Ice Prevention or Removal on the Veteran's
Glass City Skyway Cables
Prepared by: Douglas K. Nims, Victor J. Hunt, Arthhur J. Helmicki, Tsun-Ming T. Ng
Prepared for: The Ohio Department of Transportation, Office of Statewide Planning & Research
State Job Number 134489
August 2014
Final Report
1 Technical Report Documentation Page
1. Report No. 2. Government A cc ess ion No. 3. Rec ipient's Catalog No. FHWA/OH-2014/11
4. Title and Subtitle 5. Report Date (Month and Year) August 2014 Ice Prevention or Removal on the Veterans Glass City Skyway Cables 6. Performing Organization Code
7. Author(s) 8. Performing Organization Report No.
Douglas Nims, Victor Hunt, Arthur Helmicki, Tsun-Ming Ng
9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)
University of Toledo 2801 W. Bancroft St. 11. Contract or Grant No. Toledo, OH 43606 SJN 134489
12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered Ohio Department of Transportation Final Report Research Section 1980 West Broad St., MS 3280 14. Sponsoring Agency Code Columbus, OH 43223
15. Supplementary Notes
16. Abstract
The Veteran’s Glass City Skyway is a cable - stayed bridge in Toledo, Ohio owned by the Ohio DOT. Five times in the seven winters the VGCS has been in service, ice has formed on the stay cables. Ice up to 3/4” thick and conforming to the cylindrical shape of the stay has formed. As the stays warm, ice sheds in curved sheets that fall and can be blown across the bridge. The falling ice sheets pose a potential hazard and may require lane or bridge closure. Because of the specialized knowledge required, this problem required a team including experts in icing, the VGCS construction, the structural measurement system on the bridge, and green technology. The VGCS stay sheaths are made of stainless steel, have a brushed finish, lack the usual helical spiral and have a large diameter. No existing ice anti/deicing technology was found to be practical. Therefore, ODOT elected to manage icing administratively. A real-time ice monitoring system for local weather conditions on the VGCS and the stays was designed. The system collects data from sensors on the bridge and in the region. The study of the past weather and icing events lead to quantitative guidelines about when icing accretion and shedding were likely. The monitoring system tracked the icing conditions on the bridge with a straightforward interface so information on the icing of the bridge is available to the bridge operators. If the conditions favorable to icing occurred, the monitoring system notified the research team and appropriate ODOT officials. If ice has formed, the monitor tracks the conditions that might lead to ice fall.
17. Key Words 18. Distribution Statement
No restrictions. This document is available to the public through the Ice, Bridges, Cable-stayed, Hazard Mitigation, Ice Removal, Ice Prevention National Technical Information Service, Springfield, Virginia 22161
19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No. of Pages 22. Price Unclassified Unclassified 316 $ 652,894.58
Form DOT F 1700.7 (8-72) Reproduction of completed pages authorized
2 Ice Prevention or Removal on the Veteran's Glass City Skyway Cables
Prepared by: Douglas K. Nims Victor J. Hunt University of Toledo University of Cincinnati
Arthur J. Helmicki Tsun-Ming T. Ng, Ph.D., P.E. University of Cincinnati University of Toledo
August 2014
Prepared in cooperation with the Ohio Department of Transportation and the U.S. Department of Transportation, Federal Highway Administration
The contents of this report reflect the views of the author(s) who is (are) responsible for the facts and the accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the Ohio Department of Transportation or the Federal Highway Administration. This report does not constitute a standard, specification, or regulation.
3 Acknowledgments
The authors would like to acknowledge the University of Toledo graduate students Mr. Ali Arbabzadegan, Mr. Joshua Belknap, Mr. Nutthavit Likitkumchorn, and Mr. Clinton and University of Cincinnati Infrastructure Institute graduate students, Mr. Shekhar Agrawal, Mr. Biswarup Deb, Mr. Jason Kumpf and Ms. Chandrasekar Venkatesh, who played a significant role in the research and the writing of this report. Chapter 1, Introduction, was primarily written by Mr. Ali Arbabzadegan with contributions from Mr. Clinton Mirto. Chapter 3, Phase I Research, was primarily written by Mr. Arbabzadegan with contributions from Mr. Joshua Belknap and Mr. Clinton Mirto. Chapter 4, Weather Background, Modeling, and Analysis, was written by Mr. Belknap with contributions from Mr. Arbabzadegan, and Mr. Mirto. Chapter 5, Development of the VGCS Dashboard and Initial Dashboard Results was written by students from UCII. Chapter 6, New Local Weather Sensor Testing, was written by students from UCII. Chapter 7, Experimental Studies on the Sheath Specimens, primarily written by Mr. Arbabzadegan, Mr. Likitkumchorn and students from UCII. Chapter 8, Deployment of New Sensors and Upgrade of Dashboard, was written by Mr. Arbabzadegan and students from UCII. Chapter 9, Sensor Development, was written by Mr. Likitkumchorn. Mr. Mirto assisted in the editing of this report. University of Cincinnati graduate students Mr. Biswarup Deb, Mr. Chandrasekar Venkatesh, Mr. Nithyakumaran Gnanasekar, and Ms. Monisha Baskaran were instrumental in maintaining and updating the Dashboard for this past winter and delivering the standalone computer system to ODOT Dr. Sridhar Viamajala graciously allowed the Scott Park Icing Experiment Station to be built on an unoccupied portion of the concrete pad used for his sustainable energy research. This project was performed under the aegis of the University of Toledo – University Transportation Center. The continuous support of Director Richard Martinko and Associate Director Christine Lonsway has made this project possible. The authors thank Ms. Kathleen Jones and Dr. Charles Ryerson of the U.S. Army Cold Regions Research and Engineering Laboratory for the frequent discussions about the project and extensive analysis, support in developing the criteria for the ice fall dashboard, and their help in the editing of this report. This project was sponsored and supported by the Ohio Department of Transportation. The authors gratefully acknowledge their financial support. Mr. Mike Gramza, P.E. and Mr. Tim Keller, P.E. were the technical liaisons and the authors appreciate their support and input throughput the project. The author would also like to thank Mr. Mike Madry, Mr. Dave Kanavel and Mr. Matt Harvey from ODOT for access to the bridge and assistance in observing the icing events and Mr. Jeff Baker, P.E. (now retired from ODOT) for his assistance in defining criteria for the ice fall dashboard and reviewing the User Manual
4 Dedication
This report is dedicated to the late Professor K. Cyril ‘Cy’ Masiulaniec of Mechanical, Industrial, and Manufacturing Engineering of the University of Toledo. Cy was one of the initial investigators on this project and he was active until the week before his passing when he developed the final details and directions for mounting the thermistors. He will be remembered for his attention to detail and patient thorough explanations of the thermal science that made a significant contribution to this project. Cy was always willing to step up and help our students, his department and the college in many ways. UT students consistently recognized him as an outstanding teacher and he received the UT College of Engineering Award for Teaching Excellence.
5 Abstract
The Veteran’s Glass City Skyway (VGCS) is a large cable - stayed bridge in Toledo, Ohio owned by the Ohio Department of Transportation (ODOT). The VGCS carries I- 280 over the Maumee River. Five times in the seven winters the VGCS has been in service, ice has formed on the stay cable sheaths. Ice accumulations have been up to approximately 3/4” thick and the ice conforms to the cylindrical shape of the stay sheath. As the stays warm, they shed the ice in curved sheets that fall up to two hundred and fifty feet to the roadway and the pieces of ice can be blown across several lanes of traffic on the bridge deck. The falling ice sheets require lane closures or even closure of the entire bridge and could present a potential hazard to the traveling public. Because of the unique nature of the problem, the need for a quick response and the specialized nature of the icing knowledge required, this problem has been addressed with an expert team. The team includes experts in icing from the U.S. Army Cold Regions Research and Engineering Laboratory and the NASA Glenn Icing Branch, the ODOT project managers from the bridge construction, the engineers who designed and implemented the existing structural strain measurement system on the bridge, and experts in green technology. The stay sheaths of the VGCS are unique: they are made of stainless steel, have a brushed finish, lack the usual helical spiral and have a large diameter. No existing ice anti/deicing technology was found to be practical for the VGCS. Therefore, ODOT elected to manage icing administratively. To do this, the research team designed a real-time monitoring system for local weather conditions on the VGCS and the stays as well as the surrounding area. The monitoring system collects a comprehensive set of data from local sensors on the bridge as well as other sensors in the Toledo region. The study of the past weather and icing events lead to quantitative guidelines about the weather conditions that made icing accretion and shedding likely. These guidelines form the core of the algorithms in the ice monitoring system implemented on the bridge. The monitoring system tracked the icing conditions on the bridge with a straightforward interface so information on the icing of the bridge is readily available to the bridge operators. If the conditions favorable to icing occurred, the monitoring system notified the research team and appropriate ODOT officials. If ice forms, the monitor tracks the conditions that might lead to ice fall. The benefits of completing this project include observations of an icing event, review of historical icing events, a building a local weather station on the bridge and stays to collect real-time data on icing and developing the monitoring system. Because no commercial sensor for directly measuring the presence or state of ice on the sheath exists, an electrical resistance based sensor has been developed.
6 Table of Contents
Cover Sheet ...... 12 Technical Report Documentation Page ...... 2 Disclaimer ...... 3 Acknowledgments ...... 4 Dedication ...... 5 Abstract ...... 6 Table of Contents ...... 7 List of Figures ...... 12 List of Tables ...... 19 Chapter 1: Introduction ...... 21 Section 1.1: Bridge Background ...... 21 Section 1.2: Summary of Goals and Objectives ...... 24 Section 1.3: Summary of Results...... 25 Section 1.4: Organization of this Report ...... 27 Chapter 2: Goals, Objectives, Research Approach and Benefits ...... 29 Section 2.1: Overview of Chapter ...... 29 Section 2.2: Goal ...... 29 Section 2.3: Objectives ...... 29 Section 2.4: Expert Team Approach to the Research ...... 33 Section 2.5: Benefits ...... 36 Section 2.6: Chapter Summary ...... 37 Chapter 3: Phase I Research ...... 39 Section 3.1: VGCS Sheaths ...... 39 Section 3.2: Literature Review ...... 40 Section 3.2.1 Known Icing Problems on Other Bridges ...... 40 Section 3.2.2 Anti-Icing/Deicing Technologies found in literature ...... 41 Section 3.3: Technology Matrix ...... 48 Section 3.4: Sensors on the VGCS...... 50 Section 3.4.1: Sensors on the VGCS prior to the 2012 – 2013 Winter ...... 50 Section 3.4.2: Sensors added in 2012 – 2013 ...... 51 Section 3.4.3: Sensors added in 2013 – 2014 ...... 51 Section 3.5: Chapter Summary ...... 53 Chapter 4: Weather History, Modeling and Analysis ...... 55
7 Section 4.1: Introduction ...... 55 Section 4.2: Description of the basic weather that gives rise to an ice storm ...... 55 Section 4.3: VGCS Weather History ...... 56 Section 4.4: Lessons Learned from Previous Icing Events ...... 73 Section 4.5: Analysis ...... 74 Section 4.6: Chapter Summary ...... 76 Chapter 5: Development of the VGCS Dashboard and Initial Dashboard Results ...... 77 Section 5.1: Introduction ...... 77 Section 5.2: Weather Data ...... 80 Section 5.2.1: Introduction ...... 80 Section 5.2.2: Data Sources ...... 80 Section 5.2.3: Data Classification ...... 83 Section 5.2.4: Data Collection and Storage ...... 86 Section 5.3: Ice Accumulation Determination Algorithm ...... 87 Section 5.3.1: Data Update Time ...... 88 Section 5.3.2: Ice Accumulation Algorithm ...... 88 Section 5.3.3: Station Individual Weights ...... 89 Section 5.3.4: Threshold weights ...... 90 Section 5.3.5: Ice Shedding ...... 91 Section 5.4: Ice Persistence Algorithm ...... 91 Section 5.4.1: Ice States ...... 91 Section 5.4.2: Ice Accumulation Persistence Algorithm ...... 92 Section 5.4.3: Ice Presence Confirmation ...... 95 Section 5.4.4: Ice Shedding Persistence Algorithm...... 96 Section 5.5: Monitor Website ...... 99 Section 5.5.1: Dashboard Main Panel ...... 100 Section 5.5.2: Weather Map ...... 101 Section 5.5.3: History ...... 103 Section 5.5.4: Implementation Tools ...... 104 Section 5.6: Performance Testing ...... 104 Section 5.6.1: System Reliability Test ...... 104 Section 5.6.2: Ground Truth ...... 106 Section 5.7: Conclusions ...... 125 Chapter 6: New Local Weather Sensor Testing ...... 126
8 Section 6.1: Introduction ...... 126 Section 6.1.1: Geokon Thermistors ...... 126 Section 6.1.2: Dielectric Wetness Sensor ...... 127 Section 6.1.3: Solar Radiation or Sunshine Sensor ...... 127 Section 6.1.4: Rain Tipping Bucket ...... 128 Section 6.1.5: Goodrich Ice Detector ...... 128 Section 6.2: Geokon Thermistor 3800-2-2 ...... 129 Section 6.2.1: Laboratory experiment on temperature measurement using Geokon Thermistors ...... 130 Section 6.2.2: Installation of Geokon Thermistors 3800-2-2 at the VGCS on Stays 8 & 20 ...... 135 Section 6.3: LWS-L Dielectric Leaf Wetness Sensor ...... 145 Section 6.3.1: Laboratory experiment on measurement of output voltage using LWS-L Leaf Wetness Sensor...... 145 Section 6.4: Sunshine Sensor BF5 ...... 149 Section 6.4.1: Laboratory experiment on measurement of solar radiation using Sunshine Sensor BF5...... 150 Section 6.5: Met One Rain Tipping Bucket ...... 153 Section 6.5.1: Laboratory experiment on measurement of precipitation using Tipping Bucket ...... 154 Section 6.6: Goodrich Ice Detector ...... 156 Section 6.6.1: Laboratory experiment on measurement of ice presence/thickness using Goodrich Ice Detector 0872F1 ...... 157 Section 6.7: Conclusions ...... 161 Chapter 7: Field Study of Temperature Effect on Stay Sheaths ...... 162 Section 7.1: Introduction ...... 162 Section 7.2: Design of Icing Experiment Station ...... 162 Section 7.3: Design of the UT Icing Tunnel and Design ...... 164 Section 7.4: Icing Accretion and shedding Experiments at Scott Park...... 168 Section 7.5: Thermal Experiments at Scott Park ...... 172 Section 7.6: Anti/de-icing Fluid Experiments at Scott Park ...... 175 Section 7.7: Coating Experiments at Scott Park ...... 176 Section 7.8: Coating Experiments using Icing UT Tunnel ...... 178 Section 7.8.1: Testing Procedure ...... 178 Section 7.8.2: Experiments – Icing Progression ...... 179 Section 7.8.3: Result Summery of Icing Tunnel Coating Tests ...... 199
9 Section 7.9: Field Experiment Trips ...... 200 Section 7.10: Conclusions ...... 211 Chapter 8: Deployment of New Sensors and Upgrade of the Dashboard ...... 213 Section 8.1: Introduction ...... 213 Section 8.2: Self Supporting Instrumentation Tower Design ...... 213 Section 8.2.1: Tower Design ...... 213 Section 8.2.2: Anchorage System Design ...... 214 Section 8.3: VGCS Ice Sensors Bridge Installation trip (May 16-17, 2013) ...... 215 Section 8.4: Changes to the Ice Accumulation Algorithm ...... 220 Section 8.5: Changes to the Ice Shedding Algorithm...... 225 Section 8.6: Changes to the Dashboard ...... 227 Section 8.6.1: Dashboard Main Panel ...... 228 Section 8.6.2: Map (Weather Data by location) ...... 229 Section 8.6.3: New Sensors Plotting ...... 231 Section 8.7: Insights Gained from the Operation of the Upgraded Dashboard ...... 235 Section 8.7.1: Ice Events (Winter 2013/2014) ...... 236 Section 8.7.2: Sensor Performance ...... 244 Section 8.7.3: Issues and Observations from Winter Performance ...... 249 Section 8.8: Conclusions ...... 250 Chapter 9: Ice Presence and State Sensor Development ...... 252 Section 9.1: Introduction ...... 252 Section 9.2: Ice Presence and State Sensor Laboratory Testing ...... 252 Section 9.2.1: Sensors and Data Acquisition System ...... 252 Section 9.2.2: Design of Experiments ...... 254 Section 9.2.3: Laboratory Test Results ...... 257 Section 9.3: UT Icing Sensor in Full Scale Experiments ...... 265 Section 9.3.1: Specimens and Data Acquisition System Setup ...... 266 Section 9.3.2: Full Scale Outdoor Tests ...... 270 Section 9.3.3 Full Scale Experiments Result ...... 272 Section 9.4: Conclusion and Next Steps ...... 275 Chapter 10: Transition and Maintenance ...... 276 Section 10.1: Introduction ...... 276 Section 10.2: Standalone Computer System ...... 276 Section 10.3: Maintenance ...... 276
10 Chapter 11: Conclusion, Benefits, Implementation and Future Work ...... 279 Section 11.1: Summary of Goals and Objectives ...... 279 Section 11.2: Results ...... 280 Section 11.3 Benefits ...... 284 Section 11.4: Implementation ...... 285 Section 11.5: Transition and Long Term Maintenance ...... 286 Section 11.6: Archiving of Supporting Documents ...... 286 Section 11.7: Recommendations for Future Work ...... 286 Bibliography ...... 289 Appendix A: Technology Matrix ...... 306
