sled tests of improved seats 11 8.4 Modifications to improve the van conversions 11 8.4.1 The first van conversion selected for improvement 11 8.4.2 The second van conversion selected for improvement 12 8.4.3 The third improved van conversion 12 8.5 Modifications to improve the coach-built shell 12

9 Results and discussion of tests to improved minibus shells 13

10 Implications for improved regulations 14 10.1 Adaptation of current seatbelt anchorage requirements for minibuses 14 10.2 Sled tests of the complete shell using the ECE Regulation 44 crash pulse 15 10.3 The Australian coach floor/seat regulation 15 10.4 Other considerations for an improved minibus seatbelt anchorage test 15 10.4.1 A further cause for concern 15

11 Considerations for additional or future regulations 16

12 Conclusions 16

13 References 17

14 Acknowledgements 17

Appendix A: Tables 18

Appendix B: Figures 25

Abstract 37

Related publications 37

iv Executive Summary

When accidents occur involving buses, minibuses or minibuses construction, in accident situations which coaches they often involve a relatively large number of currently result in serious and fatal injuries. It was people in a single incident and this gives rise to public concluded from this that it was possible to make vehicles, concern. Despite these types of vehicles being some of the seats and seatbelts systems to withstand the test conditions safest forms of road transport in the UK, publicity given to used here. Because the methods used to improve the seats accidents involving them has led to demands for safety and vehicles were comparatively inexpensive and used measures, notably the installation of seatbelts in coaches conventional technologies, it was also considered and minibuses. This demand has been met to some extent reasonable to require minibuses to meet higher standards in the UK by the requirement for all children, on journeys in future. Therefore it is recommended that seatbelt relating to school or other child activities, to be transported anchorage regulations for these vehicles be revised to in vehicles fitted with seatbelts. Further UK legislation will require a higher standard. Three possible regulatory require all new buses and coaches, apart from those methods to require higher standards are discussed: specifically designed for urban use and standing l Adaptation of current seatbelt anchorage pull test passengers, to be fully equipped with seatbelts. However, requirements to require minibuses to have the same the requirements for seatbelt anchorage strengths for anchorage strength as required for cars (ECE Regulation minibuses and coaches are less demanding than for cars. 14). This may be the simplest measure to introduce and This projected was intended to: would make use of existing test equipment. l determine if current seatbelt anchorage requirements for l Sled tests of the complete minibus shell using the ECE minibuses are appropriate; Regulation 44 crash pulse. This would require a high l assess the performance of current designs of minibus capacity sled and the results may be more difficult to seatbelt anchorages when subjected to typical ‘serious’ interpret. accident loadings; l Adaptation of the Australian coach floor/seat regulation l demonstrate if it is practical to achieve higher standards ADR68 to require a sled test of rows of seats using test of occupant retention than currently required. dummies combined with a matching pull test of the seat mountings in the vehicle floor. This would require a The first stage of this study was to determine from high capacity sled and could be used to test for double accident and test data the impact severity in terms of loading. However, seperat testing of seat and floor has vehicle velocity change and acceleration, of minibus some disadvantages. accidents that typically result in serious and fatal injury. It was found from this analysis that these serious accidents These three methods used as a regulation would all typically produced loadings that were about twice as provide at least double the anchorage strength currently demanding as the current regulations for minibus seatbelt required for minibuses. anchorages, and even more demanding if seats with The methods used to improve the minibus seats and restrained occupants in them were also impacted from vehicle structure have been described and these would behind by unrestrained occupants (double loading). assist manufactures in producing designs to provide higher Following this a sled test method with minibus shells levels of occupant retention. was developed to replicate this serious accident situation. Other safety concerns have been identified and solutions This method was used to test the performance of seven proposed. Possible additional or future regulations have different minibuses minibus with seatbelt anchorages to been discussed along with considerations thought necessary current requirements to determine their efficacy in that for improved regulations to take account of the construction accident situation. These vehicles tested represented the methods and scale of production of minibuses. three main types of construction used to make minibuses. Minibus seatbelt anchorages are normally built into the seat and this was the case with all the vehicles selected for testing. It was found that the seats or the seats to floor fixings of all but one of the vehicles tested failed to withstand double loading and some suffered failures with just the seatbelt loading from their restrained occupants. The one exception, which suffered no significant failures, was a minibus that had been built with anchorages of at least twice the minimum strength requirement for minibuses anchorages. Next, the seats, seatbelt anchorages and floor designs of some of these vehicles were improved where necessary and tested to determine and demonstrate whether it would be practical to provide protection, with all types of

1 2 1 Introduction ECE Regulation 14 will provide useful occupant protection in less severe accidents and are unlikely to make the Buses, minibuses and coaches are some of the safest forms situation worse in more severe accidents. Because of the of road transport in the UK. However, when accidents do problems in providing and testing full strength retrofitted occur they can involve a relatively large number of people seat-belt anchorages, and the benefits of seatbelt in a single incident and this gives rise to public concern. anchorages of reasonable but non-conforming strength, it Publicity given to accidents involving these vehicles has was decided to take a more pragmatic view on the testing led to demand for safety measures, notably the installation requirements for retrofitted vehicles. The intention was to of seatbelts in coaches and minibuses. This demand has require a reasonable standard of installation, which may been met to some extent in the UK by the requirement for not always meet the Regulation’s strength requirement. all children, on journeys relating to school or other child This was achieved by subjecting the retrofitted seatbelts to activities, to be transported in vehicles fitted with seatbelts. a visual inspection. These visual inspections have been Further UK legislation will require all new buses and carried out from August 1998 by Vehicle Inspectorate staff coaches, apart from those specifically designed for urban during the vehicle’s first annual MOT test following the use and standing passengers, to be fully equipped with installation of the seatbelts. seatbelts. However, the requirements for seatbelt To provide information to assist with this retrofitting anchorage strengths in these larger vehicles are less and inspection task a study was carried out as part of this demanding than for cars. The project was intended to project. Details of this task were reported by Lawrence and determine if current regulations for minibus seatbelts B J Hardy (1997) and are not covered in this report. anchorages are appropriate, and if not to propose improved standards. The first stage of this study was to determine from accident data the impact severity in terms of vehicle 2 Analysis of accidents to aid selection of velocity change and acceleration, of minibus accidents that a crash pulse typically result in serious and fatal injury. Following this the performance of current minibus seatbelt systems was 2.1 Method (accident analysis) reviewed to determine their efficacy in that accident situation. Next, these minibuses seatbelt systems were To obtain an overall understanding of the accident situation, improved where necessary and tested to determine and a preliminary review of available data pertaining to coach, demonstrate whether it is practical to provide protection in bus and minibus accidents was undertaken by Wall (1995). accident situations which currently result in serious and Two of the data sources identified were considered to be fatal injuries. In addition, these results have been suitable for use as a basis for deciding on the test conditions considered in order to provide guidance for the proposal of appropriate for occupant protection provision in minibuses. test methods that would result in higher standards of The first, known as STATS 19 consists of computerised data occupant protection through improved floor, seat and provided by the Police on all injury road accidents occurring seatbelt anchorage performance. in Great Britain. When the Police report a road accident involving personal injury they fill in the form for STATS 19 but in the case of an accident involving a fatality they also 1.1 Retrofitted seatbelts compile a full report of the accident including witness Legislation came into effect in 1997 - 1998 to require statements, coroners’ reports, results of Police accident coaches and minibuses to be equipped with seatbelts when investigations and often photographs. Fatal accident files, carrying children on organised trips. This particularly from co-operating Police Forces around the UK, forms the affected vehicles used to transport children to and from second database of road accidents. These databases were school, or on school trips. Many vehicles then in service interrogated to provide information on typical minibus were not originally fitted with, nor designed to provide, accident situations that would be most demanding for the seatbelt anchorages. To comply with this new requirement, seat and seatbelt system. Full-scale crash test data were many of these vehicles were retrofitted with seat belts. found which best matched this accident situation and a The method used in ECE Regulation 14 for approving simplified crash pulse was derived that would be suitable for seatbelt anchorages on a new vehicle model consists of a use in sled testing of minibus shells. This work is described destructive pull test, on a representative sample vehicle. In in more detail in the second interim report of this project many cases, particularly for vehicles not originally (Lawrence et al., 1996). designed for seatbelts, it would have been difficult or very expensive to make all retrofitted seatbelt anchorages 2.2 Results (accident analysis) comply with the requirements of ECE Regulation 14. The existing vehicle fleet used to transport children was 2.2.1 UK National Accident thought to consist of many different designs with UK National Accident data were used to find the most comparatively small numbers of most types. Taking into frequent types of minibus accident. Each accident was account the destructive nature of the pull test, it was classified by the most serious minibus occupant injury. concluded that it was not practical or economical to assess The analyses showed that the most common minibus retrofitted seatbelt anchorages to Regulation 14 standards. accidents resulting in passenger injuries were from impacts However, it is generally considered that seatbelt with other vehicles accounting for 75% of all injury anchorages of a lower strength than that specified in minibus accidents. Of these accidents with other vehicles,

