Road Profile Data Quality Control The Swedish Model Atlanta, RPUG 2009
Road Profile Data Quality Control Leif Sjögren, Thomas Lundberg The Swedish Road & Transport Research Institute (VTI) Kerstin Svartling, The Swedish Road Administration (SRA)
Europe and Sweden 7 regions + main office Pavement management budget app. 300 million Euro
Sweden Sweden:» Land area: 450 000 km 2 (major variations in the density of population)» Population: 9 million people» 2000 km north to south (major variations in climate conditions) Swedish roads:» 100 000 km state roads» 40 000 km local (communal) roads» 300 000 km private roads (1/4 with state subsidies)
Road surface monitoring in Sweden Road surface monitoring of the state network has regularly been conducted since 1987 The monitoring service was outsourced 1996 and the first procurement was done. Since then a new procurement and control has been done 2000, 2004 and now 2009.
Lane kilometers measured Längd Km 90000 80000 70000 60000 50000 Km Längd Km 40000 30000 20000 10000 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 År
Main parameters» Three longitudinal profiles (0.1 meter)» Transverse profile (1 meter)» Rut depth, max, left and right (20 meter)» IRI, left and right (20 meter)» Macrotexture, MPD in three tracks (1 meter)» Megatexture, two tracks (1 meter)» Curvature, Hilliness (20 meter),» Crossfall (1 meter)» Digital images (20 meter)
A graphic view from the PMS
Road surface monitoring quality The purpose and use of road surface monitoring data: Data to support pavement management system Performance control e.g. at new constructions Active monitoring during road construction and maintenance Research purposes, e.g. studies of maintenance effects from different actions or budget levels A need to know the quality of the delivered data!
Road surface monitoring quality» We need to have an established long term (sustainable) and absolute accuracy on all indicators! 1. They are used to make decisions (allocate budget) to reach agreed road standard including safety, emissions and accessibility levels. Societal economics 2. Performance control, Commercial economics
The main monitoring difficulties» Measure and position amounts of data, maybe 100-200 km per day and vehicle» Operator skills; repeat the same road line» Deliver data in time» Internal control functions, only deliver reliable data» Data formats
Different levels of quality control» Sensor control The accuracy of a sensor e.g. laser sensor» Application control The ability to measure e.g. rut depth» Total function control The ability to deliver the relevant and reliable data in time.
Total function control Total function control includes a combination of application control and additional tests. Additional tests are» Reproducibility accuracy the accuracy of different equipment of the same type and different operator teams» Repeatability accuracy on road network This test includes the ability to measure large amounts of road surface characteristics from extensive road net work and process and delivery the result in time
Overview of quality tests Road Surface Monitoring Services» Qualified to take part in procurement process» Control during the contract period (4 years)
Procurement process, Technical tests Comparisons tests on special sections against reference device Object performance control Extensive road network measurements During the contract period control measurements are done. 5% has to be re-measured!
Quality factors Post processing Data format Internal check Data delivery
Successful concept! Results from control measurements during contract period 20 18 16 IRI % out of limits Ruth % out of limits 14 12 10 8 6 4 2 0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
Test sections, 2009» Road Network, 4 Routes 1500 km One route 5 runs, the other routes twice with cross wise changing of operators and vehicle Six test sites (800-1200 and 20-100 meter) Each site should be measured 5 times in three speeds (30, 50 and 70 km/h) Three additional sites for crack measurements Two sections for performance control
The assessment method 2.5 2 MÄTSTORHET (9999) Referens:ReferensX 1.5 Kontrollerat mätsystem: MätbilY Differens MätbilY - ReferensX (m) 1 0.5 0-0.5-1 -1.5-2 Andel mätvärden inom kontrollintervall=91.4 % Mätstorhet MätbilY 27 km/h :1 Mätstorhet MätbilY 27 km/h :2 Mätstorhet MätbilY 27 km/h :3 Mätstorhet MätbilY 27 km/h :4 Mätstorhet MätbilY 27 km/h :5 Mätstorhet MätbilY 47 km/h :1 Mätstorhet MätbilY 47 km/h :2 Mätstorhet MätbilY 48 km/h :3 Mätstorhet MätbilY 47 km/h :4 Mätstorhet MätbilY 48 km/h :5 Mätstorhet MätbilY 68 km/h :1 Mätstorhet MätbilY 67 km/h :2 Mätstorhet MätbilY 67 km/h :3 Mätstorhet MätbilY 68 km/h :4 Mätstorhet MätbilY 68 km/h :5 Tolerans 1 - Tolerans 1 + -2.5 0 1 2 3 4 5 6 7 8 9 10 Mätstorhet ReferensX (m)
Control during the contract period» 5% has to be re-measured with different operators and measurement vehicle.» All measured parameters are controlled.» The data is compared on 400 meters averages values.
Control during the contract period All parameters» Systematic errors (Control-Production)» Repeatability, correlation (Linear regression)» Within limits (+- band limits) Longitudinal profile» Longitudinal alignment until maximum correlation» Correlation» Production standard deviation/control standard deviation
Laser RST, Profilograph
Reference method Volvo TVP, cross profile and cross fall
Reference method The Primal, longitudinal profile
Reference method rod and level support for hilliness, curvature and longitudinal profile
Reference method Detailed positioning with GPS
Crack measurement reference
Crack indicator One square is 0.1*0.1 meter. If a crack is detected inside the square it turns black. The indicator is: Percentage black squares per 20 meter road in 5 zones, 2 outside ruts, 2 in ruts and 1 between ruts
Crack measurement systems in Sweden
Thank you for your attention! Thank you for your attention!