11 List of Figures Figure 1: Veteran’s Glass City Skyway (photo credit will be provided) ...... 21 Figure 2: Veteran’ Glass City Skyway’s Illuminated Glass Pylon (ODOT, 2010) ...... 22 Figure 3: Ice Accumulation on the East Side of VGCS (Baker, 2007) ...... 23 Figure 4: Ice on the Pylon and the VGCS Glass ...... 23 Figure 5: Large Piece of Ice Almost Hitting a Car ...... 24 Figure 6: Application of Superhydrophobic Coating on the Surface (Ryerson, 2008) ... 42 Figure 7: DC Bias Deicing where Electrolysis forms Bubbles (Ryerson, 2008) ...... 43 Figure 8: Pulse Electro Thermal Deicing (PETD) (Ryerson, 2008) ...... 44 Figure 9: Ice Being Released using Ice Dielectric Heating (Ryerson, 2008) ...... 44 Figure 10: Navy Vertical Launch Systems with Electrically Heated Door Edges (Ryerson, 2008) ...... 45 Figure 11: Infrared Heaters above the CRREL Entrance (Ryerson, 2008) ...... 46 Figure 12: Aviation Facility using Infrared Radiant System (Ryerson, 2008) ...... 46 Figure 13: Photonic Deicer for Deicing of Power Lines (Couture, 2011) ...... 48 Figure 14: Damaging ice storm footprint map, 1946-2014 in the lower 48 states and portions of the lower tier of Canada...... 56 Figure 15: Dashboard readout for February 21, 2011 ...... 59 Figure 16: Overview of ice accreting on stay at 10:29 PM Sunday evening ...... 60 Figure 17: Close up of ice accreting on stay at 10:29 PM Sunday evening ...... 60 Figure 18: Stay cable diagram with ice accumulation...... 61 Figure 19: Ice Accumulation up east side of stay February 22, 2011 ...... 63 Figure 20: Frozen Rivulets and bare metal on the west side of stays February 22, 2011 ...... 63 Figure 21: Thermocouple reading between ice and stay February 23, 2011 ...... 64 Figure 22: Thermocouple reading between the ice and stay February 24, 2011 ...... 65 Figure 23: Cracking in ice from chipping away ice, February 23, 2011 ...... 66 Figure 24: Section where ice was chipped away to take temperature readings February 23, 2011 ...... 67 Figure 25: Ice thickness measurements on back stay 19 February 23, 2011 ...... 67 Figure 26: Ice thickness measurements on back stay 19 February 23, 2011 ...... 68 Figure 27: Ice accumulation on pylon glazing February 24, 2011 ...... 69 Figure 28: Ice on bridge deck after 80-90% had shed, February 24, 2011 ...... 69 Figure 29: Weather Summary for the week of February 20, 2011 (Weather Underground, 2011) ...... 71 Figure 30: Solar radiation counts February 22, 2011 ...... 72 Figure 31: Solar radiation counts February 23, 2011 ...... 72 Figure 32: Solar radiation counts February 24, 2011 ...... 73 Figure 33: Process Flow Diagram ...... 79 Figure 34: Map Showing Distances of Weather Stations from VGCS ...... 83 Figure 35: Ice Determination Algorithm ...... 89 Figure 36: Dashboard Speedometer ...... 91 Figure 37: Ice Accumulation Flowchart ...... 93 Figure 38: Sample Ice Accumulation Message Alert ...... 94 Figure 39: Dashboard with Ice Accumulation Alert ...... 95 Figure 40: Ice Presence Flowchart ...... 95
12 Figure 41: Ice Shedding Flowchart ...... 97 Figure 42: Sample Ice Shedding Message Alert ...... 98 Figure 43: Dashboard with Ice Shedding Alert ...... 99 Figure 44: State Transitions possible from Red Level 3 ...... 99 Figure 45: Dashboard Main Panel ...... 101 Figure 46: Dashboard History Panel ...... 103 Figure 47: Weather Summary on Feb 20, 2011 ...... 107 Figure 48: Screenshot Showing Ice Accretion on VGCS ...... 108 Figure 49: Weather Summary on Feb 21, 2011 ...... 109 Figure 50: Weather Summary on Feb 22, 2011 ...... 110 Figure 51: Ice Accumulation on Stays on Feb 22, 2011 ...... 110 Figure 52: Weather Summary on Feb 24, 2011 ...... 112 Figure 53: Example of Ice Shedding Alert ...... 112 Figure 54: Ice Falling from VGCS on Feb 24, 2011 ...... 113 Figure 55: Weather Summary on Feb 25, 2011 ...... 114 Figure 56: Feb 24, 2011 Algorithm Performance Graph ...... 115 Figure 57 : Contribution of the Icing Criteria and Weather Stations ...... 116 Figure 58: Solar Radiation Variation on Feb 24, 2011 ...... 117 Figure 59: Features of the Past Icing Events ...... 119 Figure 60: Dec 12, 2007 Algorithm Performance Graph ...... 120 Figure 61: Mar 28, 2008 Algorithm Performance Graph ...... 122 Figure 62: Dec 17, 2008 Algorithm Performance Graph ...... 123 Figure 63: Jan 03, 2009 Algorithm Performance Graph ...... 124 Figure 64: Geokon 3800-2-2 Thermistor ...... 130 Figure 65: Naked Thermistor Bead (photo credits, John Flynn, Geokon Inc.) ...... 130 Figure 66: Canary Systems Multilogger Software ...... 131 Figure 67: Measurement trend of eight thermistors ...... 132 Figure 68: Thermistors kept in freezer ...... 133 Figure 69: Thermistors immersed in water left to freeze ...... 133 Figure 70: Readings simultaneously noted by handheld GK 404 ...... 133 Figure 71: Standard thermometer immersed in setup to record temperature ...... 133 Figure 72: Thermistor Characteristics at Freezing ...... 134 Figure 73: Side View of Gage Locations at VGCS ...... 135 Figure 74: Custon Thermistor Mount Fabricated for Installing on Stay Surface ...... 136 Figure 75: Thermistor Placed on East Side of Stay ...... 136 Figure 76: Thermistors Placed on Upper Side of Stay ...... 136 Figure 77: Far View of Thermistor Installation of Stay ...... 137 Figure 78: Thermistor Cables Being Routed to Multiplexer Inside White Box ...... 137 Figure 79: Stay Sheath Cross Section Showing Thermistor Positions ...... 138 Figure 80: Stay 20 Thermistors Temperature Trend ...... 139 Figure 81: Stay 8 Thermistors Temperature Trend ...... 140 Figure 82: Characteristics for Stay 20 Thermistors on March 15 ...... 143 Figure 83: Characteristics for Stay 20 thermistors on March 9 & 10 ...... 144 Figure 84: Leaf Wetness Sensor functional diagram ...... 145 Figure 85: Experimental setup of data logger CR1000 with LWS-L Leaf Wetness Sensor ...... 146
13 Figure 86: Droplets of Water Sprinkled on Leaf ...... 146 Figure 87: LWS-L partially immersed in cup of water ...... 147 Figure 88: LWS-L immersed in cup left to freeze ...... 147 Figure 89: LWS Wetness Test ...... 148 Figure 90: LWS Freezing Temperature Test ...... 148 Figure 91: Sunshine Sensor BF5 (side view) and Detailed Construction ...... 149 Figure 92: Sunshine Sensor BF5 Set Up on a Deck for Unobstructed Exposure to Solar Radiation ...... 150 Figure 93: Solar radiation characteristics over an extended period of 16 days ...... 151 Figure 94: A typical partly cloudy day chosen to see the daily solar radiation characteristics ...... 152 Figure 95: A typical clear sunny day taken as example to see the daily solar radiation characteristics ...... 153 Figure 96: Rain Tipping Bucket (from top left clockwise) distant view, top view and inside view ...... 154 Figure 97: Rain Bucket lab ...... 155 Figure 98: Gessler Buret ...... 155 Figure 99: Rain Bucket accuracy experiment (actual vs tipping volume) ...... 156 Figure 100: The Goodrich Ice Detector (external and function diagrams) ...... 157 Figure 101: Ice Detector Mounted for Experiment ...... 159 Figure 102: Microcare Anti-Stat Freezing Spray ...... 159 Figure 103: Probe before Spraying ...... 159 Figure 104: Probe After Spraying ...... 159 Figure 105: Frequency/Ice thickness characteristics of 0872F1 during freezing spray experiment ...... 160 Figure 106: Thickness measurement using calipers ...... 160 Figure 107: Google Earth Screenshot of Scott Park ...... 162 Figure 108: Experimental Setup ...... 163 Figure 109: Sensors on South-faced Specimen ...... 164 Figure 110: Data Acquisition System ...... 164 Figure 111: SolidWorks Design for the UT Icing Tunnel ...... 165 Figure 112: UT Icing Tunnel ...... 165 Figure 113: Testing Section of the UT Icing Tunnel ...... 166 Figure 114: Misting System in the Testing Section ...... 167 Figure 115: Panasonic HX_A100D Camera (Panasonic 2013) ...... 167 Figure 116: Mounting System of Testing Section ...... 168 Figure 117: Spraying a Mist of Water on North-faced Specimen ...... 168 Figure 118: Pattern of Ice Accumulation on Outdoor Tests ...... 169 Figure 119: Water beneath the Ice Layer before Shedding ...... 169 Figure 120: Ice Shedding Steps ...... 170 Figure 121: Stay’s Behavior in Icing Test – 2/15 to 2/18 ...... 171 Figure 122: Stay’s Behavior in Icing Test – 2/20 to 2/22 ...... 171 Figure 123: Thermal Experiment Setup ...... 173 Figure 124: Strands Configuration in Thermal Tests ...... 173 Figure 125: Deicing Pattern in Thermal Test ...... 174 Figure 126: Accumulated Ice in Anti-icing Thermal Test ...... 174
14 Figure 127: Formation of Ice in Chemical Anti-icing Test ...... 175 Figure 128: Drip Tube System used in Chemical Deicing Test ...... 176 Figure 129: Hydrobead Sprayed on Half of the Specimen ...... 177 Figure 130: Water Droplets due to Hydrobead ...... 177 Figure 131: Specimen’s Behavior in Coating Test ...... 178 Figure 132: Uncoated - 40 Micron - 0:00 min ...... 179 Figure 133: Uncoated - 40 Micron - 0:15 min ...... 179 Figure 134: Uncoated - 40 Micron - 0:30 min ...... 180 Figure 135: Uncoated - 40 Micron - 0:45 min ...... 180 Figure 136: Uncoated - 40 Micron – 1:00 min ...... 180 Figure 137: Uncoated - 40 Micron – 1:30 min ...... 181 Figure 138: Uncoated - 40 Micron – 2:00 min ...... 181 Figure 139: Uncoated - 40 Micron – 4:00 min ...... 181 Figure 140: Uncoated - 40 Micron – 6:00 min ...... 182 Figure 141: Uncoated - 40 Micron – 8:00 min ...... 182 Figure 142: Uncoated - 40 Micron – 10:00 min ...... 182 Figure 143: Uncoated - 40 Micron – After Test ...... 183 Figure 144: None Coating - 40 Micron – Shed Ice Sheet ...... 183 Figure 145: Hydrobead-Coated Specimen ...... 184 Figure 146: Hydrobead – 40 Micron – 0:00 min ...... 184 Figure 147: Hydrobead – 40 Micron – 0:15 min ...... 185 Figure 148: Hydrobead – 40 Micron – 0:30 min ...... 185 Figure 149: Hydrobead – 40 Micron – 0:45 min ...... 185 Figure 150: Hydrobead – 40 Micron – 1:00 min ...... 186 Figure 151: Hydrobead – 40 Micron – 1:30 min ...... 186 Figure 152: Hydrobead – 40 Micron – 2:00 min ...... 186 Figure 153: Hydrobead – 40 Micron – 4:00 min ...... 187 Figure 154: Hydrobead – 40 Micron – 6:00 min ...... 187 Figure 155: Hydrobead – 40 Micron – 8:00 min ...... 187 Figure 156: Hydrobead – 40 Micron – 10:00 min ...... 188 Figure 157: Hydrobead – 40 Micron – After Test ...... 188 Figure 158: Hydrobead – 40 Micron – Shed Ice Sheet...... 189 Figure 159: PhaseBreak TP – 40 Micron – 0:00 min ...... 189 Figure 160: PhaseBreak TP – 40 Micron – 0:15 min ...... 190 Figure 161: PhaseBreak TP – 40 Micron – 0:30 min ...... 190 Figure 162: PhaseBreak TP – 40 Micron – 0:45 min ...... 190 Figure 163: PhaseBreak TP – 40 Micron – 1:00 min ...... 191 Figure 164: PhaseBreak TP – 40 Micron – 1:30 min ...... 191 Figure 165: PhaseBreak TP – 40 Micron – 2:00 ...... 191 Figure 166: PhaseBreak TP – 40 Micron – 4:00 ...... 192 Figure 167: PhaseBreak TP – 40 Micron – 6:00 ...... 192 Figure 168: PhaseBreak TP – 40 Micron – 8:00 ...... 192 Figure 169: PhaseBreak TP – 40 Micron – 10:00 ...... 193 Figure 170: PhaseBreak TP – 40 Micron – After Test ...... 193 Figure 171: PhaseBreak TP – 40 Micron – Shed Ice Sheet ...... 194 Figure 172: WeatherTITE – 40 Micron – 0:00 min ...... 194
15 Figure 173: WeatherTITE – 40 Micron – 0:15 min ...... 195 Figure 174: WeatherTITE – 40 Micron – 0:30 min ...... 195 Figure 175: WeatherTITE – 40 Micron – 0:45 min ...... 195 Figure 176: WeatherTITE – 40 Micron – 1:00 min ...... 196 Figure 177: WeatherTITE – 40 Micron – 1:30 min ...... 196 Figure 178: WeatherTITE – 40 Micron – 2:00 min ...... 196 Figure 179: WeatherTITE – 40 Micron – 3:00 min ...... 197 Figure 180: WeatherTITE – 40 Micron – 4:00 min ...... 197 Figure 181: WeatherTITE – 40 Micron – 6:00 min ...... 197 Figure 182: WeatherTITE – 40 Micron – 8:00 min ...... 198 Figure 183: WeatherTITE – 40 Micron – 10:00 min ...... 198 Figure 184: WeatherTITE – 40 Micron – After Test ...... 198 Figure 185: WeatherTITE – 40 Micron – Shed Ice Sheet ...... 199 Figure 186: Stay Specimens at Different Angles and Orientations ...... 201 Figure 187: Data-logging System Setup ...... 201 Figure 188: Sunshine Sensor Setup...... 201 Figure 189: Ice Detector Placed Right Beside Stay ...... 201 Figure 190: Stay Thermistors Zip-tied on Sheath ...... 201 Figure 191: Leaf Wetness Sensor Taped on top of Specimen ...... 201 Figure 192: Ice Detector at Various Times Throughout the February 16 Experiment . 202 Figure 193: Leaf Wetness Sensor at Various Times Throughout the February 16 Experiment ...... 203 Figure 194: Ice Detector Characteristics (Toledo experiments on February 16) ...... 204 Figure 195: Characteristics of stay thermistors (Toledo, February 16) ...... 205 Figure 196: Leaf Wetness Sensor ice melting characteristics ...... 205 Figure 197: LWS-LS with Different Slants ...... 207 Figure 198: Top & Side Thermistors Setup ...... 207 Figure 199: Ice Detector Setup ...... 207 Figure 200: First Spray Shower ...... 207 Figure 201: Garden Hose mount on ladder (left) & hand held (right) for experiment on ice detector & leaf sensors ...... 208 Figure 202: Ice Detector at Various Times during Experiment (Left and Middle during ice accretion; right during deicing) ...... 208 Figure 203: Stay thermistor characteristics (Toledo experiments February 20 – 21) .. 209 Figure 204: Leaf Wetness Sensor Characteristics (Toledo, February 20 – 21) ...... 210 Figure 205: Ice Detector characteristics (Toledo, February 20 – 21) ...... 211 Figure 206: Tower Anchorage System ...... 214 Figure 207: Rohn’s Weather Tower Drawing ...... 215 Figure 208: Tower mounted near stay 19 ...... 215 Figure 209: Initial Plan by UT Research Team for Tower Mounting ...... 216 Figure 210: Leaf Wetness Sensor Zip-tied to Cross-arm ...... 217 Figure 211: Rain Bucket mounted on cross-arm using leveling bracket ...... 218 Figure 212: Sunshine Sensor attached to cross-arm with steel U-bolts ...... 218 Figure 213: Ice Detector Mounted using Steal Worm Band Clamps ...... 219 Figure 214: Ice Detector Mounted Close Up ...... 219 Figure 215: Sensor Cable Conduit ...... 219
16 Figure 216: CR1000 Datalogger Setup Insider Tower Cabinet ...... 219 Figure 217: Close up of Weather Tower ...... 220 Figure 218: Completed New Weather Station Near Stay 19 ...... 220 Figure 219: Flowchart of existing Ice Accumulation Algorithm (Agrawal, 2011) ...... 222 Figure 220: Flowchart for revised ice accumulation algorithm ...... 224 Figure 221: Flowchart of existing Ice Shedding Algorithm (Agrawal, 2011) ...... 226 Figure 222: Flowchart for revised ice shedding algorithm ...... 227 Figure 223: Dashboard Main Panel ...... 228 Figure 224: Example Snapshot of Weather Map, with Pop-up for Ice Detector ...... 230 Figure 225: Last 48 hour report of Solar Sensor (Global Radiation) ...... 231 Figure 226: Last 48 hour report of Leaf Wetness Sensor ...... 231 Figure 227: Stay 20 Thermistors plot (January 1 – July 1) ...... 232 Figure 228: Stay 8 Thermistors plot (January 1 – July 1) ...... 232 Figure 229: Ice Detector plot (June 1 – July 1) ...... 233 Figure 230: Leaf Wetness Sensor plot (June 1 – July 1) ...... 233 Figure 231: Rain Tipping Bucket plot (June 1 – July 1) ...... 234 Figure 232: Sunshine Sensor plot (June 1 – July 1) ...... 234 Figure 233: Ice Detector & LWS Characteristics during Ice Event, December 9, 2013237 Figure 234: VGCS Icing camera view before noon ...... 239 Figure 235: Ice Detector & Leaf Wetness Sensor characteristics on February 20 ...... 240 Figure 236: Ice detector & Leaf wetness Sensor characteristics on April 3 ...... 242 Figure 237: Rain Tipping Bucket & Leaf Wetness Sensor characteristics on April 3 ... 242 Figure 238: Solar Radiation & Stay Thermistor 8X08TWS characteristics on April 3 .. 243 Figure 239: Leaf Wetness Sensor characteristics winter 2013/14 ...... 245 Figure 240: Stay Thermistor characteristics winter 2013/14 ...... 245 Figure 241: Sheath thermistors warming faster than outer (March 4, 2014) ...... 246 Figure 242: Rain Tipping Bucket characteristics winter 2013/14 ...... 247 Figure 243: Ice Detector characteristics winter 2013/14 ...... 248 Figure 244: Solar radiation Sensor characteristics winter 2013/14 ...... 249 Figure 245: Relative distribution of alarms triggered by new sensors ...... 249 Figure 246: UT Icing Sensor Circuit ...... 253 Figure 247: Electro Spacing Area of the UT Icing Sensor ...... 253 Figure 248: UT Icing Sensor Connected to Data Acquisition System ...... 254 Figure 249: Dashboard of UT Icing Sensor ...... 254 Figure 250: 1-mm Electro Spacing UT Icing Sensor ...... 255 Figure 251: 7-mm Electro Spacing UT Icing Sensor ...... 255 Figure 252: Water Measurement ...... 256 Figure 253: Ice Measurement ...... 256 Figure 254: 75% Slush Measurement ...... 256 Figure 255: 50% Slush Measurement ...... 256 Figure 256: 25% Slush Measurement ...... 256 Figure 257: Ice Measurement at 6 mm thickness ...... 257 Figure 258: Ice Measurement at 13 mm thickness ...... 257 Figure 259: Ice Measurement at 19 mm thickness ...... 257 Figure 260: Resistance of Ice for 1-mm Electro Spacing Sensor ...... 258 Figure 261: Dashboard Screenshot of Ice Measurement ...... 258
17 Figure 262: Resistance of 75% Slush for 1-mm Electro Spacing Sensor ...... 259 Figure 263: Dashboard Screenshot of 75% Slush Measurement ...... 259 Figure 264: Resistance of 50% Slush for 1-mm Electro Spacing Sensor ...... 260 Figure 265: Dashboard Screenshot of 50% Slush Measurement ...... 260 Figure 266: Resistance of 25% Slush for 1-mm Electrode Spacing Sensor ...... 261 Figure 267: Dashboard Screenshot of 25% Slush Measurement ...... 261 Figure 268: Resistance of Water for 1-mm Electro Spacing Sensor ...... 262 Figure 269: Dashboard Screenshot of Water Measurement ...... 262 Figure 270: Resistance of Ice for 7-mm Electro Spacing Sensor ...... 263 Figure 271: Resistance of 75% Slush for 7-mm Electro Spacing Sensor ...... 263 Figure 272: Resistance of 50% Slush for 7-mm Electro Spacing Sensor ...... 264 Figure 273: Resistance of 25% Slush for 7-mm Electro Spacing Sensor ...... 264 Figure 274: Resistance of Water for 7-mm Electro Spacing Sensor ...... 264 Figure 275: Resistances for 6-mm Thickness and 7-mm Electro Spacing Sensor ...... 265 Figure 276: VGCS Stainless Steel Specimens ...... 266 Figure 277: HDPE Specimen and Frame Structure ...... 266 Figure 278: North Facing Specimen with 120 Stands Inside ...... 267 Figure 279: Sensors Setup on VGCS Specimen ...... 268 Figure 280: Sensors Setup on HDPE Specimen ...... 268 Figure 281: Cross Section and Sensor Setup Orientation of both Specimens ...... 268 Figure 282: UT Icing Sensor on HDPE Specimen ...... 268 Figure 283: MicroStrain V-Link ...... 269 Figure 284: MicroStrain TC-Link ...... 269 Figure 285: MicroStrain WSDA-Base (Signal Receiver) ...... 270 Figure 286: V-Link and UT Icing Sensor ...... 271 Figure 287: Ice Testing ...... 271 Figure 288: Slush Testing ...... 271 Figure 289: Water Testing ...... 271 Figure 290: UT Icing Sensor Initial Test ...... 272 Figure 291: Misting Water on VGCS Specimen ...... 273 Figure 292: Ice Accumulation on VGCS Specimen ...... 273 Figure 293: Stay Behavior in Icing Experiment ...... 274 Figure 294: Flowchart for Stand Alone System ...... 278