3 61% were due to the minibus colliding with a single identified. Out of these 13 cases, 6 cases (46%) involved the vehicle. Of these accidents, impact with car type vehicles full width of the front of the minibus and a further 5 cases were the most common and accounted for 72% of injury (38%) involved at least half the width. The remaining 2 cases accidents involving one vehicle. (15%) involved about one third of the minibus front. It can However, accidents resulting in serious or fatal injuries therefore be concluded that a large overlap was the most to minibus occupants are thought to be more appropriate frequent accident situation. Also out of the total of 51 frontal when considering seatbelt strength requirements. Again minibus accidents identified, it was possible to make a ‘best minibus accidents with other vehicles were found to be the estimate’ of the speed of each vehicle at contact and the mass most frequent accounting for 59% of all fatal or serious of the vehicles involved in 11 cases. From this an estimate of injuries accidents. Of these accidents with other vehicles, the change in minibus speed at first contact (delta V) was 57% were due to the minibus colliding with a single calculated, using the equation of conservation of momentum vehicle. Of these accidents, impact with car type vehicles and assuming no rebound. Details of the injuries sustained by were the most common and accounted for 67% of fatal or the minibus occupants were also available for these cases. It serious injury accidents involving one other vehicle. In should be noted that an accident is classified by the police as these accidents with single car type vehicles the front of fatal if one of the occupants in either/any vehicle involved or the minibus was the first point of contact in 64% of a person outside the vehicle is killed. Therefore not all fatal impacts with car types. Of these accidents involving the minibus accidents cases found involve a fatal minibus front of the minibus, 68% were with the front of the car. occupant injury. Fatal or serious minibus accidents with heavy vehicles These data have been used to plot the cumulative (trucks and buses) were less common (19% of all fatal or percentage of minibus accident according to the most serious serious accidents) but resulted in a higher injury rate. The occupant injury severity against the change in minibus speed proportion of minibus accidents that resulted in fatal or at first contact (delta V), see Figure 1. serious occupant injuries was 29% for the heavy vehicles, compared with 6% for the car type vehicles. The minibus front was most often involved in accidents fatal or serious 100 accidents with heavy vehicles (59% front, 25% side and 16% rear of the minibus). Of these fatal or serious minibus front accidents, 35% were to the front, 27% were to the 80 side and 38% to the rear of the heavy vehicle. The probability of fatal or serious injuries was found to 60 be higher if the minibus overturned. For the impacts with the car type vehicles, the proportion of minibus accidents in the fatal or serious category is 31% for the minibus 40 overturned compared with 6% for non-overturned (the numbers of overturned accidents with heavy vehicles was 20 too small to make a similar comparison). As no information is given in this database on whether ejection Cumulative percentage of minibus accidents occurred as a result of overturning it is not possible to 0 know the exact cause of the injuries. However, it is likely 020406080100 that the wearing of seatbelts would reduce the risk of Delta V (km/h) injuries within the vehicle and the risk of ejection. Ejection greatly increases the risk of serious or fatal injuries for Fatal Injury Serious Injury Fatal or Serious occupants. Nevertheless, the frequency of minibus Accidents Accidents Injury Accidents overturning accidents is low at 19% of all fatal or serious injury minibus accidents. Figure 1 Cumulative percentage of accident severity by Speed limits on the roads where the fatal or serious change in minibus accidents occurred were generally above 40 miles/hr. For accidents with the car type vehicles this speed limit accounts From this figure a test speed can be selected for a sled test for 56% of fatal or serious minibus accidents while for the of minibuses or minibus components to cover a chosen heavy vehicle type the figure is higher at 77%. However, the proportion of accidents. The figure shows that a test speed of less serious accidents are more common on roads with lower about 48 km/h would cover at least 50% of all accidents in speed limits. The results of this analysis are given in more which a minibus occupant was fatally or seriously injured. detail in Tables 1 to 6, Appendix A. Severe impacts with heavy vehicles were found to result in a large extent of crush within the minibus, causing 2.2.2 Results of Analyses of UK Police Fatal Accident serious injury to the occupants within these areas. Any Results of Analyses of UK Police Fatal Accident Files also normal style restraint system is unlikely to be of showed that most of the impacts with other vehicles were worthwhile benefit to the minibus occupants within the frontal impacts, 51 cases out of a total of 74 cases (69%). In crushed cell. However, the restraint system would still this database it was also possible to find the degree of overlap benefit occupants in positions within the area of the in 13 cases out of a total of 51 frontal minibus accidents minibus that was not crushed.

4 Rollover accidents often result in the ejection of 10 occupants, who thereby suffer fatal or serious injury. A 5 seatbelt system would help to retain the occupants within 0 the vehicle during rollover. However, a seatbelt test based -5 on rollover requirements would be less demanding than -10 one for frontal impacts. It will, however, be necessary for -15 the sensors of any emergency locking retractors to operate under an overturning mode. -20 Although it might be expected that accidents involving -25 heavy vehicles would result in larger values of minibus Acceleration (g) -30 delta V, due to the heavy vehicles having a larger mass, -35 this was not found to be the case in the fatal accident -40 database. The values of delta V for accidents with heavy -45 vehicles were found to be of a similar magnitude to those -50 of car type impacts because the heavy vehicle accidents 0 20 40 60 80 100 120 occurred at lower speeds. This finding was despite the UK Time (msec) national statistics showing that accidents with heavy vehicle most frequently occurred on roads with higher Figure 2 Full scale crash test results overlaid with ECE speed limits. The results of this analysis are given in more Regulation 44 crash pulse detail in Tables 7 to 10, Appendix A. The Police Files were first selected automatically using into the seat. Consequently it is important that a crash the STATS 19 definition for minibuses type vehicles, pulse is carefully selected to provide the optimum which also includes camper vans. Table 7 shows the reduction of injuries without being over demanding. A test number of minibuses and camper van accidents in the for minibus seatbelt systems should be based on the police files selected. This gives some indication of the loadings that occur in accidents. However, it would be proportion of camper vans accidents that are included in unreasonable to expect them to withstand the most severe the STATS 19 statistics given in the previous tables accidents. Indeed, anchorage of excessive strengths would (Tables1 to 6, Appendix A). It was not possible to remove provide little or no additional protection because in very camper vans from the analysis of STATS 19 data. violent accidents they would induce belt loadings in excess of the strength of the human frame. In these cases some additional protection may be provided through 3 Crash pulse for sled testing of deformation of the seatbelt system on overload (e.g. minibuses bending of the seat back). Because of the above considerations it is concluded that a target of a seatbelt 3.1 Crash test data system strong enough to withstand the loadings in a One full-scale minibus crash test was found which minimum of 50% of fatal and serious injury accidents, matched the most common impact condition, identified would be reasonable. As has already been discussed, this above, of 100% overlap. This was a Ford Transit minibus target would result in a sled test speed (delta V) of about crash test conducted at TRL. The vehicle, a 1983, 12 seat 48 km/h. However, it should be noted that the data are Ford Transit (line built), was subjected to a perpendicular drawn from a sample of fatal accidents. Because of this it impact into a large concrete block, with full overlap at a is likely that a number of lower speed accidents resulting speed of 49.5 km/h (30.9 mph). Figure 2 shows the time in minor or serious injuries are not included. The bias from history for that test from the fore and aft accelerometer this means that a test velocity selected from this graph is mounted on the B post. likely to represent in real life a larger proportion of all fatal The change in velocity seen by the minibus during the and serious minibus accidents. There were no accident data main phase of this impact (delta V) is equal to the test available for minibus accidents that would enable this velocity. degree of bias to be estimated. Therefore it should be noted that the estimate of 50 percent of the severe minibus accidents included in a 48 km/h test has to be considered 3.2 Selection of a crash pulse as conservative. 3.2.1 Discussion (crash pulse) The accident data discussed above indicate that the most The two main properties needed to specify a crash pulse common impact is minibus to car, front to front, with 100 for a sled test are the test velocity and the sled acceleration per cent overlap. However, accidents with heavy vehicles, time-history. The velocity can be selected from the although less common, result in a higher injury severity. accident data and the acceleration should be one typical for The impact conditions in the TRL full-scale Ford Transit all types of minibus. test would be broadly similar to the most common minibus It is difficult to provide strong seatbelt fixings in front to car front, or minibus front to front or rear of heavy minibuses because, unlike cars, for some seats there is no vehicle accidents with a delta V of about 49 km/h. adjacent vehicle structure to which they can be attached. However, it will be slightly more demanding because the Therefore, minibus seatbelt anchorages are normally built concrete block would have caused slightly higher minibus

5 accelerations due to it forcing all the stiffest frontal onto the vehicle chassis, not pre-assembled and then structures to deform. In a minibus to other vehicle impact mounted to the chassis. Typically, longitudinal galvanised these stiff structures are likely to deform slightly less steel members were attached to the top of the vehicle because they are likely to penetrate the other vehicle. main chassis members. Cross members were then fixed to Taking these considerations into account it was concluded these and a side structure attached in turn. The vehicles that the acceleration time history from this full-scale test were clad in either stainless steel sheet or aluminium. would be suitable for developing a crash pulse. Ideally it Seats or tracking fixings were normally located on the would be preferable to test a range of different makes of longitudinal and cross members of the floor. minibus against cars and heavy vehicles to select an Minibus body shells were obtained which represented average or a worst case acceleration time history. the three types of construction. Vehicles were chosen from However, taking into account the accuracy with which the the larger manufacturers or converters who were willing to delta V value could be established from accident data the co-operate in the research programme. A total of seven results from this one vehicle model are thought to be minibus shells were purchased for the initial phase of acceptable. Therefore, the sled crash pulse for tests of testing current standards of construction; one line built minibus shells in this project was based around the from manufacturer A, two coach built from manufacturers acceleration performance of the Ford Transit to concrete F and G and four van conversions. The van conversions block tests illustrated in Figure 2. selected were based on three different makes and models For a sled test some simplification of the acceleration pulse of van. Van conversions B and C used the same model of is acceptable to take account of the nature of sled arresting large long-wheel base van as a basis and were converted systems and the frequency response of seatbelt systems. In by the same converting company. Van conversions D and addition, a tolerance is required to make reasonable E were based on different models and converted by allowances for the test to test variations common to sled test different converting companies. Four of these, (one coach systems. Both of these requirements are normally taken into built by company G and three van conversions vans C, D account by providing ‘a crash pulse corridor.’ Sled test and E) were then selected for improvement. With the regulations were examined for guidance on this and it was exception of one of the van conversions (van C), which noted that the ECE Regulation 44 pulse (Economic was improved solely by TRL, the manufacturer or Commission for Europe, 1998) (for testing child restraint converter and seat manufacturers made these systems) has the same test velocity, and an acceleration improvements with various degrees of assistance and history similar to, but slightly lower than, the average level of collaboration from TRL. acceleration found in the Transit test. Figure 2 shows the Regulation 44 crash pulse corridor overlaid on the full-scale crash test result. It was decided to use this corridor for 5 Testing of current and improved minibus sled tests in this research programme. minibus shells

4 Selection of representative minibus 5.1 Method shells for sled testing The front and back of the main longitudinal members of each of the minibus shells were attached to the test sled. A study was made of the methods used to produce These attachments were arranged to cause the minimum minibuses in the UK. It was found that there were three of interference (strengthening) to the floor structure of different methods of construction: the vehicle. i Line built - A standard delivery van is normally used as These shells with six crash test dummies as passengers the base for line built minibuses. However, the extra were then dynamically tested to the ECE Regulation 44 features required for a minibus are built into the vehicle crash pulse. as they are made. These features include purpose made In each of the sled tests performed, a total of six floor reinforcements, seat attachment anchorages and passenger seats were fitted into the minibus shell in two or factory fitted windows. three rows. The seat rows were made up of single, double ii Van conversions - The modifications made to the basic and triple seat units. The exact seat arrangement depended vans by the various converters visited vary widely. In on the seat plan of the minibus concerned. For the vehicles some cases the van roof is removed and replaced by a tested there were three basic seating configurations as raised glass-fibre roof. Window apertures are cut in the shown in Figure 3. side structure of the van and this sometimes requires the In the tests dummies, with or without the seatbelts in removal of stiffening struts. Often a welded steel use, occupied all the seats. As all of these vehicles had the strengthening frame is fitted around the window seatbelt anchorages attached to the seats, these aperture in an attempt to replace the strength removed arrangements produce a combination of seat loading when the aperture was cut. conditions that enabled the maximum information to be iii Coach built vehicles - The coach built vehicles are obtained from each test. These arrangements provided: normally built on the rear of a chassis-cab produced by a i double loading, where the front seats were subjected to major vehicle manufacturer. The coach built vehicles the loading of their restrained occupants and also the examined under construction were all being built directly loading from un-restrained occupants behind;

