BUDAPEST, HUNGARY AUTOMATED RAVELING INSPECTION AND MAINTENANCE PLANNING ON POROUS ASPHALT IN THE NETHERLANDS Petra Paffen and Frank Bouman, RWS Peter-Paul Schackmann, TNO
CONTENTS RWS road network The raveling challenge R&D project The solution, from research to implementation Results The next challenge(s) Proposal
EXISTING RWS PAVEMENT CHARACTERISTICS Quantity about 15.000 lane kilometers 85% porous asphalt, 66% ZOAB 0/16, 17% ZOABTW / DGD 15% dense asphalt 9% DAB, 3% SMA, 3% others Road surface distress on porous asphalt is mainly causes by raveling and cracks ZOAB 0/16: 80% maintenance due to raveling 10% maintenance due to surface cracks (increasing the last years) ZOABTW / DGD 60% raveling 30% cracks (increasing since 2001)
THE RAVELING CHALLENGE Yearly monitoring program up to 2013 Wet skid resistance, ~10.000 km mostly right lanes, measurements mainly outsourced. Evenness, ~7500 km on right lanes, measurements done with own RWS vehicle. raveling and cracks, ~7500 km, Visual Condition Inspection (VCI) from hard shoulder for all lanes (estimation of intervention year). Yearly Visual Condition Inspection (VCI) undesired due to Safety regulations Subjectivity Functional procurement and guarantee contracts require objective unbiased measurements for control purposes Innovation for a safe and objective inspection technologyfor raveling and crack detection is needed
R&D PROJECT Goal of the project ( 2009-2013) Main goal is to be able to predict automatically the time to maintenance (intervention year) per 100 meter section. For this the amount of stone loss per rut (1 meter wide) per meter should be available 200 values per 100-meter section Secondary goal is to be able to determine the exactamount of stone loss Performed experiment: 1. Detailed measurements on the highway with different gradation of stone loss 2. Detailed visual inspection on the amount and location of stone loss 3. We defined criteria on what output of a measurement system would be necessary to be able to predict the time to maintenance of a 100 meter section Accuracy on the determination of stone loss: individual stone loss should be detectable The repeatability of the output of the measurement should be high 4. Market and technology scan on available measurement systems. Available or do we have to build out own system?
THE EXPERIMENT
R&D PROJECT (2) Market scan identified 3 (type of) commercially available candidates: 1. High accuracy (point) texture lasers 2.A state-of-the-art Laser scanner 3. A measurement system based on laser triangulation Simulations, using the specifications of the three systems, on the obtained real measurement data were performed. Stone loss and repeatability were determined for the three systems LMCS (laser triangulation, INO/Pavemetrics) system proved to be the best option, providing 3D height profiles with the accuracy and repeatability necessary to determine the amount of stone loss
FROM RESEARCH TO IMPLEMENTATION 2009-2010: 2010-2011: 2012-2013: 2013-2014: 2014-: Feasibility Proof of concept First generation New vehicle Update (colour) TRL 1 TRL 2 TRL 3 TRL 4 TRL 5 TRL 6 TRL 7 TRL 8 TRL 9
IMPLEMENTATION 3D data generation? Laser triangulation raveling 3D data Raveling Asset management Maintenance planning
RAVELING AND SURFACE CRACKS ARE VERY NICELY VISIBLE IN THE DATA
3 STEPS TO BE PERFORMED AFTER MEASUREMENT 1.Determine from 3D data automatically the type of asphalt applied on the road 2.Determine the amount of stone loss, based on the type of asphalt 3.Predict time to maintenance Extra: Exact amount of stone loss
STEP1: DETERMINE AUTOMATICALLY THE TYPE OF ASPHALT At least 99,6% van de ZOAB (+) and TWZOAB sections are recognised correctly, a percentage higher than can be achieved using databases
STEP2: DEVELOPED 3D STONE LOSS ALGORITHM (2) A 3D algorithm based on the 2D Stonewayprinciple has been developed The algorithm requires different parameters for different type of asphalt. We have determined the parameters for ZOAB 0/16 and TWZOAB (4/8)
STEP3: TIME TO MAINTENANCE For two regions we compared and analysed in detail the results of the time-to-maintenance obtained from our model with the visual inspections Count of WEG ModelPlanjaarZOAB MJPV na insp. 0 1 2 3 4 5 6 Grand Total 0 2 1 3 67% 1 2 4 5 2 13 85% 2 18 48 5 71 93% 3 13 86 110 6 215 97% 4 7 26 111 7 151 95% 5 7 127 63 197 100% 6 13 554 1146 1713 99% X 1 1 3 5 Grand Total 3 6 37 144 164 798 1216 2368 We see a very good resemblance of both predictions, based on our criteria that 90% of our predictions should be within 1 year of the VCI.
SCALING UP TO MEASURING 100% OF THE NETWORK 2012 2013 2014
LEVEL OF AUTOMATION: RAVELINGAND SURFACE CRACKS area ZOAB 0/16 (porous asphalt) TWZOAB (fine grade porous asphalt) DAB (Dense Asphalt Concrete) SMA (stone Mastic Asphalt) others raveling Surface cracks 2013 2014 To be decided To be decided To be decided To be decided To be decided To be decided
THE CHALLENGE Roads in the Netherlands consist already of 10+ types, and still increasing due too noise regulation. More quiet, very open surface layers, are applied. What level of automation is realistic/achievable? Remaining surface type are spread over the country. Visual inspection for these sections still require significant time Solutions available now on the market for automatic measurements are dedicated solutions, not open for further development or exchange A platform with open data from the measurements would provide an big advantage for development and exchange of results for all road operators
CURRENT SOLUTION, ALL STOVEPIPES NL: raveling on ZOAB 0/16 NL: ravelingon ZOAB 4/8 NL: Surface cracks on DAB? NL: Other Surface types? Hungary? Country X: ravelingon ZOAB 4/8? GER: Surface cracks on DAB? GER: ravelingon ZOAB 0/8 GER: Other damages / surface types? Denmark? Country Y:?
DESIRED SOLUTION Mutual challenge? NL: raveling on ZOAB 0/16 Mutual Development? Denmark:? 3D Height profiles of the road Need for algorithm Already developed in the Netherlands? Or new combination? GER+Hugr avelingon ZOAB 4/8?
NEXT STEPS RWS is looking for possible collaboration, e.g: Mutual development for new surface type/ surface deficit combination. the importance of the type of surface deficits vary from country to country. SMA and AC are applied in the Netherlands but (much) more abroad Exchange of algorithms: the algorithms for porous asphalt can be shared R&D in an European body / project A Questionnaire to road operators has been sent around to investigate the need of other road operators Please contact us in case of interest
THANK YOU FOR YOUR ATTENTION Petra Paffen: petra.paffen01@rws.nl Peter-Paul Schackmann: peter-paul.schackmann@tno.nl