18 List of Tables Table 1: Viable Technologies ...... 31 Table 2: Information Required to Revolve Uncertainties ...... 31 Table 3: Team Members Roles and Expertise ...... 35 Table 4: Sheath Roughness Test Data ...... 40 Table 5: Most Viable Solutions for the VGCS ...... 50 Table 6: Uncertainties that Needed Resolved and Corresponding Sensors...... 52 Table 7: Ice Accumulation Weather Conditions ...... 56 Table 8: Ice Falling Weather Conditions ...... 57 Table 9 Weather Conditions for February 20, 2011 (Kumpf et. al, Weather Underground, 2011) ...... 61 Table 10: Interstice Temperature February 23 ...... 65 Table 11: Weather conditions for February 24, 2011 (Kumpf et. al, Weather Underground , 2011) ...... 68 Table 12: Ice Accumulation Criteria ...... 74 Table 13: Ice Fall Criteria ...... 74 Table 14: Sensor System at RWIS Stations ...... 81 Table 15: Airport Information ...... 82 Table 16: Distances of Weather Stations from VGCS ...... 83 Table 17: METAR and RWIS Precipitation Measurements for Ice Accumulation ...... 84 Table 18: Ice Accumulation Criteria ...... 85 Table 19: METAR and RWIS Precipitation Measurements for Ice Shedding ...... 85 Table 20: Ice Shedding Criteria ...... 86 Table 21: Final Ice Accumulation/Shedding Criteria ...... 86 Table 22: Weather Station Weights ...... 89 Table 23: Dial States Explanation ...... 92 Table 24: Tools Used To Design Dashboard ...... 104 Table 25: Dates for Past Ice Events that were Tested ...... 105 Table 26:Weather Statistics for December 12, 2007 Ice Event ...... 105 Table 27: Summary of Events when Ice Accumulation occurred in 2011 ...... 106 Table 28: Interstice Temperature on February 23, 2011 ...... 111 Table 29: Station Comparison for the 2011 Winter ...... 116 Table 30: Overall Performance of Dashboard on Past Icing Events ...... 124 Table 31: Comparison of readings taken by all 3 methods ...... 133 Table 32: New Stay Thermistors List ...... 138 Table 33: Sky Cover and Precipitation During the Period ...... 141 Table 34: Weather Report on March 15 ...... 142 Table 35: Wetness Test ...... 146 Table 36: Impurity Test ...... 147 Table 37: Impurity Test ...... 147 Table 38: Rain Bucket Lab Experiment 1 with 5 Minute Sampling Rate ...... 155 Table 39: Rain Bucket Lab Experiment 2 with 30 Minute Sampling Rate...... 156 Table 40: Caliper Test ...... 160 Table 41: Icing Sensors Initial Observations ...... 161 Table 42: Approximated ice thickness comparison of coatings and droplet sizes ...... 199 Table 43: Event History (February 16, 2013) ...... 200
19 Table 44: Event History (February 20-21, 2013) ...... 206 Table 45: Summary of VGCS Sensor Installation Trip ...... 217 Table 46: Ice Accumulation Station Functions ...... 223 Table 47: Ice Fall Station Functions in algorithm ...... 226 Table 48: Chronology of winter 2013/2014 icing event triggers ...... 236 Table 49: Web Report Tool: Sample Icing Events and Comments, December 2013 .. 243
20 Chapter 1: Introduction Section 1.1: Bridge Background The Veteran’s Glass City Skyway (VGCS), formerly known as the Maumee River Crossing is a large cable-stayed bridge on Interstate 280 that crosses over the Maumee River in Toledo, Ohio. The VGCS is owned by the Ohio Department of Transportation and is considered as the most expensive project ever undertaken by ODOT (Wikipedia, 2013). The construction began in 2001 and the bridge was opened for service in July 2007. The entire project consists of 8,800 feet of approaches and main span. The main span is approximately 1,225-feet in length, consists of a single pylon that rises 216 feet above the bridge deck, and has a single plane of stays, seen in Figure 1 below. The VGCS, which is an important connector for multimodal transpoortation and economic development, carries three lanes of traffic and has thousands of vehicles crossing daily.
Figure 1: Veteran’s Glass City Skyway (photo credit wiill be provided) The VGCS has several novel features: the cradle system for the stays, the stainless steel stay sheathes, and the illuminated glass in thee pylon. The VGCS is one of two installations of a new cradle system (Figg, 2005). This particular system eliminates the need for cable anchorage in the pylon by carrying stays from one side of the bridge deck to the other. This allows the tower to be more slender than what is possible with a conventional anchorage arrangement. The pylon is illuminated with internal LED lighting that is infinitely variable, an example of one lighting schemes can be seen in Figure 2 below. This makes the pylon visible for miles at night and the pylon can be lit to reflecct local events or the time of the year.
21
Figure 2: Veteran’ Glass City Skyway’s Illuminated Glass Pylon (ODOT, 2010) Under some winter conditions, ice forms on the stay cables of the VGCS. Ice accumulation can exceed a 1/2 inch and may persist for several days on the stays. Ice then sheds in semi-cylindrical sheets from the cable sheaths. Shedding of an individual stay can occur in less than a minute. Ice shedding is triggered by a combination of rising temperatures and solar radiation. The sheets may fall over two hundred and fifty feet to the roadway. Due to their aerodynamic shape, they can glide or be blown across several lanes of traffic. In some instances, large ice sheets have crossed all the lanes of traffic and fallen in the river. The potential of falling sheets typically requires lane or bridge closure for the duration of the ice persistence. Lane closures result in the inconvenience to the traveling public as well as loss to economic activities. Falling ice is a safety hazard to motorists and determining ice presence remotely is problematic. Currently, ice presence is determined manually, putting ODOT personnel in harm’s way. Figures 3 and 4 show ice accumulation on the stays and pylon of VGCS in the 2011 icing event.
22
Figure 3: Ice Accumulation on the East Side of VGCS (Baker, 2007)
Figure 4: Ice on the Pylon and the VGCS Glass Figure 5, which was captured during 2011 ice fall event, shows a large piece of ice, circled in red, falling into the third lane of traffic while vehicles are still travelling over the bridge (Belknap, 2011).
23
Figure 5: Large Piece of Ice Almost Hitting a Car Section 1.2: Summary of Goals and Objectives After four icing events in the first two winter seasons of the VGCCS being open, reseaarch was undertaken to assist ODOT in implementing an icing management procedure for the VGCS. The research followed a phased approach. The first phase focused on review of available technologies, selection of potential technologies for the VGCS and costing of the potential technologies whereas the second phase focused on the development and implementation of a monitoring system and sensorss. The first phase objectives were:
Identify available technologies and procedures that coulld potentially solve the icing problem.
Assess the state of the art via literature review and consultation with icing experts.
Examine the advantages, disadvantages, and applicability of each identified technology on the VGCS.
For each viable solution, develop a detailed description of the implementation, define required validation testing, perform a bbenefit/cost analysis, develop a budget for implementation and define a time frame for implementation.
Develop a real-time icing condition monitor. This objective was added by the research team in response to a request by ODOT to make the research immediately actionable by the bridge operators. As the project progressed, as phase II was undertaken. Phase II built on the background developed in the first phase. The objectives for Phase II altered to account for the knowledge gained concerning the state of the art and practice in anti/deicing and to address the need to better understand the microclimate on the bridge. The Final Phase II objectives were:
24 Collect data to resolve uncertainties in the bridge microclimate and the conditions on the stays. To understand the icing behavior it was necessary to gain knowledge about how and when ice was forming on the stays, stay sheath temperatures and the local conditions on the bridge,
Make a recommendation on two to four viable active solutions. This required experiments on anti/deicing techniques as well as literature review and discussion with experts.
Improve the user friendliness, algorithms and error handling of the icing monitor.
Develop of an ice presence and state sensor. No such commercial senor exists and data about the ice persistence and water flow beneath the ice is essential to understanding shedding. Through experimentation, no practical active or passive anti/deicing solution was ever identified, as discussed in Chapter 7 of this report. This ultimately led to a new overall objective, which was to improve the monitoring of icing events in order to provide ODOT with the best information to manage their response to an icing event. The goals, objectives, and uncertainties will be provided in more detail in the following chapter. Section 1.3: Summary of Results Past icing events were reviewed, the mechanisms for icing where explored, and the basic conditions that are favorable to icing accretion and shedding were ascertained. Historically, roughly two icing events occur each year. Icing on the VGCS occurs when there is general icing in the area. There have been five major icing events on the VGCS. The last of which was in February 2011. Conditions are favorable for ice accretion when one of the following conditions occurs: i. Precipitation with air temperature at the bridge below 32o F, or ii. Fog with air temperature at the bridge below 32o F, or iii. Snow with air temperature at the bridge above 32o F. The ice accretion rate is generally slow because during an ice storm precipitation rates are low and much of the water runs off the stays. Once the ice accretes on the stays and pylon, it may persist until shedding conditions occur. Temperatures above 32o F and/or solar radiation cause ice fall. Water flowing beneath the ice layer was observed prior to the ice fall in 2011 and is thought to be a precursor to ice fall. If there is ice on the stay, the weather conditions that cause ice fall are: i. Air temperature above 32o F (warm air), or ii. Clear sky during daylight (solar radiation). Given the unique features of the VGCS, the paucity of literature directly on point, and the urgency of addressing the problem, an expert team was selected to address this problem. The research team that had expertise in icing, icing instrumentation, icing test facilities, the VGCS construction and VGCS instrumentation was formed to address the
25 issues of ice prevention and mitigation on the VGCS. A comprehensive review all anti/deicing technologies that could be identified regardless of their technology readiness level was performed. A matrix of over 70 potential technologies was developed. The matrix describes the advantages and disadvantages of each technology. To simulate icing events and use a test bed for experiments an icing field station was designed and built. It had three full scale sheath specimens ten feet long. One of these specimens included strand. The station had a local weather station and a wireless data acquisition. The initial set of experiments verified that ice accretion and shedding similar to that which occurs on the bridge could be replicated. The icing station was then used for experiments on anti/decing chemicals, anti-icing coating, heat for anti-icing and deicing, and tests of instruments. The technologies that were the most viable were identified. They were: i. Deicing/anti-icing chemicals which would not present a biohazard when leached into the river such a sodium chloride; agricultural products, such as beet based deicers, and calcium chloride ii. Anti-icing coatings iii. Heat. The VGCS stays are mostly hollow so there is a potential to internally heat the stays. Experiments to evaluate the efficacy of each viable technology were carried out. The anti-icing chemical experiments showed that on the stainless steel surface of the sheath the chemicals tested did not persist. The deicing experiments showed that the chemical tested was not viscous enough to sheet across the sheath surface. These results are consistent with the results in the literature. In addition, to not performing the desired anti/deicing functions, chemicals would require a distribution system so they were deemed impractical. Several anti-icing coating were tested in the icing wind tunnel and at the icing experiment station. The coatings did not significantly delay the onset of ice, which stuck to the stay specimens and most did not change the shape of ice that shed. The coating that was outdoors for an extended duration of time became opaque and gummy, therefore, it would alter the appearance of the stays. These results are consistent with the results in the literature. Additionally, coating would be difficult to apply so they were deemed impractical. Introductory heating experiments were carried out at the icing experiment station. The heating was effective at deicing and partially effective at anti-icing. The requirement to heat each stay would require an expensive heating system. At that point, heating was deemed impractical so no advanced experiments or thermal analyses were conducted. Thus, no active or passive system was identified which had sufficient level of promise to justify detailed estimates of installation, operation or maintenance costs. When it was judged that the regional weather information and the RWIS did not provide enough information to assess the microclimate and icing behavior, a local weather station was installed on the bridge. The combination of the existing sensors and the
26 local weather station gives a good picture of the conditions on the bridge. Prior to deployment in the field, experiments on the sheathing specimens at the field station and in the laboratory coupled with the literature review lead to the conclusion that the proposed sensors functioned as desired and they were recommended for installation. To make the research immediately actionable by ODOT operations, a real-time icing condition monitor was developed. The research team designed a real-time monitoring system to track icing conditions on the bridge with a straightforward interface so information on the icing of the bridge was readily available to the bridge operators. This monitoring system is referred to as the “icing dashboard” or simply “the dashboard” because the information necessary to support ODOT operations is presented on one simple visual display. When conditions favorable to icing occur the dashboard alerted the research team. If the conditions favorable to icing persisted, ODOT was notified and, as required, requests for verification of ice accretion were made. The basis of this monitoring system is the smart mix of the automated algorithm and the visual observations, which helped aid in training the system for more optimal performance. The system uses an intelligent decision making process based upon initial criteria from past weather data analysis with parameter adjustments made after visual observations. Dashboard has done well in detecting ice accumulation each time, but the analysis done on the algorithm results and onsite observations from research team members and ODOT have been used to refine the algorithm as well as the interface. The dashboard has proven to be a valuable resource for the bridge operators as well as a valuable tool for reviewing weather events. The automated ice detection and monitoring dashboard for the VGCS was developed, implemented, successfully tested, and has been transferred to ODOT. No suitable sensor to detect the continued presence of ice or the transition from ice to water exists. Therefore, development and field testing of a suitable sensor were undertaken. The resistance based sensor detects the presence of ice and can differentiate between ice and liquid water. The sensor is designed to be mounted on the sheath and can detect the layer of water which forms beneath the ice just prior to shedding. The sensor has been tested in the laboratory and at the icing experiment station. The transition of the dashboard to District Two has concluded. A local standalone computer with the dashboard on it has been provided to the District. The standalone version maintains the basic functionality of the dashboard algorithms and alert system and provides links to the icing weather instrumentation on the bridge. A person at the computer can monitor the conditions on the bridge and determine the causes of alerts. Section 1.4: Organization of this Report Chapter one described background information regarding the VGCS, introduced the problem statement of helping ODOT operations with icing problems on the VGCS, and gave summaries of goals and objectives as well as results.
27 Chapter two discusses goals, objectives and benefits as well as introduces the expert team. Chapter three describes phase I research, which involved investigating the VGCS stay sheaths, performing a literature review regarding icing events on other structures as well as potential anti/deicing technologies, constructing a technology matrix to narrow down the numerous technologies to a few viable ones was constructed, and providing the history of sensor presence on the VGCS. Chapter four looks into the basic weather that gives rise to ice storms, the VGCS’s weather history, lessons learned from previous icing events, and accretion and shedding algorithms. Chapter five thoroughly discusses the development and testing of the icing dashboard as well as its initial results. Chapter six looks into each of new sensors implemented onto the bridge as well as describes both the laboratory and field tests performed on the new sensors. Chapter seven discusses the experimental studies performed on the sheath specimens at the outdoor icing experiment station located at the University of Toledo’s Scott Park Campus. This chapter gives detailed analysis and discussion regarding the potential technologies tested as well as the new sensors that were eventually implemented. Chapter eight describes the design and implementation of the local weather tower on the VGCS. Chapter nine discusses the development of the University of Toledo ice presence and state sensor. Chapter ten looks into the transition as well as the near-term and long-term maintenance of the icing dashboard. Chapter eleven provides a conclusion and recommendations for future work.