6 and consisted of heavy steel tubes with substantial diagonal braces and large strong flat plate feet to which the lower seatbelt anchorages were attached. The large flat

1 2 1 2 3 1 2 3 ‘feet’ were bolted to threaded anchorages in the floor pan. A(r) U(r) U(r) A(r) U(r) U(r) A(r) U(r) The floor pan was heavily reinforced with longitudinal and transverse under-floor box members. In addition the sides 3 4 4 5 6 4 5 B(u) U(u) B(u) U(u) C(r) B(u) U(u) of the seat backs that were against the side wall of the vehicle, (numbers 3, 5, and 6), were also attached to the 5 6 6 minibus side rail by means of short straps made from C(r) U(r) C(r) seatbelt webbing. The seatbelts had the reel attached to the 2-2-2 Seating 3-2-1 Seating Alternative 3-2-1 foot of the seat, with a pivoted guide and upper mounting on an ‘extender’ on the side of each seat. Key A & C = Fully instrumented dummy B = Partially instrumented dummy 6.2 The van conversion minibuses U = Un-instrumented dummy 6.2.1 The first van conversion (u) = Un-restrained (r) = Restrained The first van conversion tested (vehicle BVCS) had seats made by supplier β. These seats were bolted in place in a 2-2-2 pattern. The conversion consisted of a wooden floor Figure 3 Plan view of the seating arrangements used in the placed on top of the steel floor of the van and the feet of sled tests the seats were bolted through both the wood and steel floors with square-load spreader plates underneath. Two ii rear loading only, where the middle row of seats initially pairs of legs were used for each double seat unit. had un-restrained occupants and were later loaded by the knees and head of the restrained occupant behind. 6.2.2 The second van conversion The seat plan of the line built vehicle A was such that The second van conversion tested (vehicle CVCS) again the unrestrained seats had no rear loading (alternative had seats made by supplier β. However, the seats were 3-2-1 seating plan); attached to a low-cost tracking system in a 2-2-2 pattern. iii single loading only, where the seat is only loaded by its The conversion consisted of a wooden floor placed on top restrained occupant. However, for all but the line built of the steel floor of the van and the tracking was attached vehicle A with the alternative 3-2-1 seating plan, the to the floor with bolts through both the wood and steel restrained occupant also contacted the back of the seat in floors with penny washers (washers with a large outside front, thereby reducing the loading on the seat and diameter) to spread the load underneath. In addition, restraint to somewhat less than full single loading. wood-screws were used between the bolts to fix the The instrumented dummies used for these tests are as tracking to the wooden floor. Two pairs of legs were used follows: Dummies A and C (both restrained) were Hybrid for each double seat unit. III dummies with head and thorax triaxial accelerometers, chest compression potentiometers and neck and femur 6.2.3 The third van conversion force transducers. Dummy B (un-restrained) was a Hybrid The third van conversion tested (vehicle DVCS) had the II dummy with head and thorax triaxial accelerometers, seats made by manufacturer γ, fitted in a 3-2-1 pattern. The and femur force transducers. front triple seat unit had three pairs of legs, one pair at each end and one pair in the middle. Their front and back feet rested on a sub-frame with bolts through the foot, the 6 Description of the standard minibus sub-frame and the steel floor of the van. The sub-frame shells tested was made of rectangle steel tube and was fitted on top of the steel floor. Load-spreading washers were used beneath All of the vehicles tested had the seatbelts attached to the the floor. The sub-frame was only attached to the floor by seats. the bolts of the seat feet, despite the frame extending some distance beyond the seats. Wooden in-fill was used 6.1 The line built or factory made minibus between the sub-frame members to create a flat floor area.

The factory-produced minibus (vehicle ALBS) was unusual Behind the front row of seats a continuous sheet of wood because the seatbelt anchorages had been designed to meet was placed on top of the steel floor of the van. The middle the more demanding strength requirements for a car (M1) seats and the rear seat had a combination of floor mounted rather than those of a minibus (M2 or M3). The legs, bolted through the wood and steel floor, and side anchorages were therefore at least twice the current mountings on to the vehicle side structure (two pairs of minimum strength requirement for a minibus. The seating legs and a side mount for the middle double seat unit, one arrangement, with some seats removed, was in the pair of legs and a side mount for the single rear seat). The Alternative 3-2-1 configuration, (see Figure 3). The seat, front and rear seat legs were joined by a continuous ‘U’ made by supplier α, had legs that were integral to the seat shaped foot.

7 6.2.4 The fourth van conversion devised which proved effective in similar component tests. Therefore, all the ten joints on the test shell were The fourth van conversion tested (vehicle EVCS) had the seats, made by manufacturer δ, fitted in a 2-2-2 pattern. strengthened using the same method before it was tested. This conversion was intended to be especially robust and the van shell had been heavily reinforced with a full safety cage which was made from rectangular steel tube and 7 Results of tests to standard minibus included roll-over protection. Each double seat unit had shells one pair of floor-mounted legs, on the inboard end, with the feet bolted through a strip of high-strength steel on the The results of the tests of the standard minibus shells are floor. Each side of the strip had been seam welded to the summarised in the first part of Table 11, Appendix A. floor. The outboard side of each double seat unit of seats These results show that all but one of the standard was attached to the side rail of the reinforcement cage at vehicles tested suffered serious failures of the front seats the front and back of the seat. and/or the vehicle structure at the seat attachment points. This is not surprising as the crash pulse and double loading for this seat row was more than twice as 6.3 The coach built minibuses demanding as that currently required for M2 vehicles. 6.3.1 The first coach-built minibus The only standard vehicle not suffering serious failure was the line built minibus, which had been designed to The first coach-built minibus tested (vehicle FCBS) was made from a series of modules joined to make up the exceed current regulations by a large margin. However, selected length. (This modular system enabled the one aspect of the performance of this vehicle gave rise to manufacturer to tailor the length of the coach built section to concern about the construction of the seat back. The seat fit both long and short chassises.) A wooden floor was backs inside the metal frame were filled with only fabric, mounted on top of multiple longitudinal members joined by fibre padding and wire. This provided little resistance or crossbeams. Each longitudinal member consisted of a wide load spreading when the knees of the unrestrained rear and shallow ‘U’ section pressed from stainless steel sheet. occupants struck the back of the front seats. The knees This assembly was in turn attached to longitudinal members penetrated the seat back and tore the ribs from the spine of the chassis. The seats, made by manufacturer β, were of the dummy in the front seat (seat 2). Although this fitted in a 2-2-2 pattern and were bolted through both the dummy was not instrumented to record this type of wooden floor and the underlying longitudinal members. injury, the mechanical parts are strong and this damage Two pairs of legs were used for each double seat unit. indicating that a significant concentrated force was applied to the dummy, indicating a risk of serious or fatal spine injury in a human. This emphasises the importance 6.3.2 The second coach built minibus of wearing seatbelts in the rear of passenger vehicles, The second coach built minibus tested (vehicle GCB ) was S both to protect the rear seat occupant and the occupant made up from a series of cross-members attached to two immediately in front. longitudinal members beneath, which in turn sat on the Many of the outputs from the instrumented dummies were chassis rails of the chassis cab vehicle. The seats, made by within safe limits where the tests were meaningful. Test ε manufacturer , were fitted in a 2-2-2 pattern. Each double results were the seats failed completely, where there were no seat unit had a pair of legs at each end with the feet seats in front or where dummies were ejected are not reported. attached to two lengths of tracking, which ran the full Consequently, only the HIC values from the rear ‘restrained’ length of the shell, one for each end of the seats. The dummies are reported in Table11, Appendix A. tracking assembly consisted of an aluminium extrusion on Before and after photographs of the standard minibus top of a steel ‘top-hat’ shaped section. This assembly had shells are shown in Figures 4 to 10, Appendix B. been designed by the coach builder to transfer the seat loads directly to the top of the floor cross-members. In 7.1 Results and discussion of the tests to the standard addition, the sides of the aluminium track were shaped to minibus shells provided support and location for wooden floor sections used to in-fill between the tracks. 7.1.1 The standard line built minibus It was intended to test the ‘standard’ minibus shells None of the seats or seat to floor connections in this vehicle without any modification that would affect their occupant (vehicle ALBS with alternative 3-2-1 seats) suffered serious protection performance. However, for this minibus the structural failure. method used to attach the cross-members to the The double loaded front triple seat unit and floor showed longitudinal members appeared to be inadequate to most distress with the seat backs pushed forward about 25 withstand the test loads. These joints effectively hold the degrees and the floor lifted around the rear foot fixings and complete coach-built shell to the chassis of the chassis-cab depressed around the front foot fixing. The webbing or in this case to the test sled. TRL considered that there attaching the outer side of the triple seat unit to the wall of was a high risk of the entire coach built shell detaching the minibus (side of seat back 3) had reduced the distance during the dynamic test. Samples of these joints were that end of the seat unit moved forward, causing some twist made and dynamically tested. These tests confirmed that along the row of seats. One of the rear diagonal braces the joint was inadequate. A modification to the joint was between the seat-foot and the seat–pan had also failed in

8 tension. Examination of the floor beneath this seat showed horizontal. The unrestrained dummies behind, made heavy that some of the construction joints had failed allowing one knee contact with the back of the front seats before their of the foot fixing inserts to come close to breaking free. rear feet failed. The unrestrained dummies from the middle No movement of the seats with un-restrained occupants row of seats followed the front seats as they tipped was seen (seats 4 and 5). Seat six, the single seat with a forward, finally landing on top of the front seats. restrained occupant, moved very slightly due to twisting of Following the departure of the unrestrained dummies the floor mountings, but these appeared to have further from the middle seat row, the back of seat five of the reserves of strength. middle row was hit by the knees and then the head of the restrained dummy in the single rear seat behind. The 7.1.2 The first standard van conversion middle seat row did not suffer any failures or any significant permanent distortions from this rear loading. In the test to vehicle BVCS (with 2-2-2 seats), the front feet of the first double seat unit punched through the wood and The impact of the restrained rear seat dummy with the steel floor layers under the loading from its restrained back of the middle row seat was caused by a combination occupants before the unrestrained dummies behind had hit of normal seat belt pay-out, the seat belt upper anchorage it. The seats then tipped further forwards until their backs moving forward as the back of the rear seat flexed and the small space between the seat rows. The HIC from the were almost horizontal. 36 The rear double seat unit with just the loading of their instrumented restrained dummy in rear seat from the restrained dummies distorted at the junction of the back to impact with the back of the middle seat row was 1133. the seat and the front feet dug into the floor slightly and the rear feet pulled the floor up slightly but did not fail. 7.1.5 The fourth standard van conversion

However, their dummies made heavy contact with the In the test of vehicle EVCS (with the 2-2-2 seats) the front backs of the middle row of seats, which would have off- seats initially held but, following knee loading from the loaded the rear seat to some extent. Despite this off- unrestrained dummies behind, the bolt head in the rear foot loading the rear seat back at maximum deflection was of the inboard legs first pulled through the foot followed angled forwards by about 25 degrees past the vertical. closely by failure of the seat’s rear side mount. The side mount failed due to a clamped insert pulling out of a 7.1.3 The second standard van conversion channel section in the seat. The double seat unit then tipped forward about the remaining front foot and front side mount In the test of vehicle CVCS (with the 2-2-2 seats) the feet of the front double seats unit ripped straight out of the which then both failed (the foot breaking off from the leg at tracking at the beginning of the impact. The seats, a weld and the seat side mount clamped insert pulling from complete with their dummies retained in the seats by their the seat channel). This allowed the folded forward seat to seat belts, were ‘carried’ to the netting fitted at the front of slide to the front of the minibus with its dummies still the cut down minibus shell. The front seat and occupants retained in the detached seats by their seat belts. The were closely followed to the front of the shell by the unrestrained dummies behind closely followed the front seat unrestrained dummies from the seat pair behind. as it moved forward, finally landing on top of it. Following the departure of their unrestrained dummies Following the departure of the unrestrained dummies the rear feet of the middle seats, pulled out from the from the middle seat row, the back of the middle row was tracking under a combination of the seat’s inertia and hit by the knees and then the heads of the restrained loading from the knees and heads of the restrained dummies in the double seat unit behind. The middle seat dummies in the rear seats. The seats then tipped forwards row did not suffer any failures or any significant about their front feet stopping when the back had just permanent distortions from this rear loading. passed the horizontal. The impact of the restrained dummies in the rear seats The rear double seat unit with the restrained dummies with the backs of the middle row was caused by a remained fixed to the tracking, but the tracking ripped its combination of normal pay-out of the rear seatbelts, failure retaining screws out from the floor. The seats tipped of the rear side mounting and the small space between the forward with the tracking bending, just in front of the front seat rows. The failure of the rear side mounting was again feet of the seat. The dummies, retained in the seats by their due to the clamped insert pulling from the seat channel and seatbelts, were pressed against the backs of the middle row this allowed the seat to twist forward. The HIC36 from the of seats as both sets of seats tipped forward together, the instrumented restrained dummy in rear seat from the seat stopping when the back had just passed an angle of 45 impact with the back of the middle seat row was 730. degrees to the horizontal. 7.1.6 The first standard coach built minibus 7.1.4 The third standard van conversion As described previously, the floor of vehicle FCBS (with

In the test of vehicle DVCS (with the 3-2-1 seats) the bolt the 2-2-2 seats), was made up of wood on top of multiple, heads, in the rear feet of the front triple seat unit pulled wide and shallow ‘U’ shaped, longitudinal members. In the through the feet allowing the front row of seats to tip test the front feet of the first double seat unit punched forward about their front feet, which bent but did not through the wood and caused local bending down of the detach. The seatbelts retained the dummies in the seats as steel longitudinal member below and the rear feet slightly they tipped forward stopping with their backs nearly pulled up locally the wooden floor and underlying

9 longitudinal member. This was due to both the occupant fixings finally failed and the seat and dummies then slide loading and heavy knee loading from the unrestrained to the front of the shell closely followed by the dummies behind. Following on from this the attachment unrestrained dummies from behind. between the top of the rear legs and the underside of the Following the departure of the unrestrained dummies seat failed completely and the seat pivoted forward about from the middle seat row, the back of the middle row was the joint between the top of the front legs and the hit by the knees and then the heads of the restrained underside of the seat. The seats tipped forward with their dummies in the double seat unit behind. The seat back dummies still retained in the seat belts until the seat backs flexed slightly due to this loading and the feet were were almost horizontal and were closely followed by the slightly bent, however, the damage was minor. unrestrained dummies behind. The back feet of the rear seats, with just the loading of Following the departure of the unrestrained dummies their restrained dummies, bent slightly allowing the legs to from the middle seat row, the back of the middle row was lift slightly off the floor. The rear leg attachment to the seat hit by the knees and then the heads of the restrained base started to fail but did not completely pull through the dummies in the double seat unit behind. Due to this beam forming the rear of the seat base and this beam was loading, the front feet of the middle seats row pushed also buckled. This allowed the complete seat to tip forward down the floor at the front causing slight local deformation and the dummies made heavy contact with the back of the of the wood and underlying longitudinal member and the middle row of seats. Despite this off-loading, the backs of rear feet pulled up the wooden floor and underlying the rear seats, at maximum deflection, were angled longitudinal member a significant amount. This allowed forwards by about 10 degrees past the vertical. It was the middle seats to tip forward so that the back went concluded that the attachments of this pair of seats were slightly past the vertical at maximum displacement. very close to complete failure. The HIC36 from the The feet of the rear seats, with just the loading of their instrumented restrained dummy in rear seat from the dummies, again pushed the wood floor down at the front impact with the back of the middle seat row was 1298. and pulled it up at the rear in a similar manner to that of The modified joints between the floor cross-members the middle row. However, the displacement was slightly and the longitudinal beams bent slightly as the floor more. This allowed the complete seat unit to tip forward cross-members lozenged slightly but showed no signs of and the dummies made heavy contact with the back of the failing. middle row of seats. Despite this off-loading the backs of the rear seats, at maximum deflection, were angled forwards by about 20 degrees past the vertical. The HIC 36 8 Improvements of minibus shells and from the instrumented restrained dummy in rear seat from the impact with the back of the middle seat row was 1356. seats The reason for the different pattern of floor distortion for It was decided to improve and test three of the van the front and back feet, seen between the front and rear seats conversions and one of the coach-built shells. rows, was due to the relative position of cross members which locally strengthened the longitudinal members. The local creasing of the underlying longitudinal floor 8.1 Method members about the front feet of the front seat, and to a lesser Modifications to improve the strength of the seat extent the rear feet of the rear seats, extended to the side attachments and floor reinforcements were devised to member of the vehicle and was extensive. However, the side overcome the deficiencies seen in the current vehicle tests. panels had been omitted from the side frame to aid filming For the seat assemblies, dynamic tests, again to the ECE of the tests. Nevertheless, the level of damage seen may Regulation 44 crash pulse, were carried out with modified indicate that the complete structure had insufficient strength seats mounted to a solid floor fitted to the sled. A second to withstand seat and occupant loading in a severe impact. double seat unit was positioned behind the modified seats so that two or four un-instrumented dummies could be used to 7.1.7 The second standard coach built minibus produce either single or double loading in the seats under test. Further seat modifications were carried out in the light In the test to vehicle GCB (with the 2-2-2 seats) the rear S of the failures, until the results were considered satisfactory. mountings of the front seat failed first followed by the Mathematical simulations were run in order to front mountings as the seat tipped forward. This was due to determine the likely foot to floor forces of the improved both their restrained occupant loading and heavy knee seats. A combination of component tests, calculations and loading from the unrestrained dummies behind. Although engineering judgement was used to decide if and how a the seat and fixings were identical at each end the failures vehicle needed strengthening and, if so, how to improve it. were different. The outboard rear foot failed by tearing off the bolt fixing it to the plate locked in the tracking and the front foot then tore off the leg. While the inboard legs 8.2 Modifications to the seats, seat legs and seat side failed by tearing off where they were attached to the seat mounts frame underneath the seat base, first at the rear and then at The standard versions of the four vehicles selected for the front. Due to the rear failures the seats tipped forward improvement each had seats from a different manufacturer. with their dummies still retained in the seat by the seatbelts TRL collaborated with these manufacturers to develop until the seat backs were almost horizontal, then the front improved seats and carried out seat sled tests to determine

10 their performance. The design target selected for these freedom to select the positions of pairs of legs across the improved seats were: width of the seat to match fitting requirements. As well as i For the seat to be restrained to the floor under both introducing a weak joint the practice of adjusting leg double and single loading without excessive movement positions can reduce the installed strength of the system if of the seat base. the legs are too close together, leaving parts of the double or triple seat unit cantilevered. Heavy external clamps were ii A small movement of the seat back in single loading. found to be effective in holding the legs to the front and rear iii A small to medium movement of the seat back in double cross-members of the seat base with one design. An loading to help reduce the injuries of the unrestrained alternative solution found with a second seat manufacturer rear occupant. consisted of replacing the open section cross members at the iv A deformation mode for the seat back that was unlikely front and rear seat base with box members and locally to result in it rupturing or separating leaving jagged packing the interior of the cross members in the vicinity of stumps under more severe loading than the ECE the legs with well fitting (aluminium) blocks. These blocks Regulation 44 crash pulse. were positioned so that the legs could be attached by bolting v A system that prevented excessive penetration of seat through the box and packing. A third solution found was to back by the knees of unrestrained rear occupants. fit stronger box section seat base cross-members and to weld vi Improved protection for the head impact of restrained the legs permanently in place to them. Welding the legs to occupants with the seat in front. the seat base results in a loss of fitting flexibility, however, it has the advantage that it would fix leg positions to those tested by the seat manufacturer. 8.3 Results and discussion of sled tests of improved seats Current regulations, for anchorages attached to seats, For three out of the four seat manufacturers who supplied place restrictions on deformation but this encourages rigid γ δ ε the original seats for these vehicles ( , and ) an structures that are more likely to snap off when subjected acceptable solution was found which met targets (i) to (iv) to higher loads, leaving dangerous stumps. Designs that and some also achieved to some extent targets (v & vi). are initially rigid but then fail progressively on overload Many modifications were made to the seats from the would be safer. This has been achieved with these fourth manufacturer (manufacturer β) and methods of experimental seats by using frame sections (tubes, etc.) for producing controlled energy absorption through bending the seat backs that are stronger in tension than of the back and of controlling rear occupant knee compression. This promoted progressive buckling of the penetration were found with this seat. However, ultimately front compression parts on overload. Alternatively, seat the seat frame was found to be too weak for the required back corner gussets, designed to buckle progressively on strengthening to be practical, so these seats were not used overload, were found to be effective. in the improved vehicle tests. Experimental seat backs, made from approximately As can be seen from the results of tests to standard 0.5 mm thick sheet steel attached by spot welds to the shells, there were a number of failures found with the seats and their attachments. The improvements developed for frame, were found to prevent excessive knee penetration each seat were specific to each design. Nevertheless, from unrestrained and restrained rear occupants and were certain generic problems and solutions emerged which are found to be effective in strengthening the joint between the summarised below. seat back and seat base. Seat rows are normally so close in The front and rear legs were too close together creating minibuses that knee and head contact with the seat in front high leg and feet to floor forces. Angling back the rear leg is very likely for adults, even when restrained. increased this distance and was more effective at resisting Heavy seat reinforcement was found in the seat back in the combination of shear and moments caused by the the area likely to be hit by the head of rear restrained inertia of the seat and occupant(s). occupants. Solutions to this problem were not fully The front feet were too small to spread their loads explored but it was found that reinforcement could be directly into the floor. (Under-floor washers commonly moved further forward and covered on both sides with used by converters are ineffective for compressive front padding. Alternatively, parts above the upper seatbelt foot loads). Larger feet, more feet or stronger floors are anchorage could be designed to be lightweight and to bend required. on rear impact. Cantilevered feet were used by all manufacturers to Photographs of selected seat development tests are provide room for the fixing bolts, but the feet bent and shown in Figure 11, Appendix B. reinforcing gussets often split, foot attachment welds Details of the final seat modifications used for each failed. The bolt heads were also found to pull through the vehicle are given in the following section. feet. It was also noted that some failure modes of cantilevered feet were such that floor and bolt forces 8.4 Modifications to improve the van conversions could be considerably magnified by a levering action, see Figure 11, Appendix B. A combination of thicker metal in 8.4.1 The first van conversion selected for improvement the bolt area and bigger and stronger gussets were used for The first van conversion selected for improvement, vehicle the feet of the improved seats. EVCS fitted with the reinforcing safety cage, was The clamp arrangements used to attach the legs to the seat considered to have sufficient strength in its original floor base were weak points. This option gives converters the and side mountings to withstand the loads of improved