28 Chapter 2: Goals, Objectives, Research Approach and Benefits Section 2.1: Overview of Chapter This chapter describes the overall goals of the project, the objectives that were achieved to reach those goals, the approach that was taken to reach the objectives and the benefits that accrued to ODOT from this project achieving its goals. Section 2.2: Goal Under some winter conditions, ice forms on the cables stays of the VGCS. Ice accumulations have been observed at a thickness of 3/4”. The ice accumulation depends on the temperature, precipitation and duration of the storm. The accreted ice conforms to the cylindrical shape of the stay sheath. Thus, as the stays warm, the ice sheds in curved sheets. These curved sheets of ice then fall up to two hundred and fifty feet to the roadway below and may be blown across several lanes of the bridge deck depending on wind conditions and/or ice sheet aerodynamics. The falling ice sheets require lane closures and could present a potential hazard to the traveling public. The overall goal of this research was to assist ODOT in implementing an icing management procedure for the VGCS. This procedure may be active, passive or administrative. Active procedures involve anti/deicing measures that are typically powered and activated only when needed. Passive procedures operate without power and are continuously available, and include coatings or other technologies that are permanently in place. Administrative procedures focus on obtaining information about the condition of ice on the stays and pylon and managing the response to icing incidents with or without taking anti/deicing measures. Section 2.3: Objectives The research followed a phased approach. The first phase focused on review of available technologies, selection of potential technologies for the VGCS and costing of the potential technologies. The second phase focused on the development and implementation of a monitoring system and sensors. The original objectives of this study included the conceptual design and rough costing of three to five reasonable options, which included active or passive anti-icing or deicing approaches applicable to the VGCS, for ODOT. Investigation of a wide range of technologies was completed. No practical anti/deicing technology was identified. Therefore, the objective shifted to the monitoring of icing events in order to provide ODOT with the best information to manage their response to an icing event. The original objectives as well as the modification of objectives will be described below. The initial overall objectives of this study were to present three to five reasonable options to ODOT for ice protection on the VGCS as mentioned above. The highest priority was to identify cost effective methods to prevent the formation of ice on the stays. If suitable methods for ice prevention were not identified, the secondary objective was to identify methods to safely and efficiently remove ice from the stays without damaging the structure or causing additional safety concerns and delays to the public.
29 The first phase objectives were as follows: 1) Identify available technologies and procedures that could be used to solve the icing problem. Sixteen potential technologies were identified. Fourteen ice protection technology categories are acknowledged for anti-icing, deicing, and ice detection in the work by Ryerson (Ryerson 2009). There are many technologies from Ryerson’s work that are potentially applicable to the VGCS cables, which include: chemicals; icephobic coatings; structure design; expulsive techniques; heat; high-volume water, air, and steam; infrared energy; piezoelectric methods; pneumatic boots; vibration and appropriate ice detection methods. Proprietary methods such as pulse electro-thermal de-icing (PETD), a technique incorporating nano-fibers and a piezoelectric system proposed for aircraft will also be considered (Petrenko 2009; Prybyla 2009, and Near 2009, respectively).
2) Assess the state of the art through a literature review and consultation with the icing experts. Given the unique features of the VGCS, the paucity of literature directly on point, and the urgency of addressing the problem, an expert team is a superior way to quickly gain familiarity with the state of the art as well as define testing procedures and identify available facilities.
3) Examine the advantages, disadvantages, and potential applicability of each identified technology on the VGCS.
4) Identify the most viable solutions. It is expected that the most practical solutions will be novel adaptions or combinations of existing solutions.
5) For each viable solution, develop a detailed description of the implementation, define required validation testing, (either in situ or offsite), perform a benefit/cost analysis, develop a budget for implementation and define a time frame for implementation. Because we expect that the solutions will be novel, it is anticipated that some validation testing will be required.
6) Issue an interim report providing a summary of the findings from steps 1 through 4 and the recommendations and economic analysis from step 5 (Nims, 2011).
The research from Phase I resulted in the identification of several viable technologies, which can be seen in Table 1. These technologies fell into three separate categories, which were chemical distribution, chemicals, and internal heating. The technologies deemed viable for chemical distribution included the use of drip tubes or cable climbers with supply hoses or tanks. The chemicals that were further investigated were sodium chloride, calcium chloride, and agricultural products. As for internal heating, forced air, air with piccolo tube, steam heating elements and electrical heating elements were considered.
30 Table 1: Viable Technologies Category Specific Technology Chemical Drip Tube Distribution Cable climber with supply hose or tank Sodium Chloride, Calcium Chloride, Agricultural-based deicing Chemicals products Potential options to be explored are: forced air, air with piccolo tube, Internal Heating steam heating element and electrical heating elements
As part of Phase I, any proposed implementation was investigated in such a way that the implementation would be as “green” as possible. If any of the potentially viable solutions identified above in 5) required the use of a local power source, then cleaner alternative forms of energy, such as solar power, was investigated and utilized if possible. If the recommendation involved the application of chemicals, then the potential environmental consequences were considered and avoided if possible. At an icing team meeting during Phase I work (the meeting notes are in the interim report (add cite)), it was identified that there was insufficient information concerning the ice accumulation conditions, the ice shedding conditions, the microclimate of the bridge and the effectiveness of the viable technologies to reasonably cost alternatives. Thus, the team and ODOT decided that the uncertainties listed in Table 2 needed to be resolved.
Table 2: Information Required to Revolve Uncertainties Required Information Uncertainties to be resolved Presence of ice and/or It is difficult to be certain when ice is forming on the stay, how liquid water on stay fast it is accumulating and if it is persisting. Stay Sheath The temperature of the stays during an icing event is unknown. Temperature It is considered as one of the reasons for shedding. Solar radiation may contribute to ice shed. Solar radiation Sky Solar Radiation raises the stay temperature and the temperature between the ice sheet and the sheath. The bridge has its own microclimate: precipitation amount and Local Weather type, droplet size, wind speed, wind direction, visibility needs to Conditions be determined on the bridge. Characteristics of the distribution of the heat along the stay from Heat flow along stay air flow and through the stay cross section from a local source, and across a stay and the VGCS specific constants for thermal analysis, need to section be determined. The efficacy of the chemicals, the effect of the chemicals on the Efficacy of anti/deicing brushed surface of sheaths, and a practical method for applying chemicals the chemicals are unknown. Visual record of Observation of the unquantifiable aspects of icing on the VGCS. conditions Aerodynamic effects of A drip tube is a possible chemical distribution system. How the drip tube drip tube effects the aerodynamics of the stays.
31 In response to a request by ODOT at a project progress meeting to make the research immediately actionable by ODOT operations, a real-time icing condition monitor was developed. This monitoring system is referred to as the “icing dashboard” or simply “the dashboard” because the information necessary to support ODOT operations is presented on one simple visual display. The need to resolve the uncertainties in Table 2 and build on the capabilities of the dashboard led to a modification of Phase II research, which was initially focused on the implementation of viable technologies. Final Phase II objectives were as follows: 1) Collect data to resolve uncertainties in Table 2. Some of the data may come from existing sensors while some of the data required new sensors (discussed later in this report), laboratory experiments and on-site observation. The collected information should be sufficient to allow accurate costing, resolve uncertainties to reduce the risk of deploying an icing strategy that does not work, and be useful for improving and updating the icing dashboard. The uncertainties to be resolved and the reason for resolving the uncertainty is listed in Table 2 above.
2) Make a recommendation on two to four viable active solutions. To make a decision on the viability on an active system, it is necessary to have a reasonable estimate of the cost and the practical implementation strategy.
3) Improve the icing dashboard. The dashboard tracks the icing events in a format that is easy to understand, is useful for managing icing incidents and archives data. Local condition data that is collected from the bridge will be used to increase algorithm intelligence and error handling. The improvements focused on the enhancement of the visual display, refinement of the accretion and shedding algorithms and incorporation of data for a local weather station on the bridge.
4) Development of an ice presence and state sensor. No suitable sensor exists. Therefore, development and field testing were undertaken.
5) Transition the dashboard board and local weather station to ODOT District 2 so that the functionality of the dashboard and the information from the icing sensors is available to the operators of the VGCS.
As with Phase I, any proposed implementation was to be as “green” as possible. If the recommended solution involved the application of chemicals, then the potential environmental consequences of the chemical waste stream were addressed and “green” alternatives for conventional chemicals were investigated and utilized. The experimentation of the viable technologies will be thoroughly discussed in chapter 7 of this report. Through experimentation, no practical active or passive anti/deicing solution was ever identified. This ultimately led to a new overall objective, which was to improve the monitoring of icing events in order to provide ODOT with the best
32 information to manage their response to an icing event. Section 2.4: Expert Team Approach to the Research Because of the unique nature of the problem, the need for a quick response and the specialized nature of the icing knowledge required, the VGCS icing problem has been attacked with an expert team. The primary requirement was a team of researchers who are experts in ice and professionals familiar with the bridge. These are supplemented by team members who are expert in instrumentation, “green” energy and “green” chemistry. The team includes national expertise in icing from the U.S. Army Cold Regions Research and Engineering Laboratory and the NASA Glenn Icing Branch, expertise on the VGCS design and instrumentation, and experts in green technology. This team will address the unique features of the VGCS stays and provide recommendations to the Ohio Department of Transportation for the most practical and cost effective ice sensing, anti- icing and deicing systems for the VGCS. An expert team was the best way to rapidly assess the state of the art. This approach allowed the research team to confirm that a practical solution for ice anti/deicing for the VGCS does not currently exists. The icing experts have identified the information that must be collected and understood to design an effective anti/deicing solution. The research team consists of the following members: Jeff Baker, P.E., Independent consultant who was formerly the construction manager for VGCS, familiar with all aspects of VGCS construction and operation; experience with VGCS icing incidents. Nabil Grace, Ph.D., College of Engineering Dean, University Distinguished Professor, Lawrence Technological University; Director, Center of Innovative Materials Research; director of the LTU Comprehensive Environmental Test Chamber which has large scale icing test capacity. Michael Gramza, P.E., ODOT lead, District Construction Engineer for District 2, and former construction project manager of the VGCS. Cyndee Gruden, P.E., Ph.D., Associate Professor of Civil Engineering, University of Toledo; environmental engineer with expertise in management of deicing waste streams. Art Helmicki, Ph.D., Professor, Department of Electrical and Computer Engineering, University of Cincinnati; Director, Applied Systems Research Lab, a designer of the data collection system for the VGCS; expertise in sensor and signal processing. Victor Hunt, Ph.D., Research Associate Professor, Department of Electrical and Computer Engineering, University of Cincinnati; expertise in bridge instrumentation, a designer of the existing VGCS instrumentation system. Kathleen Jones, U.S. Army Cold Regions Research and Engineering Laboratory, Expertise; expertise in static and dynamic loads on structures due to atmospheric
33 icing; leader of freezing rain survey team; wrote ice load section for ASCE7 Standard, Minimum Design Loads for Buildings and Other Structures. Richard Martinko, P.E., Director UT-University Transportation Center and Intermodal Transportation Institute; former deputy director of ODOT District 2, former assistant director of ODOT, and former ODOT project principal of all phases of the VCGS project. Cyril Masiulaniec, Ph.D., Late Associate Professor, University of Toledo, Department of Mechanical, Industrial and Manufacturing Engineering; expertise in icing and thermodynamics. Douglas Nims, Ph.D., P.E., PI of this project, Associate Professor of Civil Engineering, University of Toledo; instrumentation and structural study of the VGCS; management of engineering consulting and academic teams. Tsun-Ming “Terry” Ng, Ph.D., Professor, University of Toledo, Department of Mechanical, Industrial and Manufacturing Engineering; expertise in icing and sensor. Currently, working on a study of icing on wind turbine blades.. Andrew Reehorst, NASA Glenn Icing branch; expertise in icing sensors; experience with ice accumulation and icing test facilities. Charles Ryerson, Ph.D., U.S. Army Cold Regions Research and Engineering Laboratory, Manager of CRREL’s Icing Program, Deep; deep and broad experience with aircraft and structural icing. Familiar; familiar with icing test facilities. His 2009 study on off-shore facilities is similar to this VGCS study. Thomas Stuart, Ph.D., Professor of Electrical Engineering University of Toledo; expert in power, PI of an ODOT funded research study of a solar installation near to provide power to the VGCS site. Mario Vargas, Ph.D., NASA Glenn Icing Branch, lead. NASA Glenn has an icing wind tunnel and the researchers are familiar with the capabilities of icing test facilities. Ted Zoli, S.E., Vice President of HNTB, expertise in icing; an extensive history of working with icing issues including testing structures on Mount Washington. Currently, he is engaged on two other cable stayed bridges with icing issues.
34 Table 3: Team Members Roles and Expertise Icing Local Green Team member Brief Description of Primary Activity/Expertise Expert Knowledge Expert
Former construction manager for VGCS, familiar with all aspects of VGCS construction and operation, experience with Jeff Baker X icing incidents.
Lawrence Technological University (LTU). Director of a unique low velocity wind/ freezing/icing/rain/load testing Nabil Grace X facility.
Mike Gramza X ODOT lead, former project manager of VGCS, able to provide input on ODOT operation needs.
Cyndee Gruden X University of Toledo. Expertise in management of de‐icing chemicals
Kathleen Jones X CRREL, national icing expert, leader in icing risk, member and former chair of ASCE‐7 committee on icing
University of Cincinnati. Instrumented VGCS, expertise in instrumentation and testing, support for implementation Art Helmicki X and testing costing
University of Cincinnati. Instrumented VGCS, expertise in instrumentation and testing, support for implementation Victor Hunt X and testing costing
Rich Martinko X University of Toledo. Understanding of ODOT operations, administrative support
Cy Masiulaniec X Late of the University of Toledo. Icing expertise, lead in performing thermal analyses and experiments.
University of Toledo. Lead in administrative support. Instrumented VGCS, lead in developing background Doug Nims X information for alternative, support for thermal calculations, lead in report writing and costing.
Terry Ng X University of Toledo. Icing expertise, lead in sensor development and experiments.
Andy Reehorst X NASA Glenn, icing sensor expert
CRREL, national icing expert, recently completed oil platform study which is parallel to the present VGCS study, Charles Ryerson X familiar with other test facilities nationally
Tom Stuart X University of Toledo. A lead in the design of the VCGS solar array, expertise in power management
NASA Glenn lead, aircraft icing expert, intimately familiar with test facilities at NASA Glenn and familiar with other Mario Vargas X test facilities nationally,
HNTB. National icing expert, consultant on VGCS design and construction, experience with icing problems on existing Ted Zoli X X bridges
35 Section 2.5: Benefits The benefits accruing to the traveling public, operators of the VGCS, District 2 and ODOT in general include: benefits accruing to ODOT and D02 from the dashboard and database include. 1) Comprehensive review of existing active and passive technologies: With the support of team member Charles Ryerson and drawing extensively, on his studies of icing technology, all of the known anti/de-icing technologies were investigated. The included over 70 technologies and is described in the technology matrix summarized in this report and presented in detail in Belknap, 2011.
2) Comprehensive review of past weather events; Team member Kathy Jones reviewed the icing events in northwest Ohio for the past twenty years including the first four icing events on the bridge. Vehicles were damaged in at least two of the first four icing events. A summary of this work is presented in chapter 4.
3) Detailed study of the 2011 major icing event: This was the fifth major icing event on the bridge since its opening. The icing team was onsite from the initial rainfall through the icefall for the February 2011 major icing event. Pieces of ice several feet long and up to three-quarters of an inch thick fell. The bridge was closed for several hours. The team was able to capture video and images of the ice shedding that lead to increased understanding of the icing behavior. A summary of the study of the 2011 icing event is presented in chapter 4.
4) Ice accretion and shedding algorithms: The study of the past weather and icing events lead to quantitative guidelines concerning the weather conditions that made icing accretion and shedding likely. These guidelines form the core of the algorithms in the ice monitoring system implemented on the bridge.
5) Development and implementation of the dashboard: In response to ODOT’s request for a way to make the results of the research easily actionable by the operators of the bridge, a real-time monitoring system was implemented. Initially, the information from existing sensors on the bridge and in the surrounding region was feed into the ice accretion and shedding algorithms and the results displayed on a graphical user interface. This interface was design so it displayed information about the icing status of the bridge in a simple on screen format, much as the dashboard of a car is designed to put the information essential to the operation of the vehicle in a visually compact format. At present, the dashboard reflects information from the initial sensors as well as a bridge mounted weather station and camera as well as temperature sensors on the stay sheaths.
6) Design, installation and use of a local weather station on the bridge. When it was identified that the existing sensor system on the bridge and in the surrounding
36 regional area was not adequate to monitor the microclimate on the VGCS and icing conditions of the stays, it was decided that a suite of local sensors was required. The research team identified an array of commercially available sensors that could provide most of the required information. The team then procured the sensors. A weather tower with local sensors and a camera as well as stay as mounting brackets to attach thermistors directly to the sheath were designed and installed. The sensors were made operational and their data was incorporated into the dashboard.
7) Field and laboratory studies of anti-icing and de-icing technologies: Experiments on anti/deicing chemicals, anti-icing coating and anti/deicing application of a heating system were carried out on full scale sheath specimens at the icing experiment station and in the icing wind tunnel at the University of Toledo. These studies coupled with the literature review demonstrated that no existing technology was appropriate for anti/de-icing on the VGCS.
8) Development of an ice presence and state sensor: No commercial sensor for directly measuring the presence or state of ice on the sheath exists. An electrical resistance based sensor has been developed. The sensor detects the presence of ice and can detect the layer of water, which is a precursor to ice shedding between the ice and the sheath. This sensor has been tested in the icing wind tunnel and at the icing experiment station. It is ready for deployment.
9) Database: The dashboard collects a comprehensive set of from the regional and local sensors on the bridge. It records all the icing and shedding alerts, serves as a log for all the observations and has the capability of exporting and plotting the data. This provides a database than can be used for study of the icing behavior of the bridge.
10) Archive: In addition to the weather data, the dashboard serves as a repository of all references, reports, presentations and other documentation of this project. This allows convenient access to the information for ODOT and researchers.
Through this project, the safety of those crossing the bridge has improved, the understanding of icing events has advanced and a sensor array capable of ascertaining the state of icing on bridge has been installed. Section 2.6: Chapter Summary The overall goal of this research was to assist ODOT in implementing an icing management procedure for the VGCS.
The objectives to support this goal were to Evaluate the state of the art in anti/deicing technologies through literature review and experimentation
37 Install sensors to understand the microclimate and icing on the bridge Design a real-time monitoring system to track icing conditions on the bridge with a straightforward interface so information on the icing of the bridge was readily available to the bridge operators.
An expert team approach was followed. A team with local expertise in the VGCS and expertise in anti/decing was formed and carried out the tasks to achieve the objectives. The objectives to support this goal were to Evaluate the state of the art in anti/deicing technologies through literature review and experimentation Install sensors to understand the microclimate and icing on the bridge Design a real-time monitoring system to track icing conditions on the bridge with a straightforward interface so information on the icing of the bridge was readily available to the bridge operators.