11 seats, so a second, similar example was obtained. This was instead of the original three pairs of legs. The sub-frame fitted with seats that had been improved by the seat mounted on the floor for the front set of three seats had manufacturer (manufacturer δ) based on the results in the two additional fore/aft members to take the extra seat legs. TRL seat sled tests. The seats were again arranged in a The sub-frame also had additional fixings to the floor 2-2-2 pattern. The improved vehicle is referred to as where it extended beyond the seats. vehicle EVCM. The seats had heavier front and rear seat base The vehicle mountings for the middle double-seat unit and cross-members to which a stronger outboard pair of legs and the rear single seat were unchanged as they were considered an improved side mounting had been welded instead of the satisfactory for the purposes of this test programme and had original clamping system. The design of the seat back to seat not suffered failures in the original tests. base joint was revised to improve the transfer of the seat Other modifications to the seat had been made in back bending moments to the front and rear seat legs. The addition to the extra legs and side mount for the front set rear foot of the improved pair of legs was reinforced to resist of three seats. The legs themselves were stronger with a pull-through of the head of the bolt. more robust clamp arrangement to attach them to the seat base. The rear cantilevered feet were also modified with a 8.4.2 The second van conversion selected for improvement triangular gusset arrangement replacing the original ‘U’ The second van conversion selected for improvement, section. The original junction of the seat and seat-back had been reinforced by welding straps along the underside of vehicle CVCS, was considered to require strengthening of the seat mountings. The vehicle was essentially undamaged the seat and up the back of the seat side members to from its first test when the tracking failed. It was therefore reinforce the parts subjected to tension. Sled tests had improved and re-used. It was modified by TRL to make shown that this had the effect of forcing the seat back corner gusset plates to buckle and absorb energy on improved vehicle CVCM. The modifications made were to strengthen and provide additional seat mounting points on overload rather than snapping-off, when tested in the shell. Originally the floor was supported by a conjunction with strengthened legs. The seat backs were combination of heavy and light cross-members. Firstly the made from 0.5 mm sheet steel welded to the seat frame, to light cross-members were replaced by heavy ones. Then two control penetration of the knees of rear occupants, and strips of steel, running the length of the floor, were seam- were faced with cloth covered hardboard. welded on top of it. These strips were positioned to take the middle and outboard pairs of seat legs. Finally a heavy ‘U’ 8.5 Modifications to improve the coach-built shell shaped reinforcement was fabricated to fit inside an existing The coach-built shell selected for improvement was stiffening channel in the inner side panel of the van to vehicle GCBS. In the original test the failures had all been provide an additional side mounting point. The standard with the seats and not the vehicle or tracking. However, version of this vehicle had been fitted with seats from mathematical simulation of improved seats predicted that a β manufacturer . As it had been found impractical to tensile load of about 70 kN on the rear leg, and a strengthen these to the required standard, alternative seats compressive load of about 50 kN on the front leg would be were fitted to the improved shell. These seats were modified caused by double seat loading in the test (see Figure 12). δ seats made by manufacturer and were similar to those used This was well in excess of the minimum strength that is in the improved van conversion vehicle EVC described M typically required of this type of tracking when used with above, but they also had: seats with M2 anchorages. Therefore, a tracking pull test l a thin sheet-steel back to prevent excessive knee was devised to assess the strength of the tracking. The penetration; track was mounted to a solid base and a static load, l the seat back reinforcement in the region of rear head developed by a hydraulic ram attached through a impact, which was originally only covered by cloth, was cantilever, was applied to the ‘seat mounting plate’ locked moved 25mm forwards and covered at the back with a into the track. The first static tensile loading test resulted in 25mm layer of dense energy absorbing foam and at the failure of the bolts holding the tracking at a force of about front with the original seat upholstering foam; 40 kN. A repeat test using higher grade bolts resulted in a l a second pair of legs welded to the seat base close to the catastrophic failure of the system, with the ‘seat mounting middle of the double seat unit, aligned with the floor plate’ bursting out of the tracking at a load just below 50 kN. reinforcing strip. It was concluded from these results that the strength of the original two track seat fixings would probably be Each double seat unit was therefore retained by a total inadequate. A coach-built shell was therefore obtained of two pairs of legs and a side mounting. The seats were fitted with three lengths of tracking. The new shell, again arranged in a 2-2-2 pattern. referred to here as Vehicle GCBM, also had the improved joint, between the floor cross-members and the underlying 8.4.3 The third improved van conversion longitudinal members, that was used in the tests to the standard shell. Additional floor support was provided in The third improved van conversion (vehicle DVCM) was produced by the converters following discussions with the modified shell in the vicinity of the rear wheels to TRL. The seats, again from manufacturer γ, were fitted in a bridge the gap left for the tyres. Two equally spaced cross- 3-2-1 pattern as before with the front set of three mounted members were used, thus avoiding contact with the tyres. on a sub frame. However, the front triple-seat unit was The modified shell was again fitted with seats from now fitted with five pairs of legs and a side mounting manufacturer ε in a 2-2-2 pattern. In addition to the extra

12 Axial seat member force 9 Results and discussion of tests to (kN) improved minibus shells

The results of the tests to the improved minibus shells are summarised in the second half of Table 11, Appendix A. Before and after photographs of the improved minibus shells are shown in Figures 13 to 16, Appendix B. Out of the four modified vehicles tested two had no failures of the seats, the seat to vehicle attachments or the seatbelts.

One, vehicle EVCM, the first of the improved van conversions tested, suffered a seat failure due to a poor quality weld. On this vehicle the double loaded front seat double unit suffered a partial failure when the foot of the inboard rear leg pulled-off at the start of the impact. Examination of the foot welds showed that they had very poor penetration. Despite this failure the seat was still retained by the remaining front leg and side mounting. Seats to the same design had shown no distress when subjected to double loading in the seat development sled tests. The floor and side mountings of this vehicle were also considered adequate, so it was concluded that the seat in the full vehicle test would not have failed had the welds been of good quality. The second improved van conversion tested, vehicle CVC suffered a failure of both the seatbelts on the rear Figure 12 Computer simulation force results of double M double seat unit. The mounting of the seatbelt reel was such loaded seat subjected to Regulation 44 pulse that belt forces caused the belt mounting to distort and bend the ‘U’ shaped seatbelt reel frame causing it to spring open middle pair of legs fitted to each double seat unit to and allow the reels to pull out. This failure was due to a accommodate the third length of tracking, other combination of poor reel and mounting design. This modifications had been made to the seats. The legs problem had not been identified in the seat development themselves were stronger and were attached to the cross- tests because the seat manufacturer had made subsequent members of the seat base by both welding and bolting modifications to the reel mounting. However, it was through the box member. The front and rear box cross- concluded that minor changes to the design of the reel, the members of the seats were also packed with an aluminium reel mounting on the seat or to both would eliminate this block at each leg position to further reinforce it and to problem. It may seem surprising that the reels of the front prevent crushing due to the leg bolt. The original foot, double-loaded seats did not fail as double loading in cars which had linked the front and back legs, was replaced normally increases seatbelt loads. However, when the with a twin rear foot and a single front foot all reinforced upper anchorage is attached to the seat back, as in this with triangular gussets. The fore/aft seat-base members case, double loading will reduce seatbelt forces as the seat were reinforced with lengths of steel angle to help them back and anchorage is pushed forwards. The high-speed withstand the bending moments from the seat back. The film of this test showed that the seatbelts of the rear original junction of the seat base and seat back had also double-seat unit offered some restraint in the initial stage been reinforced by welding steel strips along the underside of the impact before failing. The HIC36 of the rear seat of the seat and up the back of the seat side members to ‘restrained’ dummy on impact with the back of the middle reinforce the parts subjected to tension. Sled tests had seat row was 1102. These seats had been modified to again shown that this had the effect of forcing the seat- improve the rear head impact area. The top of the seat back back corner gusset-plates to buckle and absorb energy on had deformed as intended, therefore, the HIC36 would overload rather than snapping-off when tested with probably have been below the normal criterion of 1000 strengthened legs. The seat backs were again made from had the seatbelts not failed. Apart from this, it was not 0.5 mm sheet steel welded to the seat frame to control possible to compare the effects of improving rear head penetration of the knees of rear occupants. The top area of contact with the standard vehicle because of its tracking the seat back had been modified to improve rear occupant and seat failures. However, the improved design appeared head protection. The box cross-members at the top of the to be a practical and effective solution. seat back were replaced with an upper and lower steel strip The improved seat back head impact area for the designed to bend on head contact. The headrest fixing improved coach-built minibus appeared to be effective sockets were also attached to these strips. with the HIC36 falling from 1298 to 761. However, again it is difficult to compare with the standard results where the rear and middle seats both moved.