The overall benefit is increased safety for the traveling public. The benefits of completing this project were: Comprehensive review of existing active and passive technologies. Identification that no existing technology was suitable for anti/deicing the VGCS. Comprehensive review of past weather events. Detailed study of the 2011 major icing event. Ice accretion and shedding algorithms. Making real time icing information about the bridge available to the bridge operators. Development and implementation of the dashboard. Design and installation of a local weather station on the bridge. Field and laboratory studies of anti-icing and de-icing technologies. Development of an ice presence and state sensor. Creation of an icing database. Creation of an information archive.
38 Chapter 3: Phase I Research The research performed as well as the findings for Phase I of the VGCS project will be presented in this chapter. Phase I research included the VGCS’s stay sheath analysis, literary review, the completion of a technology matrix, and the identification of uncertainties as well as sensors that will resolve them. The research performed in Phase I allowed for better understanding of icing events that have occurred on the VGCS as well as viable ice protection technologies. During Phase I research, a number of uncertainties pertaining to the microclimate of the VGCS were identified. These uncertainties must be resolved in order to provide a practical anti/deicing technology and/or better monitoring of icing events, thus, several sensors were proposed. The Phase II research is discussed in chapters 5 through 9. Section 3.1: VGCS Sheaths The stay sheaths of the VGCS are unique. Typical stay sheaths are high-density polyethylene (HDPE). However, for the VCGS, stainless steel sheaths were chosen over HDPE due to their low life cycle cost. The brushed stainless steel surface also is an aesthetic enhancement for this signature bridge. The reflection of the light off the stays provides a unique appearance and enhances the effect of the illuminated pylon. Some characteristics of the stays sheaths, such as being made of a 1/8 inch thick 316L stainless steel, a brushed finish, and having a larger than typical diameter, may contribute to the icing problem the VGCS is experiencing. The stainless steel has a brushed finish. A Bendix Profilometer Peak Counter was used to determine the surface roughness on a sample piece of the VGCS sheath. The surface roughness could be a factor for ice clinging to the stays. After a team visit to the NASA Glen Icing Branch, it was determined that the sheath was roughly comparable to the smoothness of an aircraft, and therefore, did not facilitate icing. The test was run by The University of Toledo machine shop supervisor John Jaegly in the Material Science Lab room 1061 North Engineering. The machine uses a carbide tip moved across a surface. Table 4 shows readings taken across the grain of the brush finish and with the grain. Multiple readings were taken on multiple areas. Readings are in mirco-meters.
39
Table 4: Sheath Roughness Test Data Across the Grain With the Grain Test Area Test Area Test Area Test Area Test Area Test Area
1 2 3 1 2 3 1.269 1.212* 1.372* 0.228* 0.353* 0.470* 1.282 1.185* 1.347* 0.229 0.397* 0.485* 1.289 1.157 1.315* 0.231 0.405 0.507 1.146 1.345 0.218 0.409 0.503 1.150 1.374 0.206* 0.427 0.503 1.173* 1.367 0.381* 0.506* 1.171* 1.347* 0.390* 0.499* 1.373* 0.401*
Average Average Average Average Average Average 1.280 1.151 1.362 0.226 0.413 0.504
Avg. of 3 1.264 Avg. of 3 0.381 Note: * measurements are not included in the calculation of the average because the apparatus was being moved during the readings.
Section 3.2: Literature Review Icing is a worldwide problem for large bridges and other industrial facilities in cold climates; therefore, a broad literature regarding both structures that have been affected and anti-icing/deicing technologies were reviewed. This section will first discuss known icing events that have been found in literature, then anti-icing/deicing technologies found in literature, and finally the technology matrix for the Veterans Glass City Skyway. Section 3.2.1 Known Icing Problems on Other Bridges Leonard P. Zakim Bunker Hill Bridge: This particular bridge is an A-type cable stayed bridge that crosses the Charles River in Boston, Massachusetts. In March of 2005, the Boston area experienced winter conditions that caused the cables sheaths of the Leonard P. Zakim Bunker Hill Bridge to ice. The ice then fell off of the stays in large sheets and onto the roadway below. Officials and design engineer considered this weather to be a “fluke,” thus, no technology was investigated (Daniel, 2005) Penobscot Narrows Bridge: The Penobscot Narrows Bridge is an I-type cable stayed bridge that allows traffic to cross over the Penobscot River between Verona Island, ME and Prospect, ME (Penobscot Narrows Bridge and Observatory, 2014). The bridge was completed and opened in 2006 and experienced weather that caused icing for the first time in 2014. Due to the irregular occurrence of icing, the state DOT has taken an observation approach, thus, no technology is currently being investigated or deployed
40 (Gluckman, 2014) Port Mann Bridge: The Port Mann Bridge is an A-type cable stayed bridge that allows traffic to cross the Fraser River in Vancouver, B.C.. The bridge was opened to traffic in 2012 and experienced winter conditions that resulted in the accretion and shedding of “wet” snow in December 2012. Several technologies have been investigated, which includes: heating of the stays, the use of water, the use of a helicopter, cable collars, coatings, chemicals, sensors, etc. Currently, cable collars have been deployed on several stays (Meiszner, 2013). They have been relatively successful, but at times they get stuck on their way down the stays. In addition to the cable collars, a dashboard has been set up in order to provide real-time conditions on the bridge. Ravenel Bridge: The Ravenel Bridge is an A-type cable stayed bridge that connects Mt Pleasant, SC and peninsular downtown Charleston SC. In late January 2014, the bridge experienced weather that caused the cable stays to ice. Once the stays warmed up, ice began to shed causing damage to numerous vehicles passing below. There are reports of ice sheets as large as 8 to 10 feet falling from the cable sheaths (ABCNews4 WCIV- TV., 2014). Currently, there is not a technology deployed on this bridge. Severn Bridge: consisting of the Aust Viaduct, Severn Bridge, Beachley Viaduct, and Wye Bridge; stretches from England to Wales. The bridge was closed for ice falling off the stay cable on two occasions, February 6, 2009 and December 22, 2009 (Severn Bridge, 2011). Svinesund Bridge: The Svinesund Bridge connecting Norway and Sweden has an arc section of 188 meters (~617 feet). The single arc superstructure supports two-lane bridge decks on either side. To prevent ice formation on the arc during the winter, a temperature sensor controlled electric cable system was installed in the top section (Net Resources International, 2011). Southern Quebec, western New Brunswick, and eastern Ontario were covered with thick ice in 1998 due to a significant ice storm. Bridges and tunnels were closed because of the weight concerns as well as falling from superstructures (Countryman Electric). Uddevalla Bridge: The Uddevalla Bridge is an A-type cable stayed bridge located in Uddevalla, Sweden that crosses the Sunninge sound (Uddevalla Bridge in Sweden, 2012) .This bridge was opened to traffic in 2000 and has been experiencing frequent icing problems ever since (Bowers, 2014). Pulse electro-thermal de-icing (PETD) has been deployed on one cable and one pylon of the bridge for testing. Field testing, though successful, revealed a mechanical design flaw (Petrenko, 2011). This technology has been proven to be successful in icing conditions. In addition to PETD technology, the bridge consists of a large array of sensors that relay data back to a dashboard, which gives an early warning based on the microclimate of the bridge.
Section 3.2.2 Anti-Icing/Deicing Technologies found in literature Chemicals are considered as both deicing and anti-icing technologies. They are widely available commercially and are used in several industries such as aviation, off- shore oil industry, transportation departments, and marine structures. Chemicals have both dry
41 and wet applications. In anti-icing, chemicals are either used to reduce the adhesion strength between ice layer and the substrate or prevent the formation of ice. In deicing technology, chemicals are usually used to melt ice layers during or after ice storms. The most widely used chemicals are calcium chloride, magnesium chloride, potassium chloride, calcium magnesium acetate, urea, and agricultural based chemiccals. The major concerns regarding chemicals are environmental and corrosion issues as well as application persistence. Coating is another technology that is an active subject of development and testing. Coatings are a passive anti-icing technology, which means theey are applied to surfaces to reduce the adhesion strength of ice to the surfaces and to prevent the formation of ice on the surfaces (Ryerson, 2008). In Kulinich’s work, ice repellency of hydrophobic coating in different materials with different surface topographies was evaluated (Kulinich, 2011). A new approach of this technology would be the development of super hydrophobic material which causes water to bead into small drrooplets. Menini has developed a new coating with ice phobic characteristics on aluminum alloys which is widely used for several industries such as transmission lines, aircraft wings, and fuselage (Menini, 2011). Figure 6 shows water droplets on a surface which is covered by a superhydrophobic coating.
Figure 6: Application of Superhydrophobic Coating on the Surface (Ryerson, 2008) Concerns with coatings are efficacy in preventing of ice formation on the surfaces and their persistence through the winter. New designs, materials, and details can be considerred for new cable-stayed bridges in cold climates to prevent ice accumulation on their stays. Recently, innovative deicing technologies have been developed which use electricity as a technique to melt ice. These techniques cause ice to de-bond at the ice/substrate interface and then use external forces, such as gravity or wind drag, to remove the ice from the substrate. Electrical techniques have developed in three fundameental subcategories. These techniques include the following: 1) the application of a DC bias voltage to the ice/substrate interface, 2) pulse electro-thermal deicing, and 3) ice dielectric heating (Ryerson, 2008).
42 In the first technique, a small DC current is applied to the ice/substrate interface through conductors. The DC current, through electrolysis, ablates the ice into cavities filled with an oxygen and hydrogen gas mixture. The electrolysis-driven aablation reduces the contact area between the ice and substrate, and consequently reduces the adhesion strength. Figure 7 shows the presence of bubbles in the interface which helps in removing the ice. Experiments have demonstrated proof of concept, but the conductors are typically destroyed due to the high currents.
Figure 7: DC Bias Deicing where Electrolysis forms Bubbles (Ryerson, 2008) In pulse electro thermal deicing (PETD), a thin film conductor is used in the interface which melts the thinnest ice layer and external energy, wind or gravity, is used to remove the ice. Instead of heating the substrate continuously as done by most electrothermal deicing systems, the PETD rapidly heats the ice-subtrate interface for a few seconds with a high current, low voltage pulse. The rapid rise of temperature to only a few degrees warmer than freezing melts a thin layer of ice and causes ice debonding. Thhe short electrical pulse is the source of the energy savings. The best reference of that technology is a paper which was written by Petrenko (Petrenko, 2011). In this paper, Petrenko presents the PETD method, its theory, results of computer simulations, and extensive data from laboratory tests as well as several large-scale implementations. One of the promising features of this method would be the low energgy for deicing. Figure 8 shows a thin metal-foil heater which is used in the ice/substrate interface to melt the ice layer.
43
Figure 8: Pulse Electro Thermal Deicingg (PETD) (Ryerson, 2008) One of the implementations of PETD is the Uddevalla Bridge in Sweden (Petrenko, 2011). A system in current usage was installed on one cable, which is over 200m in length and 25cm in diameter, and one pylon. The system is no longer used on the Uddevalla Bridge. The third electrical technique is ice dielectric heating. That metthod uses high frequency excitation from 60 kHz to 200 kHz to melt an ice layer. Figure 9 shows an ice layer releasing from electrical transmission cable using the ice dielectric heating method.
Figure 9: Ice Being Released using Ice Dielectric Heating (Ryerson, 2008)
44 The next technology, which is considered as new teechnology, 20 years old, is electro- expulsive deicing systems (EEDS). EEDS uses a variety of technologies to create small amplitude and short duration mechanical pulses to remove the ice from the substrate. The most applicable design for cable stayed bridges is an electrically actuated system which was designed and developed by NASA Ames (Ryerson, 2008). There are several technologies which use heat as a method to deice or anti-ice structures. The most applicable ones are electrothermal, hot air, and water deicing. Electrothermal heating is using electrical resistance as a sourcce in either deicing or anti- icing. In electrothermal heating, heating of substrate occurs as a result electrical resistance in wires such as nichrome wires or carbon layers. Though this method is considered as an efficient heating system considering all of the energy conducted through the wires is converted to heat, its use of electrical energy is costly. An example of this method is the heating of automobile rear windows. Figure 10 shows an application of heating cables to prevent icing of hatches and bulk- head doors which is used by the Navy.
Figure 10: Navy Vertical Launch Systems with Electrically Heated Door Edges (Ryerson, 20008) Another source of heat for deicing and anti-icing is hot air. The example of this technology is the automobile windshield defroster. Hot air is used widely in the aviatioon industry. The U.S. Air Force has used jet engines mounted onto trucks as a source of energy to blow warm air across the wings of iced aircraft (Ryerson, 2008). Heated pressure washers and steam nozzles use hiigh pressure jets to remove ice from surfaces. The Navy considers this technology to be viable and less expensive for deiicing of ships (Ryerson, 2008). The efficiency of a hydraulic system is dependent to variety of factors such as: nozzle size, flow rate, wind, and distance. Though viable, steam is less available today than when steam power was common.
45 Another technology, which is considered unique, is iinfrared deiicing. This method is a kind of remote technology which heats the objects through absorption of infrared energy. There are still additional experimental and analytical needs to understand the deicing process using infrared heaters (Koeing, 2011). Figure 11 shows the use of electric infrared heaters to deice the door entrance at the Cold Regions Research Laboratory (CRREL) facility. Figure 12 shows the application of infrared deicing in the aviation industry.
Figure 11: Infrared Heaters above the CRREL Entrance (Ryerson, 2008)
Figure 12: Aviation Facility using Infrared Radiant System (Ryerson, 2008) Millimeter wave technology is another technology for deicing and detecting the presence of ice on surfaces. The ice naturally absorbs microwave energy and heats. That technology is applicable to any situations where water is available to absorb microwave energy and heat the surfaces (Ryerson, 2008). The topic of atmospheric icing on cables has been studied inteensively in the past. However, most of these studies were conducted for icing on to transmission and power lines, which behave differently than cable stay sheatths. Atmospheric icing is a general description for various phenomena which include: rimme, fog, freezing rain, and wet snow.
46 An enormous amount of research has been conducted to understand the nature of the problem and to predict the icing load on the structure. In order to accomplish this goal, numerous ice and snow models have been developed in order to predict the thickness of ice and/or snow as well as the weight associated to the ice and/or snow. By being able to predict the aforementioned values, it becomes possible to examine the effect of this load on man-made structures. Some scholars such as Makkonen (2010), Admirat (2008), Sakamoto (2000), Finstad (1988), and Nygaard (2013) have developed their own icing and snow models. Although they have developed different models, the significant parameters in the icing process are nearly the same for all of them. Additionally, all of the models follow the ISO standard equation for icing of structures (ISO12494 2001): Where,
α : Collision efficiency - the ratio of the droplets that hit the cable to the total number of droplets in the windward side (Dubach et al., 2005);
α : sticking efficiency - the ratio of the droplets that stick on the cable to the total number of droplets that hit the cable (Dobesch et al., 2005);
α : Accretion efficiency - a representation of the amount of the droplets that will freeze to the total number of droplets that hit the surface;
A : Cross-sectional area perpendicular to object;
V : Particle impact speed perpendicular to object;
w : Water content [mass concentration of the ice particles]. The major difference between each model lies in the parameters of collision, sticking, and accretion efficiency; all of which depend on the event itself. Therefore, given the above equation, the factors that affect the formation of ice include: wind speed, precipitation type, precipitation amount or visibility (visibility can be substituted if the amount is unknown), and the size of the cable. In addition, some other factors that can be used for detecting the right conditions for an icing event are temperature and humidity. In the future, the existing analytical studies could be advanced to address the torsional rigidity and geometry of the stays versus power lines. The stay specific models would be used primarily for forecasting ice accretion. The forecast weather data would be input in to the models and total accretion predicted. This forecasting of the ice accretion could be used to reduce the false alarms of the monitoring system. Several anti/deicing technologies have been tested as passive or active solutions such as deicing or anti-icing. These solutions prevent or diminish the accumulation of ice on the surface of a cable, or in this case a stay sheath. Two examples of these technologies
47 include internal heating systems and the super hydrophobic coatings, which were tested in the indoor icing tunnel and the outdoor experimentation station at the University of Toledo. These experiments will be discussed in more detail later in this report. Another method that was recently introduced by Couture (2011) utilized a photonic deicer to melt accumulated ice on power lines; this is displayed in Figure 13 below.
Figure 13: Photonic Deicer for Deicing of Power Lines (Couture, 2011) Cable collars are another technology that may be used in order to remove material such as snow and ice from cable stay sheaths. Cable collars work by using gravity to force them down the stay, ultimately removing the material, as seen in the video in the article by Meiszner (Meiszner, 2013). They are typically in the shape of a ring that has a significant amount of weight. This active technology has been used on the Port Mann Bridge to remove “wet” snow and has been relatively successful at doing so. Common issues associated with this technology include getting stuck in the process of removing the material from the stay, sliding over ice rather than removing ice, and the potential for damaging the sheath. Section 3.3: Technology Matrix The objective of literature review was to assemble a comprehensive list of all potential solutions for VGCS. To reach that goal, a technology matrix which has a description of the technology, a discussion of its advantages and disadvantages, a rough estimate of the cost of each technology, and the status was developed (Belknap, 2011, Ali 2013)). The technology matrix lists 75 potential technologies in 13 different categories. The categories of technologies include: 1- Chemical and chemical distribution: Anti- icing/deicing chemicals such as salt or agricultural base products and practical systems to distribute them.
48 2- Coatings: A layer applied to the surface of the sheath which prevents ice accretion on surfaces.
3- Design: Changing the shape of stay’s sheath to prevent ice accumulation.
4- Electrical deicing systems (electro-expulsive): Using repelling forces between conductors to produce an explosive force that ejects ice from the sheath.
5- Pneumatic expulsive deicing systems: That is the system which an inflatable boot covers the stays. When the boot is inflated, the ice cracks and falls down.
6- Heat: Use of thermal systems to prevent formation of ice or remove accumulated ice.
7- Infrared radiant heat: Use of radiant infrared heating to warm up the stays to prevent ice accumulation or remove accumulated ice.
8- Heating the ice-substrate interface: Applying heat directly to the interface between the ice and sheath. This reduces energy demand.
9- High-velocity water, air, or steam: Use of a high velocity stream of fluid to force the ice to fall off from stays.
10- Manual deicing methods: Chip or scrape ice off the stays.
11- Piezoelectric: Attach a piezoelectric actuator to the sheath surface to break the bond of ice to the stay causing shedding to occur.
12- Vibration or covers: Using vibration to break the ice-surface bond or covering the stays to prevent ice formation.
13- Ice detection: Sensors can monitor ice accumulation or detect the presence of ice.