13 No significant movement of the seats or seat backs was As it has been shown that it is necessary and practical to seen in the middle or rear seat rows in any of the tests to provide improved seatbelt systems it is recommended that the modified vehicles. The front seats of all the modified seatbelt anchorage regulations for these vehicles be revised vehicles were seen to bend forward when impacted by the to require a higher standard. unrestrained dummies behind. However, the seat back Improved standards for seatbelt systems should take into movement was controlled, absorbed energy, and restrained account many factors, including: or partially restrained the rear occupants. i the direction and magnitude of forces within the seat and The fitting of thin sheet metal backs appeared to be vehicle mountings; effective in preventing excessive knee penetration. The ii the need to confirm that there are no weak links in the femur forces of the restrained and unrestrained dummies load path from seatbelt anchorages to vehicle structure. were also low with this design of seat back. It was not possible in these tests to determine the injury risk to the Two options are proposed below that would improve front occupant from knee penetration. minibus safety. A third possible option could be based on the It was concluded, in the seat development tests, that seat method used in Australia for coaches, however, this method has some disadvantages which are also discussed below. backs, that bent forwards and absorbed energy when overloaded, were beneficial (in the case of double loaded seats). If this is accepted then it can be concluded from the 10.1 Adaptation of current seatbelt anchorage results of the tests to the standard line-built vehicle and the requirements for minibuses modified vehicles that it is possible to make minibuses, The analysis of minibus accident data and the full-scale seats and seatbelt systems capable of withstanding a ECE minibus crash test have shown that minibus seatbelt Regulation 44 crash pulse. This conclusion applies to the anchorages should ideally be able to withstand the forces three manufacturing methods (line-built, van conversion or generated by the ECE Regulation 44 crash pulse. It can be coach-built) and all combinations of occupant loading. As calculated that for single loading this is approximately the methods used to improve the seats and vehicles were equal to the ECE Regulation 14 seatbelt anchorage comparatively inexpensive and used conventional requirement for M1 vehicles (Economic Commission for technologies, it was also considered reasonable to require Europe, 1994). However, currently the anchorage strength minibuses to meet the test conditions used here. requirements for M2 and M3 vehicles (minibuses) are half The methods used in this project to improve the seats and one third respectively of that required for M1 vehicles. and vehicles have been shown to be effective and should An improved seatbelt regulation for M2 and M3 vehicles be of help in designing improved systems. could simply be to require them to withstand the same anchorage strength requirements as M1 vehicles. The fact that the factory-produced minibus, with anchorages to M1

10 Implications for improved regulations requirements (vehicle ALBS), performed well in the sled tests supports this argument. It could be argued that M1 The analysis of fatal minibus accident data when combined anchorage requirements would be over demanding for M3 with full-scale crash data indicates that the current vehicles which, due to their high mass, are less likely to requirements for minibus seatbelt anchorages are suffer high accelerations when impacting other vehicles. insufficient to provide protection in some of the more severe However, it was found practical to improve the larger accidents. Vehicles with seat and seatbelt systems that can coach-built shells to meet the Regulation 44 loading. A withstand ECE Regulation 44 loading would provide further possible criticism of the ECE Regulation 14 protection for more than 50% of serious minibus accidents, anchorage pull tests is that it is quasi-static and produces taking into account the bias towards serious accidents in the longer duration anchorage loading than those in a crash. fatal accident database. Improved regulations based on the However, this will only have the effect of introducing a loading caused by a Regulation 44 crash pulse would small additional factor of safety. As the pull test is far require anchorages of approximately twice the strength simpler, easier to interpret and less expensive than a currently required for M2 vehicles. Some may conclude that dynamic test it has many advantages. the test severity chosen for this test program is inadequate Therefore, requiring minibuses to meet the M1 for improved regulation for minibus seatbelt anchorages anchorage standard may well be sufficient to ensure because it only covers 50% of the fatal and series minibus improved anchorage strength. accidents analysed. However, the cost of providing anchorages in minibuses to M1 requirements is thought to be relatively small but it would increase rapidly if the test 10.2 Sled tests of the complete shell using the ECE were more demanding. A more demanding test may well Regulation 44 crash pulse enter the area of diminishing returns by resulting in high A sled test of complete shells or representative sub- costs for very few extra injuries saved. For the more violent systems could be used evaluate seatbelt anchorage accident the ultimate limit on providing additional injury strength. The Regulation 44 crash pulse could be used or, savings is, as already discussed, the seat belt loadings that if considered necessary, a crash pulse could be derived by the human frame can withstand. This limit has not been carrying out a programme of crash tests of representative established for the three point belt but it is thought to be minibus models. Sled tests would have the advantage of close to the M1 anchorage requirement. producing more realistic dynamic loading of the complete

14 system and could provide information on knee and rear Although this test method has virtues, separate testing of head-impact protection. However, a large capacity sled the floor and seats also has the above disadvantages, however, would be required to take a full minibus shell fully these disadvantages are comparatively insignificant. equipped with seats and occupant dummies. 10.4 Other considerations for an improved minibus 10.3 The Australian coach floor/seat regulation seatbelt anchorage test The essence of the Australian ADR68 (Federal Office of Other considerations for an improved regulatory test Road Safety, 1994) is that the performance of three rows should include the following: of double seats, with restrained dummies in the middle row i the small-scale of production of minibuses, seats and and unrestrained dummies in the rear row, is first assessed seatbelts when compared with car manufacture; on a sled. The seats are mounted on a sled and subjected to ii the methods used to manufacture minibuses; a 20 g deceleration pulse. The test assesses the iii the tailoring of each vehicle to match the customer performance of the seats with regard to withstanding requirements for seat location plan, seat quality and seat double loading and protecting the heads of the restrained to floor fixing options (direct bolting to the floor or via dummies if they contact the seat in front. In addition to tracking). this, by mounting the double-loaded seat to the sled by a separate force frame, the foot to floor forces and moments Some flexibility in the requirement for testing of all for the seats under assessment are found. Also the seat seating options is recommended. This could be achieved if mounting points, in the floor of the coach that the seats are the approval authority is allowed to use engineering intended for, are subjected to similar forces and moments judgement to only test the worst cases. Worst-case testing to those found in the sled tests. could, for instance, allow families of seat plans and seats This method has the advantage that it would require types, for a particular vehicle model, to be approved with effective protection for restrained occupants, restrained one or two tests for a particular manufacturer or converter. occupants who still hit the seat in front and restrained occupants whose seat back is hit from behind by 10.4.1 A further cause for concern unrestrained occupants. It would also have some A further cause for concern was the failures or potential advantages for minibus manufacturers. It would probably failures seen in the test programme. The failure of the poor help to share test costs with the seat manufacturers and quality foot-to-leg weld highlighted a similar problem also alternative seats could be fitted to one floor design if they found during the seat development tests where welds of generated a force no larger than that with which the floor standard components were also found to be faulty due to was tested and had similar foot fixing geometry. However, poor weld penetration. Weld quality is difficult to control it was clear from the modes of failure seen in the testing of for the small-scale simplified seat construction methods standard minibuses, and the component tests of seat and used by minibus seat manufacturers. However, methods tracking, that local distortions of the seat and floor will should be devised to ensure the consistent quality of normally occur when seats are tested fitted in a minibus. critical welded joints. These could result in significant additional stress The failure of the seatbelt reels in the test to the concentrations in the seat and floor that could precipitate improved van conversion, vehicle CVCM, could have been failures that would not have been observed when the seats prevented by improvements to the reel frame or to the reel and floor were tested separately. This is because: mounting on the seat. The seatbelts used for minibus seats l For the sled test the seat must be attached to a rigid are normally seatbelt reels for cars, adapted by fitting frame to measure accurately the force transfer and it different lengths of webbing. The limited buying power of would not be practical to mimic any deformation of the minibus seat manufacturers may make it difficult for them real vehicle floor in the test procedure. Fixing the seat to to require improvements such as riveted bars to close the a rigid floor would effectively strengthen the seat, open end of the ‘U’ frame. Therefore, improvements to the particularly the feet. mounting are the most practical option. In the test to improved vehicle CVC it was only obvious after failure l M For the floor, it would be difficult first to establish and had occurred that the design of the reel mounting could then mimic on the floor, the distortions and/or partial cause the reel to fail. However, once alerted, a test failures seen in seats tested on a flexible floor. authority would be able to reject this type of design by l Overall, floor and seat distortion would reduce seat and visual inspection. Due to the compact nature of minibus floor stress by allowing some movement, which would seating, many seat manufacturers make use of unusual belt reduce peak acceleration. However, local floor routing and additional guides and brackets. These can distortion could increase the local stresses in seat joints suffer local distortions, which can affect the performance such as those between the seat feet and the seat leg. This of the whole seatbelt system. In a sled test for approval of could precipitate failures in the feet gussets, which in the whole minibus the standard belts and reels would be turn could increase local floor stress. Failure modes of tested on the standard seat. For seat-anchorage pull tests feet side gussets were observed in seat tests, which heavier webbing is often used attached directly to the reel increased local floor stress by acting as a lever, see anchorage point. This method will not detect the effect of Figure 11, Appendix B. minor local distortions of the anchorage on the seatbelt reel

15 assembly. Therefore it recommended that for pull tests on trajectory could be found from mathematical simulations minibuses with seat mounted anchorages the actual belt or full-scale tests and the pass criterion of HIC36 1000 installation that will finally be used should be tested with could be used. Similarly a knee impactor could be the reel lock activated. developed to test for rear knee impact. The protection criteria could be aimed at reducing seat penetration to 11 Considerations for additional or protect the seat occupants from injury or to protect both future regulations the occupant and the impacting knee from injury. Alternatively, sled based tests could be used to explore As already concluded, it would be beneficial to have seat the impact with the seat back. A combination of restrained backs that: and unrestrained test dummies could to some extent be used i bent forward and absorbed energy rather than breaking to explore seat-back overload performance due to double off when subjected to overloading from impacts of loading. However, dummies of different statures would be unrestrained occupants or in more severe accidents; required to fully explore the head impact area with the seat back. Instead, seat-back safety could be explored by a ii protected the seat occupant when the seat back is combination of sled test and sub-system tests. impacted by the knees of restrained and unrestrained occupants behind; iii protected the head of restrained and unrestrained 12 Conclusions occupants when they make contact with the back of the seat in front. 1 Accident data have been used to identify the most The seat back overload behaviour could be found by a common types of serious minibus accidents. These have pull test using an adaptation of the M1 anchorage test also been used to establish the relationship between the equipment. The following features would be beneficial. proportion of serious and fatal accidents and the velocity l The seat back should not deform significantly under the change seen by the minibus. This relationship has been normal M1 loads. used to determine an appropriate velocity change for sled tests of minibus seatbelt systems, if the more l Once the seat back starts to deform, it should have a serious accidents are to be catered for. reasonably high and constant stiffness so that it absorbs energy. 2 Suitable data from a full scale minibus crash test have been located and used in conjunction with the accident l The seat back should not fail catastrophically. data to define an acceleration pulse for sled tests of l The initial load at which bending starts should be set to minibus seatbelt systems. limit the seatbelt force on the occupant or to limit the force on an unrestrained occupant behind. 3 The ECE Regulation 44 frontal ‘crash pulse corridor’ was selected for sled testing of minibus shells, seats and Regulations on compulsory wearing of seatbelts would occupant restraints because it matched the velocity and obviously have implications for the need to consider seat acceleration pulse selected and provided suitable back overload. Compulsory wearing of seatbelts is already simplification and tolerances for this type of test. required in the UK, for all occupants of minibuses with an 4 The test conditions selected here, from accident data, are un-laden weight of less than 2540 kg. For these vehicles an at least twice as demanding as the current international overload test may be thought unnecessary if high wearing anchorage strength requirements for minibuses. rates can be achieved. However, seat-mounted-anchorages are used almost exclusively in minibuses and it is difficult 5 The standard minibus shells (made to the UK standards with these to provide upper anchorage strengths much in that were applicable at the time of their manufacture in excess of the M1 requirement. Therefore, improved 1996) suffered serious failures of the front double- minibus upper anchorages will probably have less loaded seats and/or the vehicle structure at the seat overload capacity than normally found in cars, where it is attachment points when subjected to the ECE easier to provide strength in excess of the minimum Regulation 44 crash pulse. Some also suffered failures requirements because the upper anchorages are normally of the less heavily loaded seats. part of the car structure. Consequently, for minibuses, even 6 Methods used to improve minibus seats and vehicle if compulsory seatbelt wearing is required, an energy structure have been described. absorbing seat back would provide additional protection in 7 The tests of the improved minibuses and the ‘line-built’ more serious accidents. minibus designed to exceed current standards showed that The concerns about seat-back behaviour when impacted it was possible to make vehicles, seats and seatbelts by the head and knees of passengers sitting behind could systems to withstand the test conditions used here. Because be addressed by additional sub-systems tests to the seat the methods used to improve the seats and vehicles were back. The area of the seat back likely to be struck by the comparatively inexpensive and used conventional head could be tested using an adaptation of the European technologies, it was also considered reasonable to require Enhanced Vehicle-safety Committee WG17 pedestrian minibuses to meet higher standards in future. Therefore, it headform tests (EEVC Committee, 1998). Appropriate is recommended that, subject to a full assessment of the adult and possibly child head zones, impact velocity and likely costs and benefits, consideration should be given

16 towards revising seat belt anchorage regulations in these vehicles to meet a higher standard, possibly in line with existing M1 ‘car’ standards. 8 The implications of these results for improved design standards for minibus seats and seatbelt anchorages have been discussed and possible improvements to approval methods have been outlined.

13 References

Economic Commission for Europe (1999). Uniform provisions concerning the approval of vehicles with regard to safety-belt anchorages. ECE Regulation No. 14. Geneva: United Nations.

Economic Commission for Europe (1998). Concerning the adoption of uniform technical prescriptions for wheeled vehicles, equipment and parts which can be fitted and/or be used on wheeled vehicles and the conditions for reciprocal recognition of approval granted on the basis of these prescriptions. ECE Regulation No. 44. Geneva: United Nations.

EEVC Committee (1998). Improved test methods to evaluate pedestrian protection afforded by passenger cars. EEVC Working Group 17 report. Delft, Netherlands: TNO Crash-Safety Research Centre.

Federal Office of Road Safety (1994). Occupant protection in buses. Australian design rule 68/00. Australia: Federal Office of Road Safety.

Lawrence G J L, Lowne R W, Thornton S, Wall J G and Brett M (1996). Bus and coach safety and study for minibus component tests. Project Report PR/VE/201/96. Crowthorne: TRL Limited. (Unpublished report available on direct personal application only)

Lawrence G J L and Hardy B J (1997). Tests of retro-fit seat belts and their relevance to visual approval methods. Project Report PR/SE/347/97. Crowthorne: TRL Limited. (Unpublished report available on direct personal application only)

Wall J G (1995). Coach, bus and minibus safety preliminary accident data review. Project Report PR/VE/ 173/95. Crowthorne: TRL Limited. (Unpublished report available on direct personal application only)

14 Acknowledgements

The work described in this report forms part of a Department for Transport, Local Government and the Regions funded research programme conducted by the Safety and Environment Division of TRL. The help and co-operation of the minibus and seat manufacturers involved is gratefully acknowledged.

17 Appendix A: Tables

It should be noted that in the following tables of accident statistics Fatal and Serious accidents have been classified by the most serious injury to a minibus occupant rather than the usual method of by the most serious injury of any persons involved in the accident.

Table 1 STATS 19 - Types of minibus accident against minibus accident injury severity category for all accidents in the years 1991 to 1994 inclusive †

Number of minibus accidents

Minibus overturned Minibus not overturned Total

Minibus accident with: F F+S All F F+S All F F+S All

Object in carriageway * 1 6 22 0 8 43 1 14 65 Object off carriageway * 4 39 114 7 76 265 11 115 379 Objects on and off carriageway * 2 11 33 4 16 59 6 27 92 Minibus hit no object or vehicle 0 20 82 3 71 1,178 3 91 1,260 One other vehicle 2 21 68 10 179 3,177 12 200 3,245 Two or more other vehicles 2 3 15 3 55 953 5 58 968 One or more other vehicles plus 3 13 57 7 80 1,041 10 93 1,098 object on or off carriageway

Total 14 113 391 34 485 6,716 48 598 7,107

† Minibus classification includes motor caravan * Object other than another vehicle

Table 2 STATS 19 - Other vehicle type against minibus accident injury severity category for minibus impact with one vehicle only, in the years 1991 to 1994 inclusive †

Number of minibus accidents

Minibus overturned Minibus not overturned Total

Minibus accident with: F F+S All F F+S All F F+S All

Pedal cycle 0 0 1 0 0 305 0 0 306 Moped 0 0 2 0 0 40 0 0 42 Motor scooter 0 0 0 0 0 11 0 0 11 Motor cycle 0 0 2 0 2 203 0 2 205 Combination 0 0 0 0 0 3 0 0 3 Invalid tricycle 0 0 0 0 0 1 0 0 1 Other three-wheeled Car 0 0 0 0 0 1 0 0 1 Taxi 0 0 0 0 1 28 0 1 28 Car (four-wheeled) 1 11 40 2 111 2,121 3 122 2,161 Minibus/Motor caravan 0 2 4 0 12 66 0 14 70 PSV 0 0 1 1 6631 664 Light goods < 1.5 Tons 1 1 3 0 9 157 1 10 160 Heavy goods > 1.5 Tons 0 6 10 7 32 134 7 38 144 Other motor vehicle 0 1 5 0 6 38 0 7 43 Other non-motor vehicle 0 0 0 0 0 6 0 0 6

Total 2 21 68 10 179 3,177 12 200 3,245

† Minibus classification includes motor caravan

18 Table 3 STATS 19 – Type of vehicle hit against first point of contact on minibus, broken down by accident injury severity category, for minibus impact with one vehicle only, in the years 1991 to 1994 inclusive †

Number of minibus accidents

Front Side (Off & Near) Rear Rest Total

Minibus accident with: F F+S All F F+S All F F+S All F F+S All F F+S All

Minibus overturned Taxi 0 0 0 0 0 0000 000000 Car (four-wheeled) 1 4 14 0 6 24 0 1 2 0 0 0 1 11 40 Light goods <1.5 Tons 1 1 2 0 0 1000 000113 Heavy goods >1.5 Tons 0 2 2 0 2 5023 0000610 PSV 000 000001 000001

Minibus not overturned Taxi 0 0 15 0 0 10 0 1 3 0 0 0 0 1 28 Car (four-wheeled) 2 75 1,271 0 25 529 0 11 309 0 0 12 2 111 2,121 Light goods <1.5 Tons 0 5 83 0 3 40 0 1 31 0 0 3 0 9 157 Heavy goods >1.5 Tons 4 20 68 2 8 36 1 4 30 0 0 0 7 32 134 PSV 1 4 19 0 1 27 0 1 17 0 0 0 1 6 63

Total 9 111 1,474 2 45 672 1 21 396 0 0 15 12 177 2,557

† Minibus classification includes motor caravan

Table 4 STATS 19 - Distribution of minibus accident by accident injury severity category (in accidents with cars type group and heavy vehicle type group, for minibus impacts with one vehicle only, first point of contact on minibus front, accidents in the years 1991 to 1994 inclusive †)

Proportion of minibus accidents

Minibus overturned Minibus not overturned Total

Injury accidents Injury accidents Injury accidents Number as percent Number as percent Number as percent of injury of injury of injury Minibus accident with: accidents F F+S ALL accidents F F+S ALL accidents F F+S ALL

Car type group * 16 12.5 31.3 100 1,369 0.15 5.8 100 1,385 0.29 6.1 100 Heavy vehicle type group # 2 0.0 100.0 100 87 5.80 27.6 100 89 5.60 29.2 100

† Minibus classification includes motor caravan * Car Type Group = Car (Four-Wheeled) + Taxi + Goods < 1.5 Tons # Heavy Vehicle Type Group = Goods > 1.5 Tons + PSV

Table 5 STATS 19 - Distribution of accidents by road speed limit and accident injury severity category, (first point of contact on minibus front, minibus impacts with one vehicle only, for accidents in the years 1991 to 1994 inclusive †)

Number of minibus accidents per speed limit band ‡ Minibus accident Minibus accident with: injury severity 1-30 mph 31-40 mph 41+mph Total

Car type group* Fatal 0 0 4 4 Fatal + Serious 27 10 48 85 All 770 117 498 1,385

Heavy vehicle type group # Fatal 1 1 3 5 Fatal + Serious 3 3 20 26 All 29 9 51 89

† Minibus classification includes motor caravan ‡ Includes both overturned and non overturned minibus accidents * Car Type Group = Car (Four-Wheeled) + Taxi + Goods < 1.5 Tons # Heavy Vehicle Type Group = Goods > 1.5 Tons + PSV

19 Table 6 STATS 19 - Distribution of minibus accidents by point of first contact on other vehicle against vehicle type for serious + fatal minibus accident injury severity category, (first point of contact on minibus front, for accidents in the years 1991 to 1994 inclusive †)

Number of front minibus Serious + Fatal accidents category ‡, against point of first contact on other vehicle

Minibus accident with: Front Side Rear Total

Car type group * 58 18 9 85 Heavy vehicle type group # 9 7 10 26

† Minibus classification includes motor caravan ‡ Includes both overturned and non overturned minibus accidents * Car Type Group = Car (Four-Wheeled) + Taxi + Goods < 1.5 Tons # Heavy Vehicle Type Group = Goods > 1.5 Tons + PSV

Table 7 Police fatal accident files selected as involving minibus or camper-van and manually classified by most serious injury to minibus or camper-van occupant