Solving the icing problem of the VGCS is considered as applied research instead of basic research. Due to this fact, a technology selection meeting was held in June 2010. The notes for advantages and disadvantages of the technology matrix are available in a report which was submitted to ODOT, Innovation, Research and Implementation Section (Nims, 2010). Table 5 summarizes the most viable technologies which seem to apply for the icing problem of VGCS.
49 Table 5: Most Viable Solutions for the VGCS
Category Specific Technology Sodium Chloride Agricultural Products Chemicals Beet Heat Calcium Chloride Hydrobead Coating Internal Heating
Heat - Forced air - Air with piccolo tube
Section 3.4: Sensors on the VGCS A key advance from phase I to phase II was the addition of a local weather station on the bridge. Prior to the 2012 – 2013 icing season, the VGCS had several existing sources of which data could be collected to better understand the local weather. The existing sources included several local airports as well as an ODOT Road/Runway Weather Information System (RWIS). In the research team meeting at the end of phase I, it was agreed that the existing sensor array was not adequate to capture the microclimate of the bridge, which is needed knowledge for understanding the VGCS icing events. To improve data collection, thermistors were added at several location on the bridge in the fall of 2012 and a weather tower was added in the summer of 2013. The detail of the new sensors is in chapter 6 and a discussion of the weather history from the sensors is in chapter 4. Section 3.4.1: Sensors on the VGCS prior to the 2012 – 2013 Winter The main existing local source was the ODOT RWIS site 142, which was installed on the VGCS. The RWIS station on the VGCS included the following sensors which were used to understand the weather conditions prior to the 2012-2013 icing season: 1. Linux RPU – RWIS Elite weather station platform. A full-feature weather station capable of sensing a variety of road weather conditions, gathering traffic data and activating roadside devices. This platform supports a full range of atmospheric sensors as well as pavement temperature and condition sensors.
2. Wireless Pavement Sensor X6 – Collects traffic and weather information. A self-contained, in-pavement sensor that utilizes Vehicle Magnetic Imaging
50 (VMI) technology to detect vehicle count, speed and classification. In addition, the sensor measures pavement temperature and condition.
3. RM Young Ultrasonic Wind Sensor – Measures wind speed and direction.
4. RM Young Air Temp/Dew Point Sensor – Measures humidity and temperature. Has a special plate to block direct and reflected solar radiation while allowing air passage.
5. Weather Identifier and Visibility Sensor (WIVIS) – Precipitation Identifier/Classifier with Visibility. The Weather Information and Visibility Sensor (WIVIS) determine the type, intensity and rate of the precipitation that is occurring, as well as the visibility.
Additional weather information was also collected from Toledo Express Airport and the Toledo Executive Airport. Information from these sources included temperature, dewpoint, wind speed and direction, cloud cover and heights, visibility, barometric pressure, precipitation amount, lightning can be collected from airport stations (Agrawal, 2011). Section 3.4.2: Sensors added in 2012 – 2013 In October of 2012, an array of thermistors were installed on several of the VGCS stays. The installed thermistors allow a local measurement of stay temperature before, during, and after icing events to be obtained, which is vital to the understanding of ice accretion and shedding. The thermistors locations, uses, etc. is further explained in Chapter 6 of this report. Section 3.4.3: Sensors added in 2013 – 2014 In the past, the information on icing events was limited to direct observation on the bridge and/or restricted to data regarding weather conditions at the RWIS and airport stations. In order to help ODOT anticipate icing events, take the necessary action to inform the public in order to keep them safe, and to improve the performance of the dashboard for managing icing events, weather data pertaining to the microclimate of the VGCS was required. Table 6 summarizes the required information for managing upcoming icing events and the sensors that correspond to resolving the uncertainties.
51 Table 6: Uncertainties that Needed Resolved and Corresponding Sensors Required Uncertainties that need to be resolved Sensor Information It is difficult to be certain when ice accumulates on the stays except field observation. Ice Goodrich Ice Presence of Ice thickness also triggers the criteria in falling Detector conditions The temperature of the VGCS stays during icing Stay events is unknown. This temperature is Thermistors Temperature considered as one of the reasons for falling conditions
Solar radiation can cause the sheath surface Sky Solar temperatures to go above freezing even if the Sunshine Radiation ambient temperature is below freezing. This Sensor can trigger shedding of ice off the stays
The VGCS has its own climate. Type and Local Weather LWS/ Rain amount of precipitation, wind speed and Conditions Tipping Bucket direction need to be determined
Visual records of Observation of the stays condition during icing Camera icing conditions events can be valuable
Taking the above information into consideration, several sensors were requested by the research team in order to reach the aforementioned objectives. The sensors requested and brief descriptions are as follows: 1. Goodrich Ice Detector: The Goodrich ice detector detects ice accumulation on an ultrasonic axially vibrating tube. It also measures precipitation transitions between liquid and solid condition (Goodrich, 2009). One of the unique features of this sensor is that it differentiates rain from freezing rain.
2. Leaf Wetness Sensor (LWS): Leaf wetness is a parameter which is used to describe the amount of dew or precipitation left on the surface. The LWS has a potential to detect if water is liquid or frozen.
3. Sunshine Sensor: It has been observed that solar radiation on the VGCS stays is a condition that can trigger ice shedding. This particular sensor measures both global and diffuse radiation as well as sunshine duration. The
52 sunshine sensor uses photodiodes with a computer generated shading pattern for measuring solar radiation (Delta-T Devices, 2002).
4. Electrically Heated Rain and Snow Sensor – R. M. Young’s Model 52202 (Campbell Scientific): It is an electrically heated precipitation gage which provides year-round measurement of rain and snow. It has been observed that ice accumulation and ice persistence depend on the rate and amount of precipitation. This sensor also uses a wind screen to minimize the effect of wind on the rain measurement.
5. Weatherproof Camera: The goal of having active weatherproof camera on the self-supporting instrumentation tower is to observe the unquantifiable aspect of icing events and track the performance of the dashboard. Reviewable visual record of icing events gives valuable information before, during, and after storms.
The aforementioned sensors were installed in the summer of 2013 with the weather tower. They provide critical information pertaining to the microclimate of the VGCS, thus, aiding in the understanding of icing events. A more detailed description may be found in Chapter 6 of this report. Section 3.5: Chapter Summary This chapter thoroughly described phase I research. During phase I research, the VGCS stay sheaths were investigated, literature review regarding icing events on other structures as well as potential anti/deicing technologies was performed, a technology matrix to narrow down the numerous technologies to a few viable ones was constructed, and the history of sensor presence was given. The VGCS stay sheaths are very unique when compared to the typical HDPE sheath. The unique features of the stay sheaths include being made of a 1/8 inch thick 316L stainless steel, a brushed surface, and a larger than typical diameter. The stay characteristics were chosen for aesthetic reasons and due to the low life cycle costs of stainless steel. Initially, the brushed surface of the sheaths was thought to potentially increase ice adhesion. After a team visit to the NASA Glen Icing Branch, it was determined that the sheath was roughly comparable to the smoothness of an aircraft, and therefore, did not facilitate icing. A roughness test was performed, which confirmed this belief. Phase I literature review included icing events on structures, mostly bridges, from all around the world as well as potential anti/deicing technologies. The literature review showed that this is a relatively common problem for cable stayed bridges that are located in areas that experiences weather that gives rise to icing storms; this is described in detail in the following chapter. Additionally, it allowed for all potential technologies to be investigated and discussed in order to determine a viable technology to be used. However, it was determined that no such technology existed. Technologies that were investigated are as follows: chemicals, coatings, the application of a DC bias voltage to
53 the ice/substrate interface, pulse electro thermal deicing, ice dielectric heating, electro- expulsive deicing systems, electro thermal heat, hot air, water deicing, infrared deicing, millimeter wave technology, photonic deicer, and cable collars. A technology matrix was then assembled in to gather, organize, and describe all of the technologies. This technology matrix included 75 potential technologies that were separated into 13 different categories. These technologies were then narrowed down to the most viable technologies at a technology selection meeting in June, 2010, which can be seen in Table 5 above. The entire technology matrix can be seen in Appendix X. Additionally, at the technology selection meeting, it was determined that there was insufficient information concerning the ice accumulation conditions, the shedding conditions, the microclimate of the bridge and the effectiveness of the viable technologies. This ultimately led to the addition of sensors to the existing sensor array, which had been deemed inadequate. The existing array included the following sensors: Linux RPU – RWIS Elite weather station platform, Wireless Pavement Sensor X6, RM Young Ultrasonic Wind Sensor, RM Young Air Temp/Dew Point Sensor, Weather Identifier and Visibility Sensor, and a Tipping Bucket Rain Gauge. Thermistors were added prior to the 2012 – 2013 icing season and a Goodrich Ice Detector, Leaf Wetness Sensor, Sunshine Sensor, Electrically Heated Rain and Snow Sensor – RM Young’s Model 52202 (Campbell Scientific), and a weather proofing camera was added prior to the 2013-2014 icing season as part of the weather tower installation.
54 Chapter 4: Weather History, Modeling and Analysis Section 4.1: Introduction A brief look into the causes of ice storms as well as the history of icing events that have occurred on the VGCS will be presented in this chapter in order to provide a better understanding of icing events, nature of ice accretion and ice shedding, and what happens during ice storms. Since the VGCS was opened for service in July of 2007, there have been five major icing events that have occurred. Jones’ report describes the first four icing events and the weather conditions which preceded them (Jones, 2010). The last icing event, which occurred in February 2011, was directly observed by the research team. Since February 2011, there have been minor icing events that the Dashboard has collected data on. These minor icing events will also be discussed. The weather history and the observed icing behavior described in this chapter serves as the basis for the ice accretion and shedding algorithms in the Dashboard. Section 4.2: Description of the basic weather that gives rise to an ice storm Freezing rain has been the cause of four of the five icing events on VGCS. A prolonged freezing rain event requires a layer of cold (below-freezing) air at ground level, warmer air aloft, a high pressure system to hold the cold air in the place, and precipitation. The duration of a freezing rain event depends on how long the high pressure stays in the place and variations in the thickness of the cold air layer at that location. For a major icing event in Toledo, typically, the warm air aloft comes from the Gulf of Mexico and the cold air originates in Canada. Liquid precipitation in the warm air layer (that may have fallen from the clouds above as snow) supercools as it falls through the cold air layer at the surface. If that layer is not too thick the supercooled drops remain liquid. These freezing rain drops are likely to freeze when they impact a cold surface, primarily because of convective and evaporative cooling, with some contribution from supercooling. If the surface cold air layer is thicker the drops might freeze as they fall, forming ice pellets. These particles of ice accumulate on the ground, but are likely to bounce or slide off the bridge stays. On the cold side of a freezing rain event the precipitation is likely to be falling as snow, and on the warm side the precipitation is plain rain. As an event like this evolves, the precipitation type at one location may change between rain, freezing rain, ice pellets, and snow, sometimes with different types of precipitation occurring at the same time.
55
Figure 14: Damaging ice storm footprint map, 1946-2014 in the lower 48 states and portions of the lower tier of Canada. Section 4.3: VGCS Weather History Five icing events have occurred on the VGCS. Kathleen Jones, CRREL expert and research team member, has prepared a report that describes the first four icing events as well as the conditions which caused them (Jones, 2010). The last event, which occurred in February 2011, was documented by thee University of Toledo graduate students. The icing events and basic features are listed in Table 7 and Table 8.
Table 7: Ice Accumulation Weather Conditions
Ice Event Precipitation EEvent December 2007 Freezingg rain and Fog March 2008 Snow, rain, and fog December 2008 Snow and fog; freezing rain and fog January 2009 Freezing rain and fog February 2011 Freezing rain, cleaar
56 Table 8: Ice Falling Weather Conditions
Ice Event Ice Fall Weather
December 2007 Rain with temperature above freezing
March 2008 Sun with temperature above freezing
December 2008 Rain, gusty winds and temperatures above freezing
January 2009 Gusty winds, temperature above freezing
February 2011 Light wind, overcast, and temperature above freezing
It is also possible for wet snow to accumulate on the stays and cause an ice event. This wet snow, that includes both snowflakes and liquid water can accumulate on the stays. That means that an icing event can begin with the air temperature above freezing. Icing on the cables of VGCS may also occur in supercooled clouds or fog. The Liquid water content of the fog is inversely proportional to visibility. Fog droplets, with typical diameters of a few 10s of microns, have essentially zero terminal velocity, so move only when they are carried by the wind. Therefore, in cloud icing is likely to be significant only in a thick fog and high winds. To have a better understanding of icing events, nature of ice accretion and shedding, as well as what happens during ice storms, a brief summary and lessons learned from past icing events on the VGCS is presented below. Jones report describes the first four icing events and weather conditions that preceded them. The last icing event, February 2011, was captured and documented by the University of Toledo graduate students. December 2007: The data from Toledo Express Airport and Metcalf Field indicated freezing rain and fog occurred on December 9 - 10, which is believed to have caused ice accretion on the stays. Rainfall with temperatures above freezing triggered the ice shedding from stays, which took place on December 12. Ice shedding resulted in the closure of two out of three lanes of traffic as well as damaged vehicles. (Jones, 2010) March 2008: Weather data revealed that a snow and rain mixture with temperatures falling below freezing, concurrent with a fog, caused ice formation on the stays on the evening of March 27. The shedding of ice occurred in the afternoon of March 28. Clear
57 skies and air temperatures above freezing on March 28 were considered to be the shedding triggers. During the ice fall event, the center and left lanes in both directions had to be closed and at least one vehicle was damaged (Jones, 2010) December 2008: On December 17, ice was first observed on the stays. Data gathered from Toledo Express Airport, Metcalf Field, and Toledo Blade indicated that freezing rain, snow, and fog were the conditions that caused ice accretion on the stays. On December 24, ice shedding occurred with temperatures above freezing and gusty winds. It should be noted that ice persisted on the stays for 7 days. Throughout this event, the left and center lanes were closed for 5 days, starting on December 19. January 2009: Ice first was observed on January 3 and shedding occurred on January 13. The data from airport weather stations showed that freezing rain accompanied by fog caused ice accretion on the stays. Temperature rising above freezing and gusty winds on January 13 triggered the ice shedding from the stays. Note, the left lanes were closed until January 21, 2009. This could be due persistence of ice on the stays after the initial shedding event, which then melted over the next week (Jones, 2010). February 2011: This icing event was observed and recorded from the time of ice accretion, which started on the evening of February 20, to the time in which ice shedding occurred, which began on the morning of February 24. This event will be given as a day by day account in order to provide a clear and better understanding of icing on the VGCS. Researchers were on the bridge regularly throughout the event to capture photographs and video as well as properly document the behaviors between the accreted ice and the VGCS stay sheath. A detailed description of this event is in Belknap, 2011. Photos and video from this event are permanently archived by ODOT Research. The forecast for the night of February 20 was freezing rain, followed by a drop in temperature. On a local television station’s weather website (Storm Tracker 11, 2011), the forecasters predicted snow changing to freezing rain. The update for the overnight was scattered rain or freezing rain with additional ice accumulation. With low temperatures and precipitation, the conditions were conducive to ice accumulation. The ice dashboard was monitored that night by the research team. RWIS Station 582016, located on the bridge, was tracked watching both the readouts from the dashboard and the cameras on the stays. At 9 pm, the dashboard ticker was reporting Y3, indicating that the conditions had been conducive to ice accumulation for 6 hour and visual observation should be made to ascertain if ice had accumulated. At 10:29 PM, a researcher and the ODOT shift supervisor visited the bridge and confirmed the presence of ice. With this confirmation, the dashboard was put on alert status. Figure 15 shows a screenshot of the ice dashboard on February 21. The “Record of the last 48 hours” at the bottom of the dashboard shows the progression of conditions from “clear” to “alert”.
58
Figure 15: Dashboard readout for February 21, 2011 Also, at 9 pm, ODOT personnel and researchers obbserved ice on the stays. At 10:30pm, the ODOT shift supervisor and the researchers returrned to the bridge to observe the stays. At that time, the air temperature was 30° F and the precipitation was cold blowing rain and the prevailing wind was from the east. Rain droplets were of average size annd liquid water was clinging to or blowing off the stays. On the stay checked, the ice was 1/4 inch thick on top of the stay, about ½ inch thick on the east side of the stay, there were small icicles and the ice was about 1/2 in thick on the bottom of the stay, and on the west side the stay was bare except for occasional frozen rivulets. Clear ice was accumulaating. Clear ice with few air bubbles deposited by freezing rain or freezing drizzle is referred to as glaze ice. This is opposed to white rime ice with more air bubbles which results from super-cooled fog or cloud drops carried by the wind (Ryerson, 2011). There was no sign of melting - rain was running down and blowing off the icicles. The deck outside the parapet was covered with slushy ice. The parapet had crusty wet ice. Inside of the parapets on the median was an accumulation of wet ice that was 1/2 to 1 inch thick. The ice covered the eastern face of the cable. There was little to no ice on the western face. The conditions on the stays appeared uniform as far up the stay as visible. Figure 16 shows an overview of the ice accumulation and Figure 17 shows a close up of the ice accumulating on the east face of the stay. Figure 18 is a schematic of the distribution of the accumulated ice.
59
Figure 16: Overview of ice accreting on stay at 10:29 PM Sunday evening
Figure 17: Close up of ice accreting on stay at 10:29 PM Sunday evening
60
Figure 18: Stay cable diagram with ice accumulation Table 9 compares the I-280 bridge RWIS station weather readiing with the ice dashboard and Toledo Metcalf Field weather reading.
Table 9 Weather Conditions for February 20, 2011 (Kumpf et. al, Weather Underground, 2011) 20/Feb/2011 5-6 pm 6-7 pm 7-8 pm 8-9 pm 9-10 pm 10-11 pm RWIS Temp 30° F 30° F 30° F 30° F 30° F 30° F Precip None Rain Rain Rain Raain Rain Surface Wet Wet Wet Wet Wet Wet Surface Temp 30° F 30° F 30° F 30° F 30° F 30° F Dashboard Icing Conditions Clear Y1 Y2 Y3 Alert Alert Toledo Metcalf Temp 27° F 28° F 28° F 28° F 26° F 28° F
On February 21, 2011, Storm Tracker 11 (2011) recapped that 1/4” to 1/2” of ice accumulated Sunday. They called for an additional mix of snow, freezing rain, and sleet for the morning and 3-5 inches of snow in the evening. The wintery mix and snow fell throughout the day and accumulated on the stays. A layer of snow was between the ice already on the stay and the new accumulation in somme areas. ODOT placed barrels out at the inside shoulder. This allows the barrels to be qquickly reconfigured to close lanes. The research team stopped a total of four times, three times on main span and once on the back span. They chipped a hole in the ice to meeasure thickness at the main span near stay 6, which was inspected from the median. The ice on the east side was roughly ½ inch thick with closely spaced icicles on the bottom. Then the team looked for any variation in conditions along the length of the bridgee, however, they appeared to be consistent when viewed near stay 14 from the north bound side. Near stay 10 visual inspection from inside the truck on from the south bound side it was seen that below the
61 damper collar the ice appeared very thin or even possibly bare spots and above the damper collar ice appeared thicker with pronounced frozen rivulets. The team stopped one time on the back span near stay 8 where conditions were seen that roughly matched what was observed near stay 10. The wind was from the east as it has been throughout the storm. Generally, on the east side, the ice appeared to be thicker than on the west side. The ODOT supervisor felt the coating on the east side was thicker than he had seen before. On the west side, there were some spots that appeared to have a very thin coat below the damper collar. Above the damper collar, ice was thicker and the frozen rivulets appeared more pronounced than on the east side. The ice above the collar appeared uniform as high as it could be seen on both the east and west sides of the stay. However, it is impossible to discern anything more than gross icing further up than about mid-height. On February 22, 2011, the temperature started at 15° F and climbed to 21° F by mid- afternoon then dropped to the teens again at sundown (Weather Underground, 2011). Although the air temperature was below 32° F, the sun was out and the solar radiation was 575 Watts/m^2 at 1:35 pm. At this time, the researchers observed liquid water under the ice on the stays and ice fall due to solar radiation seemed imminent. However, the ice did not fall when due to the solar radiation and the liquid water refroze at the end of the day. The research team noted that water was dripping from the icicles on the bottom of the stays. There was ice covered snow in many spots along the stays. The ice cover was uniform from the bottom up to the pylon, as observed with binoculars. Cracking in the ice was observed on stay 4 and there was significant cracking of the ice at bottom of the sleeve. There was ice on the east face of the pylon glass, but there was not any ice on the west face of the pylon glass. ODOT moved the orange barrels out closing the inside lane. Figures 19 and 20 portray the ice accumulation on the stays for February 22. The eastern face was coated and the western face was bare except for frozen rivulets.
62
Figure 19: Ice Accumulation up east side of stay February 22, 2011
Figure 20: Frozen Rivulets and bare metal on the west side of stays February 22, 2011 On February 23, 2001, the Weather Underground (2011) archived the outside temperature starting at 19°F in the early morning and rising thrroough the day to 28° F in the late evening. No precipitation occurred and there was an overcast sky the entire day. Wind speed reached 9 mph mostly South to Southwest (Weather Underground, 2011).
63 The research team continued to monitor the conditions on the bbridge. Temperature readings and pictures were taken at intervals on the stays. Liquid water was seen under the ice again even though there was an overcast skky and the ambient temperatures in the mid-20’s. As the team checked later in the day, the ice was breaking off easily and liquid water was observed under the ice. The temperature in the interstice, gap between the ice accumulation and stay sheath, was measured throughout the da. A Fluke 561 Infrared Thermometer was used to measure surface temperatures and the temperatures in the interstice. The interstice temperature was measured with a type K contact thermocouple. The lead of the contact thermocouple is approximately 0.1 inch in diameter and during the day it would slide easily into the interstice. Sometimes the lead could slide in as far as 6-10 inches with little resistance. Figure 21 shows a measurement being taken. In Figure 22, the indicated temperature of the stay surface, measured with the infrared thermometer, is 27.0 F. Because of the high emissivity of the brushed aluminum surface, it is likely that the temperature of the stay surface is overestimated. In laboratory trials with reflective surfaces, the infrared thermometer consistently read higher than the actual temperature. The temperature indicated in the interstice, measured by the contact thermocouple, was 32 F. The lead for the contact thermocouple can be seen inseerted into the interstice. Table 10 shows the variation in the temperature in thhe throughout the day.
Figure 21: Thermocouple reading between ice and stay February 23, 2011
64
Figure 22: Thermocouple reading between the ice and stay February 24, 2011
Table 10: Interstice Temperature February 23 Time Interstice Air StayNote Temperature Temperature 8:15 am 24 F 21 F 20B No visible liquid water 8:50 am 24 F 21.9 F N/A No visible liquid water 9:20 am 24 F 21.9 F 20B 9:30 am 28 F 21.9 F 15B 9:45 am 26 F 21.9 F 11B Liquid water under ice 12:15 30 F 24.1 F 19B Liquid water under ice. Large pm pieces of ice broke free easily. 1:00 pm 31 F 24.1 F 20B 1:00 pm 32 F 24.1 F 19B 1:15 pm 35 F 24.1 F 18B Liquid water under ice. Sheets break free easily. 2:55 pm 32 F 25 F 20B Liquid water that had bled from under the ice refroze on the stay. 3:57 pm 32 F 27 F 20B 5:23 pm 31 F 27 F 19B Liquid water under ice. Sheets break free easily.
Even though the day was overcast, the UV radiation penetrated the clouds and passed through the clear glaze ice. It then warmed the surface of the sheath under the ice. The resulting infrared radiation could not escape the interstice, which resulted in a significant increase in the temperature within the interstice. Due to this dramatic temperature
65 increase in the interstice, a patch of ice was able to be removed easily. Figure 23 portrays the fragility of the ice when the interstice temperature is significantly increased. Note the liquid water on the surface of the stay. The contact thermocouple indicated temperature in the interstice was 36 F. This seems a bit unreasonable because the water in the gap was in close contact with the overlaying ice. However, water moving under the ice was sometimes visible and would run out from under the ice in small streams.
Figure 23: Cracking in ice from chipping away ice, February 23, 2011 At 5:23 pm stay 19B was examined closely and it was observeed that there was liquid water under most of the ice. Figure 24 shows the extent of the layer of water. Only the small section of ice at the bottom of the stay was frozen solidlyy to the sheath. Ice this precariously attached could easily be dislodged. All the research observations were made at ground level, so it is possible that the ice layer further up the stay was materially different. Figures 25 and 26 are schematics of the ice buildup and liquid water observed at different stays and times.
66
Figure 24: Section where ice was chipped away to take temperaturee readings February 23, 2011
Figure 25: Ice thickness measurements on back stay 19 February 23, 2011
67
Figure 26: Ice thickness measurements on back stay 19 February 23, 2011 On February 24, 2011 at 5am, a senior researcher was on the bridge making observations. The research team met on the bridge around 8am. Cameras were given to two junior researchers to capture the ice fall event. At 8:40am, the ice fall event began. The ice fell in large chunks, at random intervals and from random stays. Whether conditions during the shedding event were temperatures above freezing (32° F), overcast sky, and light wind. Traffic was stopped at 9:30am because pieces were falling close to vehicles in the outer lane as well as off the bridge and into the Maumee River. The video footage showing the actual lane closure is MOV04201 with the last semi-truck driving over the bridge at 5:28, film time. The ice continued to fall until 11am. By this time, 80-90% of the ice had already fallen. Some ice remained on the pylon glass as well as on the stays close to the pylon. Temperature at the time of the ice fall was above freezing (32° F), overcast, and light wind. The south bound lanes were opened before 11:30 am; the north bound lanes remained closed for several hours because of the ice on the pylon. Table 11 compares the I-280 bridge RWIS with the iice dashboard and Toledo Metcalf Field weather readings. Figure 27 show ice accumulation on the pylon glazing at the very top. Figure 28 show the bridge deck after a majority of the ice had fallen off.
Table 11: Weather conditions for February 24, 2011 (Kumpf et. al, Weather Underground , 2011) 24/Feb/2011 6-7 am 7-8 am 8-9 am 9-10 am 10-11 am RWIS 280 Bridge Temp 33° F 33° F 34° F 34° F 34° F Precip None None None None None Surface Wet Wet Wet Wet Wet Surface Temp 33° F 33° F 34° F 34° F 34° F Dashboard Icing Conditions R1 R2 R3 R3 R3 Toledo Metcalf Temp 33° F 33° F 34° F 34° F 34° F
68
Figure 27: Ice accumulation on pylon glazing Februaary 24, 2011
Figure 28: Ice on bridge deck after 80-90% had shed, February 24, 2011
69 In summary, ice shed from the stay cables was confirmed. Videos from the ice fall event documented the size and fall pattern of the ice. Large rectangular pieces of ice from stays 20B-16B and 20A-16A, roughly, had pieces making it to the north bound outer lane of traffic and even completely off the bridge. When the ice started accumulating on the stays Sunday night, barrels were set out one lane of traffic. As the days went by and the ice was predicted to fall, the barrels were moved out so that only one of three lanes of traffic could cross the bridge in the northbound direction and leaving two of the three lanes of traffic open in the southbound direction. This was decided because the ice had accreted most heavily on the eastern side of the stays and the typical wind direction throughout the persistence and shedding period was coming from the west. On Thursday February 24th, after about a half hour of ice falling, ODOT closed the lane behind stay 20B on the north bound side. Vehicles were routed off the bridge that could turn around. Once 80-90% of the ice was down, the vehicles remaining on the bridge were instructed to cross with caution. Once all the vehicles were off the bridge was closed until all the ice had fallen. The last ice to fall fell from the pylon glass in the early afternoon. Figure 29 shows the weather for the week of February 20, 2011. Figures 30, 31, and 32 graph the solar radiation for February 22, 23, and 24, 2011, respectively.
70
Figure 29: Weather Summary for the week of February 20, 2011 (Weather Underground, 2011)
71 Solar Radiation Feb. 22, 2011
700 600 500 400 Series1 300
Watts/m^2 200 100 0 1:29:00 AM 3:00:00 AM 4:10:00 AM 5:42:00 AM 7:31:00 AM 9:13:00 AM 1:22:00 PM 2:46:00 PM 4:00:00 PM 5:33:00 PM 6:57:00 PM 8:10:00 PM 9:34:00 PM 12:00:00 AM 10:25:00 AM 11:48:00 AM 10:54:00 PM Time
Figure 30: Solar radiation counts February 22, 2011
Solar Radiation Feb. 23, 2011
250
200
150 Series1 100 Watts/m^2 50
0 1:34:00 AM 2:52:00 AM 4:01:00 AM 5:30:00 AM 6:51:00 AM 8:08:00 AM 9:28:00 AM 1:31:00 PM 2:44:00 PM 4:05:00 PM 5:28:00 PM 6:38:00 PM 8:01:00 PM 9:13:00 PM 12:00:00 AM 10:32:00 AM 12:02:00 PM Time
Figure 31: Solar radiation counts February 23, 2011
72 Solar Radiation feb. 24, 2011
350 300
^ 250 200 Series1 150
Watts/ m 100 50 0 6:01:00 AM 6:44:00 AM 7:28:00 AM 8:10:00 AM 8:56:00 AM 9:48:00 AM 1:14:00 PM 1:50:00 PM 2:36:00 PM 3:16:00 PM 3:59:00 PM 4:55:00 PM 5:52:00 PM 6:26:00 PM 10:35:00 AM 11:17:00 AM 12:04:00 PM 12:37:00 PM Time
Figure 32: Solar radiation counts February 24, 2011 Section 4.4: Lessons Learned from Previous Icing Events There are several lessons to be learned from the five previous icing events. The first is how ice is accretes onto the stay sheaths. In all five events, accretion occurred in conditions where freezing rain and/or snow were present. These events are typically, followed by a sharp temperature drop and accompanied by fog. Additionally, it has been noted that in some events there is minimal precipitation, yet significant amounts of ice still accumulates, therefore it is possible that some of the ice that is accreted onto the stay sheaths comes from supercooled drizzle or cloud droplets (Jones, 2010). Another lesson to be learned is how the ice sheds off of the stay sheaths. In four of the five events, shedding which cleared the ice off the stays occurred when the air temperature warmed to above freezing and was accompanied by gusty winds, clear skies or sunshine, and rain. The exception was the January 2009 event. Then ice shed intermittently and partially. Although, this was true for the last major icing event, which was the only to be directly observed and documented, it was learned that the temperature doesn’t necessarily need to be above freezing in order for shedding to occur. This is due to a greenhouse effect in the interstice between the ice layer and sheath surface. It was observed that considerable water flowed under the ice layer at temperatures in the mid-20’s. This effect caused ice to be removed easily, thus, displaying the possibility that shedding could potentially occur at temperatures below freezing (32 F). The ice detector has proven to be sensitive and accurate for ice accretion. It may be possible to substitute
73 Section 4.5: Analysis The common weather conditions prior to the previous ice events on the bridge led to the development of criteria to use when checking for icing event conditions. Weather conditions (for at least 6 hours) that would likely cause Ice Accumulation: 1. Precipitation with air temperature at the bridge below 32o F, or 2. Fog with air temperature at the bridge below 32o F, or 3. Snow with air temperature at the bridge above 32o F. Weather conditions that would likely cause Ice Fall: 1. Air temperature above 32o F (warm air), or 2. Clear sky during daylight (solar radiation). In order to automate the process of predicting ice fall events, an algorithm was developed based upon the above criteria to evaluate weather data. The weather data collected consists of RWIS measurements and METAR data from the local airports. Taking these criteria and the available data into account, a specific set of criteria was developed for Ice Accumulation and Ice Fall (Tables 12 and 13).
Table 12: Ice Accumulation Criteria Source Condition Description METAR or Freezing (Air Temp. <= 32o F & Precipitation type RWIS Rain is Rain) OR (Precipitation type is Freezing Rain) METAR Freezing Air Temp. <= 32o F & Precipitation type Fog is Fog METAR or Wet Snow Air Temp. > 32o F & Precipitation type RWIS is Snow
Table 13: Ice Fall Criteria
Source Condition Description METAR Warm Air Air Temp. >= 32o F METAR or Clear Sky Condition type is RWIS Clear
The data sources used, including the secondary sources used for redundancy, are: METAR Data Sources 1. http://www.wunderground.com/history/airport/KTOL/2011/05/13/DailyHistory.html ?format=1 2. http://www.wunderground.com/history/airport/KTDZ/2011/05/13/DailyHistory.html ?format=1
74 3. http://weather.noaa.gov/weather/current/KTOL.html 4. http://weather.noaa.gov/weather/current/KTDZ.html RWIS Data Source
1. ftp://ftp.dot.state.oh.us/pub/doit/ssi_rwis/ 2. http://www.buckeyetraffic.org/reporting/RWIS/results.aspx The RWIS stations report only four precipitation types (Rain, Snow, Fog, None/Other) while the METAR stations report more than 30 types, many of which are similar and could be grouped into the four RWIS types. A similar grouping is applied to the METAR data for the Ice Fall criteria. The only available metric for sky cover is from the METAR data. This metric had several values, four of which are used to classify the sky cover as “Clear”, with all other values grouped as “Not Clear”. Several assumptions guided the design of the dashboard. The assumptions below are based on Kathleen Jones’ report (Jones 2010a), discussions with research team members Kathleen Jones and Jeff Baker and Mike Madry, ODOT Northwood Outpost, concerning the icing events on the VGCS. The assumptions are rough guidelines. There will be exceptions to the assumptions. Assumptions: Ice accumulates in a discrete time period and does not fall during that period. The threshold of concern is radial ice accumulation of ¼” It takes a long event (roughly 12 hours or more) for ice to accumulate. If the sky is overcast and the temperature is less then 32°F, the stay sheath temperature after the icing event remains below freezing. (Note: On Tuesday February 23, 2011 this assumption was revealed to be flawed.) Ice can accumulate on the stays from fog or precipitation other than freezing rain, e.g., wet snow accumulations can lead to ice accumulation on the stays. Air temperature can be used as a reasonable approximation for stay sheath temperature. Wind by itself does not trigger an ice fall. Previously accumulated ice falls when either of these conditions occur o The stay sheath temperature rises above freezing. o If the sky is clear, sunlight could trigger the ice fall at temperatures below freezing.
75
Section 4.6: Chapter Summary
This chapter described the weather that gives rise to ice storms, the VGCS’s weather history including previous icing events, lessons learned from those previous icing events, and accretion and shedding algorithms.
The weather system most often associated with major icing is warm air from the Gulf of Mexico overriding cold air from Canada. This leads to liquid water falling on a cold surface. However, other conditions and weather systems have also lead to ice accretion. Historically, roughly two icing events occur each year. The last major icing event on the VGCS was in 2011. Once the ice accretes, it persists until shedding conditions occur. Temperatures above 32 and/or solar radiation cause ice fall. The ice fall in four of the five previous events was accompanied by temperatures rising above 32. If the solar radiation level is high enough, ice can shed. In February of 2011, copious amounts of water flowing beneath the ice were observed when the outside temperature was several degrees below freezing. This layer of water below the ice is a precursor to ice shedding. There is a greenhouse effect that occurs when the solar radiation passes through the ice and heat is trapped between the sheath and the ice. Ice accretion and shedding do not occur simultaneously. The findings from the weather study and observation include: The VGCS is not a special icing structure which accretes ice at a rate different from the surrounding structures. The hazard arises because the aerodynamic ice sheets from the stays can fall on vehicles crossing the bridge. Ice can come directly from precipitation or supercooled drizzle or fog. The development of ice accretion and shedding rules
76 Chapter 5: Development of the VGCS Dashboard and Initial Dashboard Results Section 5.1: Introduction When no existing ice prevention or removal technology appeared to be practical and/or economical for the VGCS, ODOT elected to proceed with a monitoring system to assist them in managing icing incidents. Literature review and contact with experts revealed no existing monitoring technology. Therefore, a novel monitoring system was designed, implemented and executed which is the first step in finding a long-term solution. The goal was to integrate a complex set of data and produce a concise graphic interface that put actionable information at the bridge operator’s fingers tips. Much as a dashboard puts key information about automobile operation in the visual field of a driver. A dashboard was developed that helps in monitoring ice events and other related parameters at VGCS for continuous flow of weather data. It also helps in getting actionable information in the hands of those who must anticipate and respond to an icing event. The working of the initial dashboard was based on a set of newly designed algorithms, with the following key features:
Dashboard is only using existing sensors, thus, there is no additional cost of instrumentation or installation. Many are maintained by others. An existing suite of local weather stations is utilized to form a virtual 2-D area or network of weather data in order to detect approaching weather conditions or patterns conducive to ice accretion and/or shedding. The dashboard works on a set of algorithms, which is the result of an extensive study of the causes and patterns of icing. The solution used in the dashboard is very flexible and can be modified as per user’s requirements. The solution can be used for any location/site and not just restricted to VGCS. This chapter explains the working, algorithm, user interface and performance of the dashboard in detail. It can also be seen that since designing and implementing a practical solution will take several winters, the current solution needs the regular attention of the user (i.e., Ohio Department of Transportation) at several parts of the solution for a successful implementation. The primary objective of this aspect of the larger project was to leverage existing weather data from sources available on the web in order to develop a virtual instrument. This virtual instrument allows weather researchers, infrastructure researchers, and transportation personnel all to monitor for potential icing events from any Internet connected device. Listed below is the list of tasks for this phase of the project:
Add weather data to existing VGCS web interface and database for possible use in algorithms below. Develop a “check engine” light (e.g., green (no ice), yellow (ice, but no shedding), red (ice with possibility of shedding) that responds to the algorithm and sends out corresponding alerts to a list of ODOT’s choosing.