Accident type Fatal Serious Slight Total

Minibus accidents 15 6 53 74 Camper-van accidents 15 5 4 24

20 15 accident α Rear passengers (1 fatal, 1 serious) 16 occupants?Driver (slight) number Driver partial (fatal) 2 8 Passenger (fatal) 17 Ejection TRL Slightly / of any none Fatally Seriously 6 1 0 4 Front passenger partial (fatal) bus hit Limit 120‡UnknownPerpendicular to minibus 112 0 96 96-24 2 0 0-80‡ 6 5 1 6 No 0 112 96 1 No 5 1Total number of occupants 6 3 32 1 5 39 Passenger partial (fatal) No 54 4 10 12 Speed (km/h) Number of minibus occupants injured minibus Of object impact* at impact*road of 112‡ 0 112 1 0 10 vehicle Of FrontFront 45 64‡ FrontFront 102Front 80‡Front 72‡ 72‡ 0 96 56‡ Slow 0 64‡ 9 112 64 1 0 1 0 No 96 0 0 5 0 1 No 0 1 No 9 14 Area (60% O/S) frontfront Nearside Nearside 51‡ 56‡ 64‡ 48 0 1 0 No 3 minibus Side (50% centre) – 96‡ 0 80 2 2 3 2 Rear Passengers (both fatal) 21 Area of of other olls after the impact? impacted impacted at Yes Side – 88‡ 0 112 1 1 9 5 - Yes Side – underran r Minibus Minibus denotes ejected casualty was wearing a seat belt by the minibus other vehicle? CarCarCarCarCarCar NoCar NoLand Rover No NoHGV No No NoHGV No NoHGV N/K NoHGV NoHGV Full front No NoHGV Front (50% O/S) Yes No Yes Full N/K YesHedge Full YesSafety barrier Full front Front (60% N/S) NoSafety barrier Front (100%) No N/K No NoGround - NoTree - Front (100%) NoTree Offside Front (30% N/S) Signpost Front (30% N/S) No - Front (75% N/S) No 32 Rear N/K Rear Rear - Front (100%) Yes - 102 - Side (75% N/S) 112 Yes Offside rear wheel Side N/K – N/K No 88 Yes Side Yes 86‡ N/K Front (50% centre) – Front (50% centre) – 0 – 96 – 0 80‡ 96 N/K 112 1 56 4 112 96‡ 1 0 0 0 4# 0 1 2 0 1 No 5 0 0 9 80 N/K No 96 No 112 1 1 3 0 1 4 6 0 1 6 7 13 No Driver (serious) 11 No 20 19 18 Table 8 Summary of all Police files located minibus accidents resulting in Fatal or Serious Injury to occupant α N/K = Not known ‡ denotes speed just before accident (cruising speed) # 2 killed by fire * Impact speeds are best estimates in Police files; negative speed indicates other vehicle travelling same direction as Mini Object impacted

21 Table 9 Vehicle speeds at first contact corrected, where necessary, to allow for typical emergency braking and/or typical witness over estimate of speed along with approximate vehicle masses, (Police files of minibus frontal accidents resulting in Fatal or Serious Injury to minibus occupant)

Best estimate of Best estimate of TRL accident minibus speed at other vehicle speed Minibus mass (kerb + number first contact (km/h) # at first contact (km/h) # passengers) (kg) Other vehicle mass (kg)

1 45 102 2455 1210 2 35 83 2550 1475 3 37 35 1855 1065 4 34 0 1855 990 5 102 Slow, perpendicular to MB 2020 1120 6 32 112 1440 1340 8 51 45 1855 1925 9 51 42 1495 16000 (HGV) 11 54 0 1550 Mass assumed infinite (HGV) 14 8 35 2290 17000 (HGV) 20 56 0 1925 Mass assumed infinite (Tree)

# Correction factors for typical overestimates of speed by independent witnesses and for the effects of emergency breaking were used as necessary when Police estimates of speed at impact were not available.

Table 10 Calculated change in Minibus velocity (Delta V), over period of main impact, using impact speeds and mass data from Table 9

TRL accident number 1 2 3 4 5 6 8 9 11 14 20

Delta V, (∆V) (km/h) 48 43 26 11 37 69 50 85 54 38 56

Number of occupants Fatally injured 0 0 0 0 0 0 1 1 4 5 1 Seriously injured 9 5 1 1 2 1 0 0 1 1 1 In minibus 9 11 1 1 8 1 1 1 5 6 3

22 , completely , seat tipped , seat tipped long way Continued .... impact Yes Yes Yes failed , rear, No, but seats , outer , back No , rear , rear, No, but seats track front legs failed Yes feet pulledoff bolt headsway long a tipped forwards Yes inner legs pulled off seat base Yes Yes foot pulledoff bolt &rear seatthen forwards side mount front foot and pulled off side mount by knees from behind failed leg to seatbase joint tipped forwards a long way , front No, front No, but No t ripped beginning of Yes legs wentthroughYes forwards a trackfixing off plate distorted penetrated heavy Partial Yes tippingforward moderate at rear backs bentforward and feet seats tipped No No, but , rear No No, but No No No No slight distortion Yes No No No No, feet at rear moderatefront feet floor seat some front and distortion caused distortion Partial failure? failure? detached? failure? failure detached? 36 1133 1298 1356 seat long way moved forwards ripped out forwards fee No, but tipped forward late forward and light floor caused tipping bent Rear seats Middle seats Front seats , rear, No, but seat 730 No No No No (single loaded) (no occupant loading) (double loaded) and rear a Partial feet bent Yes mounting seat base mount failure Seat Seat Floor Seat Seat Floor Seat Seat light Nobut No, ted backs , rear No No, but N/A ripped Yes at rear Partial failure? failure? detached? HIC and Floor Alt. 3-2-1 # No No No 453 No No No No, but 2-2-22-2-2 No, but No, but No 843 No, but No, but No 2-2-2 2-2-2 No 2-2-2 No 3-2-1 No No, but No (3L)(2L+S)(1L+S) back bent forwards γ (2LT)(2LT) (2LT) part of tracking off floor tipped and a long wayfeet seat of track seats tipped and back a long way out of track detached at β (2L)(2L) β distor (2L) (2LT)(2LT)(2LT) failure at rear leg to seat base tipped and moved forwards off tore legs pulled forwards then ε α δ (1L+S)(1L+S) (1L+S) side pulled off due to side (2LT)(2LT)(2LT) distortion front and moderate moderate β Seat code van conversion van conversion van conversion coach built line built van conversion coach built S S S S S S S Table 11 Summary of results sled tests to standard and modified minibus shells EVC DVC CVC FCB BVC GCB Vehicle code fitting plan Standard minibus shells ALB

23 , rear No, remaining Yes pulledfoot off due topoor weld attachments retained seat relatively well mounting. he triple seat unit had four front and rear legs. The double three sisted of one front and rear leg per pair, some with without diagonal braces etc. No No No No No No No No No No No No No No No No failure? failure? detached? failure? failure detached? 36 1102 (improved 761* No No No No No No 1506* 2099* Rear seats Middle seats Front seats , bent No (single loaded) (no occupant loading) (double loaded) mountingreels to fail back)* Yes Seat Seat Floor Seat Seat Floor Seat Seat failure? failure? detached? HIC and Floor 2-2-2 No 2-2-2 No No No 2-2-2 No No No 3-2-1NoNoNo (2L+S)(2L+S) seatbelt caused δ (2L+S) ε (3LT)(3LT)(3LT) (improved back) δ (1L+S) (1L+S) (1L+S) γ (5L) (2L+S) (1L+S) Seat code legs and two rear with one side mounting the single seat had front legs, leg Unsatisfactory performance. van conversion van conversion coach built van conversion M M M M Table 11 (Continued) Summary of results sled tests to standard and modified minibus shells DVC (S) = Seat with side mounting on one end. (T) = Seats attached to floor by tracking. (nL) =indicates number of pairs legs per seat row. Leg type ‘n L’ con Seat with pair/s of legs attached to the seat base, ‘n’ # = These seats had legs that were integral with the seat construction, all large feet plates and multiple diagonal braces. T *Bold = = Modified results cannot be compare with the standard where rear and middle seats both moved. GCB EVC Vehicle code fitting plan Modified minibus shells CVC

24 Appendix B: Figures

Before test After test

Damage to back of front seat from knee of unrestrained Damage to rib of front dummy from knee impact rear dummy Figure 4 Standard line built minibus test

25 Before test After test

Floor failure under feet/legs of front double loaded seat Head/knee impact damage from restrained dummy

Figure 5 First standard van conversion tested

26 Before test After test

Failure of rear seat attachments

Figure 6 Second standard van conversion tested

27 Before test After test

Failure of front seat attachments (triple seat unit)

Figure 7 Third standard van conversion tested

28 Before test After test

Failure of outer front seat attachment Failure of inner front seat attachment

Failure of one of the attachments of the rear seat

Figure 8 Fourth van conversion tested

29 Before test After test

Failure of front seats and attachments

Deformation of rear seat attachments

Figure 9 First standard coach built minbus tested

30 Before test After test

Failure of outer front seat attachment Failure of inner front seat attachment

Figure 10 Second standard coach built minibus tested

31 Stronger seat backs and feet find next weak link Stronger seats cause track failures (test load exceeded tacking design strength)

Weak energy absorbing seat backs Stronger energy absorbing seat backs

Sheet steel backs protect occupant from knee impacts from Foot failure mode causes levering action rear

Figure 11 Seat development tests – All double loaded with Regulation 44 pulse, e.g. more than twice current minimum requirements

32 Before test – Improved foot and leg to seat joint Before test – Improved side mount to seat joint

Before test After test

Failure of poor quality foot weld causes front seat unit to tip and twist

Figure 13 First modified van conversion tested

33 Before test – Two pairs of strengthened legs on reinforced Before test – Seat side mount attached to reinforced side floor rail of van

Before test After test

Failure of rear seatbelts – reels pulled out of reel frames

Figure 14 Second modified van conversion tested

34 Before test After test

Triple front seat unit after test (seat cusions removed. Note Rear double and single seat units after test improved/more legs)

Figure 15 Third modified van conversion tested

35 Before test After test

After test showing bending of energy absorbing seat back and triple track to leg fixings

Figure 16 Modified coach built minibus test

36 Abstract

There is considerable interest in improving the crash safety of buses, minibuses and coaches despite their comparatively good safety record. Publicity given to accidents involving these vehicles has led to demand for safety measures, notably the installation of seatbelts in coaches and minibuses. This demand has been met to some extent in the UK by the recent requirement for all children, on journeys relating to school or other child activities, to be transported in vehicles fitted with seatbelts. Further UK legislation will require all new buses and coaches, apart from those specifically designed for urban use and standing passengers, to be fully equipped with seatbelts. However, the requirements for seatbelt anchorage strengths in these larger vehicles are less demanding than for cars. An evaluation of samples of seats, seatbelt systems and minibuses, the potential for improving their performance and the appropriateness of current regulations was undertaken. This paper describes: a study of the available UK national accident statistics, an in-depth analysis of the UK minibus accidents held in the TRL database of police files of fatal accidents, the selection of a suitable crash pulse for testing minibus seatbelt systems which represents the real life accident situation, tests of standard minibuses in order to establish the performance of current seatbelt systems, the development and testing of improved minibus seatbelt systems and suggestions for improved test methods for minibus seatbelt systems.

Related publications

CR285 Fatal accidents involving forward control vans, minibuses and motor caravans in 1986 by S J Rattenbury and P F Gloyns. 1991 (price £20, code C) CT87.1 Bus coach and minibus safety update (1995-1999) Current Topics in Transport: selected abstracts from TRL Library’s database (price £20) CT9.3 Seatbelts and in-car safety update (1999-2001) Current Topics in Transport: selected abstracts from TRL Library’s database (price £20)

Prices current at December 2001

For further details of these and all other TRL publications, telephone Publication Sales on 01344 770783, or visit TRL on the Internet at www.trl.co.uk.

37 38