77 Develop a reporting function that will allow ODOT to: o Verify that alerts are responded to o Declare an icing event o Capture time stamp and observation notes/comments Develop database of ground truth field data collected during actual icing events to compare against Dashboard performance. Develop export function for historical data archived on the VGCS weather website. Run calibration studies based on historical/archived/ground truth data and characterize probabilities of false alarms and missed detections (i.e., false positives and false negatives). To accomplish this, a dashboard was developed which included the virtual instrument to deliver the right information based on the task list mentioned above. The dashboard also provides a rich toolset for more detailed monitoring and assessment based on regular collection and storage of weather data from multiple sources. In addition the dashboard provides a way to interact with all data collected by location on a map and plotting the different types of measurements over time. Icing is a complex problem, and no solution has been designed to remedy it, therefore, to design an online monitoring system of continuous weather data is especially helpful, a divide and conquer approach was followed. This approach is explained in Figure 33 by a process flow diagram. Following are the major steps followed: 1. Icing Experts: First step is to get necessary information from icing experts that includes researchers in Cold Region Research and Engineering laboratory (CRREL) and ODOT. The information is then analyzed to determine the icing criteria for all the three stages in icing namely ice accumulation, ice persistence, and ice shedding. 2. Data Collection: This is a crucial phase and involves three steps: Realizing reliable weather sources Choosing appropriate weather parameters that need to be considered in the algorithm design, since different sensor system measure different weather parameters. Collecting weather data from different sources and make it available to use 3. Data Processing: This phase involves the main analysis and design; it is explained through a flowchart below. The main steps are: Ice accumulation checks for the last hour, get the results and store in the database Ice accumulation check for the last few hours to determine ice persistence In case ice is detected, manual check should be done followed by reporting to dashboard Ice shedding checks for the last one-hour, get the results and store in the database Ice shedding check for the last few hours to determine ice persistence 4. User Interface: The final step is the design of a simple user interface that lets the user to check icing status at just one-click. It also provides lot of useful data to researchers pertaining to ice events.
78
Figure 33: Process Flow Diagram Overview of this chapter: Section 2 explains different criteria for ice accumulation and ice shedding. It also explains about different types of weather stations and the parameters that are important for ice event determination. Section 3 explains the algorithm used to determine ice accumulation and ice shedding for one-hour time. Section 4 explains the algorithm for calculating likelihood of iciing using the results from Section 3. Section 5 explains the features and design of the dashboard. Section 6 deals with the performance testing of dashboard for the past and recent icing events.
79 Section 5.2: Weather Data Section 5.2.1: Introduction To accomplish the objectives mentioned in the Section above, the first step is to gather information regarding weather conditions pertinent to icing. Based on statistical analysis of meteorological data [Savadjiev], there are various meteorological variables that affect the process of ice accretion and ice shedding. Extensive analysis was done on 57 icing events, which occurred in the period between February 1998 and January 2000 at the Mont Bélair in Quebec, Canada. Savadjiev also declared that freezing rain is the most important reason behind ice accretion. K.F. Jones performed a pervasive analysis of the icing events’ that occurred in Toledo between 2000 and 2010, and found some of the common properties that were persistent during each ice events. Some of the results taken from Jones’ work are listed below:
Ice accumulation occurred in both freezing rain and snow, both accompanied by fog. Ice shedding occurs when the air temperature warms to above freezing, which may be accompanied by rain, sunshine, or gusty winds. Freezing rain was associated with ice accumulation on the Skyway stays in three of the four ice events. The results from Savadjiev and Jones led to the development of criteria to use when evaluating for ice events conditions: Criteria that would likely cause Ice Accumulation: 1. Freezing Rain: Precipitation with air temperature below 32oF. 2. Freezing Fog: Fog with air temperature below 32oF. 3. Wet Snow: Snow with air temperature above 32oF. Criteria that would likely cause ice fall are as follows: 1. Warm Air: Air Temperature above 32oF. 2. Solar Radiation: Clear sky during daylight. Section 5.2.2: Data Sources To capture the icing criterions mentioned above, meteorological data are taken from different weather sources. They are: (1) Road Weather Information System (RWIS): RWIS can be defined as a combination of technologies that uses historic and current climatologically data to develop road and weather information (for example, now casts and forecasts) to aid in roadway-related decision making. The three main elements of RWIS are: Environmental sensor system (ESS) technology to collect data. Models and other advanced processing systems to develop forecasts and tailor the information into an easily understood format.
80 Dissemination platforms on which to display the tailored information. A typical RWIS contain data for the air temperature, dew point temperature, surface temperature, relative humidity, wind speed and direction, and precipitation type. The reading is taken from buckeye traffic website and it shows one of the readings for the RWIS station at Veterans Bridge. The station name, temperature and precipitation readings are labeled. Here’s listed the links to the RWIS data including the second sources for redundancy and reliability: ftp://ftp.dot.state.oh.us/pub/doit/ssi_rwis/ http://www.buckeyetraffic.org/reporting/RWIS/results.aspx To determine the icing events at Veterans’ Glass Skyway Bridge, weather data from four RWIS stations are considered. The sensor system of the four RWIS stations are tabulated below:
Table 14: Sensor System at RWIS Stations Site # 140 141 142 150 ID 582013 582014 582016 582024 Site I-475 @ US-23 I-75 @ I-475 Split I-280 @ VGCS I-280 @ Libbey Description Split - Lucas co. - SLM 4.9 Lucas Road NLF ID SLUCIR00475**C SLUCIR00075**C SLUCIR00280**C SWOOSR00420**C Latitude 41.68768° 41.67463° 41.65845° 41.52236° Longitude -83.69355° -83.57298° -83.51022° -83.46285° Atmospheric WIVIS Hawkeye WIVIS Generic Precip Sensor Wind Sensor RM Young RM Young RM Young RM Young R/H Temp Theis Theis Theis Theis Sensor Pavement 9 GH / 2 2 FP2000/ 2 GH / 6 GH 1 FP2000 Sensor Repeaters 2 Repeaters
(2) Meteorological Terminal Aviation Routine (METAR): METAR is a format for reporting weather information. A METAR weather report is predominantly used by pilots in fulfillment of a part of a pre-flight weather briefing, and by meteorologists, who use aggregated METAR information to assist in weather forecasting. METAR typically come from airports or permanent weather observation stations. Reports are generated once an hour, but if conditions change significantly, a report known as a SPECI may be issued several times in an hour. A typical METAR contain data for the temperature, dew point, wind speed and direction, cloud cover and heights, visibility, barometric pressure, precipitation amount, lightning, and other information. The sensor data is taken from wunderground website. The links to the METAR data including the second sources for redundancy and reliability are:
81 http://weather.noaa.gov/index.html http://www.wunderground.com/history/airport To determine the icing events at Veterans’ Glass Skyway Bridge, weather data from two Airports are considered. The sensor system of the two airports are tabulated below:
Table 15: Airport Information Site # Toledo Express Airport Metcalf Field Airport ID KTOL KTDZ Latitude Degree 41.5886° 41.5631° Longitude Degree -83.8014° -83.4764° Observing LAND SURFACE COOP LAND SURFACE ASOS Program AB ASOS ASOS-NWS ASOS-FAA
Source of the above data is: http://www.faa.gov/air_traffic/weather/asos/?state=OH The observing system in both the airports are of the type Automated Surface Observing Systems (ASOS). It is a joint effort of the National Weather Service (NWS), the Federal Aviation Administration (FAA), and the Department of Defense (DOD). The ASOS system serves as the nation's primary surface weather observing network and is designed to support weather forecast activities and aviation operations and, at the same time, support the needs of the meteorological, hydrological, and climatological research communities. The basic weather elements measured by ASOS observing systems are: Sky condition: cloud height and amount (clear, scattered, broken, overcast) up to 12,000 feet Visibility (to at least 10 statute miles) Basic present weather information: type and intensity for rain, snow, and freezing rain Obstructions to vision: fog, haze Pressure: sea-level pressure, altimeter setting Ambient temperature, dew point temperature Wind: direction, speed and character (gusts, squalls) Precipitation accumulation Source of the above data is: http://www.weather.gov/ost/asostech.html As explained above the implemented ice monitoring system for the Veterans’ Bridge uses a total of six weather stations which is tabulated below:
82 Table 16: Distances of Weather Stations from VGCS Weather Station Station Typee Arial Distance from Veteraans 140-IR 475 @ US 23 Split RWIS 6.4 miles 141-IR 75 @ SLM 4.9 475 Split RWIS 11.2 miles 142-I-280 @ VGCS RWIS 0 miles 150-I-280 @ Libbey Road RWIS 10.2 miles Toledo Express Airport Aiirport 12.2 miles Metcalf Field Airport Aiirport 10.3 miles
There are a couple of reasons behind choosing the above six weather stations for pooling weather data for the ice monitoring system, some of them are listed below:
Figure 34: Map Showing Distances of Weather Stations from VGCS (1) Distance from the site: The distances of the weather stations from the Veterans’ are shown in the table above. It shows that all the choseen stations are within 12 miles radius of the experimental site. Figure 34 is a map showing the weather stations. (2) Reliability: The local weather stations are operated by various sources and data is obtained via Weather Underground, which would have varying levels of reliability. The airports are run by the National Weather Service (NWS), which is the moost dependable weather service in the country. Section 5.2.3: Data Classification The weather data collected consisted of RWIS measurements and METAR data as explained above. Since the two weather stations measure a variety of weather
83 parameters, it’s important to filter out the pertinent ones that matches the ice criterion need. As we know that Ice events can be classified in two main stages namely: Ice Accumulation: The criterion for ice accumulation was given in above sections. Criteria 1: - Freezing Rain: Precipitation with air temperature below 32oF. Criteria 2: - Freezing Fog: Fog with air temperature below 32oF. Criteria 3: - Wet Snow: Snow with air temperature above 32oF. We can see in all three criterions that temperature is a common parameter, whose absolute value can be received by any weather station but the precipitation type can have different values. Table 5.4 lists all the precipitation types values for RWIS and METAR measurements. Then we will try to classify the precipitation types that can be used for the ice accumulation determination. It can be seen that there are 30+ precipitation types measured by METAR data and couple of them can be used to determine criteria 1, 2 and 3 as shown in Table 17. On the contrary, RWIS measures only ‘Rain’ and ‘Snow’ for criterions 1 and 3 respectively. Taking these criteria and the data being collected into account, a specific set of criteria was developed for Ice Accumulation shown in Table 18.
Table 17: METAR and RWIS Precipitation Measurements for Ice Accumulation METAR Precipitation Types RWIS precipitation Types Used For CRITERIA 1 Mist Rain, Light Rain, Heavy Rain Light Freezing Rain, Light freezing Drizzle Used for CRITERIA 1 Used for CRITERIA 2 Rain Fog, Light Freezing Fog
Common to CRITERIA 1 and 2 Ice Pellets Light Ice Pellets Used for CRITERIA 3 Snow, Light Snow, Heavy Snow, Blowing Snow Ice Pellets, Light Ice Pellets Used for CRITERIA 3 Snow Light Freezing Fog, Light Freezing Rain Light Freezing Drizzle Unused types Unused types Clear, Haze, Partly Cloudy Other Scattered Clouds Unknown Overcast Mostly Cloudy
84 Table 18: Ice Accumulation Criteria Source Condition Description RWIS Freezing Atmospheric Temp. <= 32o F & Precipitation type is Rain Rain RWIS Wet Snow Atmospheric Temp. > 32o F & Precipitation type is Snow METAR Freezing (Atmospheric Temp. <= 32o F & Precipitation type is Rain) Rain OR (All precipitation types listed under criteria 1 in the above table) METAR Freezing (Atmospheric Temp. <= 32o F & Precipitation type is Fog) Fog OR (All precipitation types listed under criteria 2 in the above table) METAR Wet Snow (Atmospheric Temp. > 32o F & Precipitation type is Snow) OR (All precipitation types listed under criteria 3 in the above table)
Similar to the ice accumulation, data classification can be done for ice shedding. Ice Shedding: The criterion for ice shedding was given in above sections. Criteria 1: - Warm Air: Air Temperature above 32oF. Criteria 2: - Solar Radiation: Clear sky during daylight. We can see that one of the criteria needs temperature as a parameter, whose absolute value can be received by any weather station but the second criteria require precipitation type giving information for the sky cover, which can have different values. Let’s first go through the complete list of the precipitation types that METAR and RWIS measure and then try to classify the precipitation types that can be used for the ice shedding determination.
Table 19: METAR and RWIS Precipitation Measurements for Ice Shedding RWIS precipitation METAR Precipitation Types Types Fog, Light Freezing Fog Mist Rain, Light Rain, Heavy Rain, Light Freezing Rain Rain Thunderstorm Heavy Thunderstorms Light Thunderstorms, Thunderstorms and Rain Ice Pellets, Light Ice Pellets, Light Freezing Drizzle Snow, Light Snow, Heavy Snow, Blowing Snow Snow Unknown, Overcast Other Mostly Cloudy None Used for CRITERIA 2 Clear Haze Partly Cloudy Scattered Clouds
85 For the purposes of the dashboard, the sky is considered clear if one of the following cases are true: 1.) the sky is clear or 2.) the obstruction of the solar radiation is small. The precipitation types for the dashboard algorithms include: clear, haze, partly cloudy or scattered clouds. Based on the two ice shedding criteria previously mentioned and the available precipitation types reported, the only metric for sky cover available was from the METAR data. This metric had several values, four of which are used to classify the sky cover as clear whereas all other values are evaluated as sky cover not clear. Taking these criteria and the data being collected into account, a specific set of criteria was developed for Ice shedding shown in Table 20.
Table 20: Ice Shedding Criteria
Source Condition Description RWIS Warm Air Atmospheric Temp. >= 32o F METAR Warm Air Atmospheric Temp. >= 32o F METAR Clear (Sky Condition type is Clear) OR (Any precipitation types listed under criteria 2 in the above table)
Table 21 summarizes different checks need to be done according to the algorithm and checks actually being doing in the dashboard.
Table 21: Final Ice Accumulation/Shedding Criteria Type of Ice accumulation check Ice shedding check station RWIS ☒ Temperature less than 32°F and precipitation type: ☒ Temperature greater rain than or equal to 32°F ☒ Wet snow with temperature greater than 32°F ☐ Clear sky ☐ Fog with the temperature less than 32°F Airports ☒ Temperature less than 32°F and precipitation type: ☒ Temperature greater rain than or equal to 32°F ☒ Wet snow with temperature greater than 32°F ☒ Clear sky / Scattered ☒ Fog with the temperature less than 32°F Clouds / Partly Cloudy during day time (8am to 6pm) Legends ☒ - Conditions checked in dashboard ☐ - Conditions not checked in dashboard
Section 5.2.4: Data Collection and Storage Once the relevant weather data from RWIS and METAR measurements have been identified, we collect them in the local database. Since METAR records are updated every 1-hour, the automated program runs every 1 hour for the data collection. RWIS measurements are updated every 10-minutes, so the automated program runs every 10 minute for the data collection. The automated program is written in the language
86 Python. Data being collected is stored in MySQL database in the UCII server. The tables in the database used in data storage are listed below: (a) METAR: Store METAR data. (b) RWISatmos: Store RWIS atmospheric measurements. (c) RWISsurface: Store RWIS surface measurements. (d) RWIStraffic: Store RWIS traffic measurements.
Table (a) METAR Data: The fields in this table are as follows: “Unixtime” – Time of record (in Unix time) “Temperature” – Atmospheric temperature reading (in o F) “Events” – Precipitation type/Sky cover in detail “Conditions” – Precipitation type/Sky cover in detail “Airport” – Airport KTOL or KTDZ There are few other fields recorded in this table, which are not used in the algorithm. They are: “Dewpoint”, “Humidity”, “Pressure”, “Visibility”, “wind_dir”, “wind_speed”, “gust_speed”, and “precipitation”.
Table (b) RWIS Atmospheric Measurements: The fields in this table are as follows: “unixtime” – Time of the record (in Unix time) “Sysid” – System-id 1 for the RWIS station. For the station 582014 Sysid is 582. “Rpuid” – System-id 2 for the station. For the station 582014, Rpuid is 14. “ApAir_T” – Atmospheric temperature (in o F) “Pc_Type” – Precipitation Type (1 – Rain, 2 – Snow, etc.) There are other fields recorded in this table, which are not used in the algorithm. They are: “RcdType”, “ApAir_Dewpoint”, “ApAir_RH”, “ApW_SpdAvg”, “ApW_SpdGust”, “ApW_DirAvg”, “ApW_DirMax”, “ApPrs_Barometric”, “Pc_Intens”, “Pc_Rate” “Pc_Accum”, “Vis_Distance”. *Tables (c) and (d) are stored only for future use and not used in the current algorithm. Section 5.3: Ice Accumulation Determination Algorithm Once the data sources and the criteria are decided, we need to use them to determine the potential for Ice Accumulation or Ice Shedding occurrences. The determination of ice conditions at each of the weather stations can be done and used to further evaluate the likelihood of an icing event.
87 Section 5.3.1: Data Update Time It must be noted that each of the weather stations has its own data collection time. This has a considerable significance on the time between which the algorithm is run. Since METAR data is important in both the Ice accumulation and Ice Fall determination and its update time is 1 hour, the algorithm cannot run for less than one-hour time difference to avoid checking same records for consecutive algorithm run. That’s why the least count between two runs in this algorithm is one hour. Section 5.3.2: Ice Accumulation Algorithm Sensors in any environment can occasionally misread the actual measurement so for each of the six weather stations, we evaluate all the records for the last hour. If at least 80% of the total records from the last hour meet any (or combination) of the three ice accumulation criteria, then the station has met the icing criteria as a whole for the last hour and is given a Boolean value ‘1’. If this condition is not satisfied by a weather station, the respective station is provided a Boolean value ‘0’. This is then used to find the likelihood of ice accumulation by multiplying the condition of each weather station (0 for not met, 1 for met) by the station weight and summing each result. If the total weight calculated, as above, is greater than a set threshold, we consider that potential icing conditions have been met for the last hour. The pictorial representation of the algorithm for determination of Ice accumulation is shown in Figure 35. The mathematical equation representing the algorithm is: