Weigh-in-Motion (WIM) systems improve the capacity of weigh station operations significantly by screening trucks while traveling at high speeds and only requiring trucks within a threshold of a maximum permissible gross of axle weight to be weighed on more accurate static scales. Consequently, the operation of a weigh station is highly dependent on the accuracy of the screening WIM system. This thesis develops a procedure for relating axle accuracy to gross vehicle accuracy and develops a field and modeling framework for evaluating weigh station operations. The WIM scale operation at the Stephens City weigh station in Virginia is examined to demonstrate how the field and modeling framework can be applied to evaluate the operation of a weigh station. Specifically, the field evaluation evaluated the accuracy of the WIM technology in addition to the operations of the weigh station in terms of service time, system time, and delay incurred at the static scales. During the field evaluation of the Stephens City WIM load cell system, the WIM technology was found to estimate truck weights to within 6 and 7 percent of the static weights 95 percent of the time. The modeling framework provides a methodology that can be used to determine the effects of the truck demand, the WIM accuracy, the system threshold, and the WIM calibration on system performance. The number of vehicles sent to the static scale and bypass lanes as well as the amount of delay experienced were analyzed for various system characteristics. The proposed framework can be utilized to estimate vehicle delay at a weigh station. / Master of Science

Deteriorating roadway conditions have drawn attention to the need to develop an accurate and practical system to control increasing excessive traffic volumes and traffic loads. In practice, traffic volumes often exceed predicted volumes, and truck overloading occurs frequently. Overloading pavements can result in premature deterioration, early or mistimed maintenance activities and eventually higher life cycle costs. As a part of an Intelligent Transportation Systems (ITS), especially in the area of Commercial Vehicle Operations (CVO), Weigh-In-Motion (WIM) has been focused on using state-of-the-art sensing technology to continuously collect vehicle weights, speeds, vehicle classes, and various types of traffic data as vehicles travel over a set of sensors (embedded or portable), without interruption of traffic flows. It is the process of measuring the dynamic tire forces of a moving vehicle and estimating the corresponding tire loads of the static vehicles. WIM technology is imperative for weight enforcement, road network design and management, as well as road safety. The overall purpose of this thesis is to examine the feasibility of using WIM in northern environments such as Canada’s. In response, one contribution of the thesis is to develop an economic model for WIM values that include costs due to premature pavement deterioration, benefits of weight enforcement and traffic data collection, benefits of WIM compared to conventional static weigh stations, and benefit-cost ratios of WIM values from road users and non-road users’ perspectives. Another contribution is to examine the technical performance (accuracy) of a particular WIM system. Results of field data collection and analysis are presented in this examination. This thesis also compares the advantages and disadvantages of different WIM systems, with respect to cost, accuracy, applicability, reliability, and sensitivity. Future trends and research potential of WIM are also discussed.

Deteriorating roadway conditions have drawn attention to the need to develop an accurate and practical system to control increasing excessive traffic volumes and traffic loads. In practice, traffic volumes often exceed predicted volumes, and truck overloading occurs frequently. Overloading pavements can result in premature deterioration, early or mistimed maintenance activities and eventually higher life cycle costs. As a part of an Intelligent Transportation Systems (ITS), especially in the area of Commercial Vehicle Operations (CVO), Weigh-In-Motion (WIM) has been focused on using state-of-the-art sensing technology to continuously collect vehicle weights, speeds, vehicle classes, and various types of traffic data as vehicles travel over a set of sensors (embedded or portable), without interruption of traffic flows. It is the process of measuring the dynamic tire forces of a moving vehicle and estimating the corresponding tire loads of the static vehicles. WIM technology is imperative for weight enforcement, road network design and management, as well as road safety. The overall purpose of this thesis is to examine the feasibility of using WIM in northern environments such as Canada’s. In response, one contribution of the thesis is to develop an economic model for WIM values that include costs due to premature pavement deterioration, benefits of weight enforcement and traffic data collection, benefits of WIM compared to conventional static weigh stations, and benefit-cost ratios of WIM values from road users and non-road users’ perspectives. Another contribution is to examine the technical performance (accuracy) of a particular WIM system. Results of field data collection and analysis are presented in this examination. This thesis also compares the advantages and disadvantages of different WIM systems, with respect to cost, accuracy, applicability, reliability, and sensitivity. Future trends and research potential of WIM are also discussed.

No description available.

Weigh-in-Motion (WIM) systems have improved the process of collecting data from heavy vehicles on the U.S. highway system and enforcing the laws that govern vehicle weights. The benefits of WIM are reaped by everyone from highway designers and voernments officials, to truck drivers and transportation industry owners. The data collected by WIM devices is essential for proper pavement design, developing pavement management systems, weight enforcement strategies, modeling traffic improvement projects, and predicting load-related distresses and performance. While WIM offers many advantages over its alternative, static weighing, the technology is limited by problems associated with the accuracy of its measurements. Weigh-in-Motion systems that lack accuracy require vehicles to travel slower and can result in higher queues, longer delays, and potential hazards. For these reasons, WIM system performance must be improved in order to adequately serve its purpose. In order to evaluate WIM system performance and determine what vehicle characteristics have the most affect on it, two systems in the Commonwealth of Virginia were evaluated. The first system was an in-service WIM system at the Troutsville weigh station on I-81. The Troutsville station had bending plate WIM scales located in both the northbound and southbound directions. The second system in a newly developed WIM system manufactured by Omni Weight Corporation (OWC) and was installed at the Virginia Smart Road for evaluation. The OWC scale is a completely sealed and buried system that has ten strain gauge sensors in its interior. Evaluation of both scales was performed by conducting a number of test runs under varying load conditions. Testing at Troutsville was performed using four different test vehicles with multiple loads on each. Variation in load was achieved by loading the test vehicles with various numbers of concrete Jersey Walls. Testing on the OWC scale was performed using only two test vehicles while varying the speed, load, tire pressure, and direction of travel over the scale. The study showed that the scales at the Troutsville weigh station yielded 10% error or less on only 77% of the tests, not complying with the required 95% set forth by ASTM E-1318. In comparison, using the manufacturer's processed data for the OWC scale yielded only 18% of its tests with 10% error or less, far below the ASTM standard. A model was developed to re-calculate the axle weights using the raw sensor data from the OWC scale; and an evaluation of the accuracy of this data showed that the OWC scale performed much better. While compliance with the ASTM standards was still not achieved, it rose from 18% to 71% of the tests having 10% error or less. Repeatability of the Troutsville scales and OWC scales was found to be comparable. / Master of Science

This dissertation proposes a system for weighing commercial vehicles in motion using acoustic emission sensors attached to a metal bar placed across the roadway. The signal from the sensors is analyzed by a computer and the vehicle weight is determined by a statistical model which correlates the acoustic emission parameters to the vehicle weight. Such a system would be portable and low-cost, allowing for the measurement of vehicle weights in much the same way commercial tube and radar counters routinely collect vehicle speed and count. The system could be used to collect vehicle speed and count data as well as weight information. Acoustic emissions are naturally occurring elastic waves produced by the rapid release of energy within a material. They are caused by deformation or fracturing of a solid due to thermal or mechanical stress. Acoustic emission sensors have been developed to detect these waves and computer software and hardware have been developed to analyze and provide information about the waveforms. Acoustic emission testing is a common form of nondestructive testing and is used for pressure vessel testing, leak detection, machinery monitoring, structural integrity monitoring, and weld monitoring, among other things (Miller, 1987). For this dissertation, acoustic emission parameters were correlated to the load placed on the metal test bar to determine the feasibility of using a metal test bar to measure the weight of a vehicle in motion. Several experiments were done. First, the concept was tested in a laboratory setting using an experimental apparatus. A concrete cylinder was mounted on a frame and rotated using a motor. The metal test bar was applied directly to the surface of the cylinder and iv acoustic emission sensors were attached to each end of the bar. As the cylinder rotated, a motorcycle tire was pushed up against the cylinder using a scissor jack to simulate different loads. The acoustic emission response in the metal test strip to the motorcycle tire rolling over it was detected by the acoustic emission sensors and analyzed by the computer. Initial examinations of the data showed a correlation between the force of the tire against the cylinder and the energy and count of the acoustic emissions. Subsequent field experiments were performed at a weigh station on I-95 in Flagler County, Florida. The proposed weigh-in-motion system (the metal test bar with attached acoustic emission sensors) was installed just downstream of the existing weigh-in-motion scale at the weigh station. Commercial vehicles were weighed on the weigh station weigh-in-motion scale and acoustic emission data was collected by the experimental system. Test data was collected over several hours on two different days, one in July 2008 and the other in April 2009. Initial examination of the data did not show direct correlation between any acoustic emission parameter and vehicle weight. As a result, a more sophisticated model was developed. Dimensional analysis was used to examine possible relationships between the acoustic emission parameters and the vehicle weight. In dimensional analysis, a dimensionally correct equation is formed using measurable parameters of a system. The dimensionally correct equation can then be tested using experimental data. Dimensional analysis revealed the following possible relationship between the acoustic emission parameters and the vehicle weight: ! = " # $% &2 , '(, )% *+( , ,% +( , ,-% +( , %3 +( . (/0% , 1% +( 2 v The definitions of these variables can be found in Appendix A. Statistical models for weight using the laboratory data and using the field data were developed. Dimensional analysis variables as well as other relevant measurable parameters were used in the development of the statistical models. The model created for the April 2009 dataset was validated, with only 27 lbs average error in the weight calculation as compared with the weight measurement made with the weigh station weigh-in-motion scale. The maximum percent error for the weight calculation was 204%, with about 65% of the data falling within 30% error. Additional research will be needed to develop an acoustic emission weigh-in-motion system with adequate accuracy for a commercial product. Nevertheless, this dissertation presents a valuable contribution to the effort of developing a low-cost acoustic emission weigh-in-motion scale. Future research needs that were identified as part of this dissertation include: ! Examination of the effects of pavement type (flexible or rigid), vehicle speeds greater than 50 mph, and temperature ! Determination of the best acoustic emission sensor for this system ! Exploration of the best method to separate the data from axles which pass over the equipment close together in time (such as tandem axles) ! Exploration of the effect of repeated measures on improving the accuracy of the system.

Structural health monitoring (SHM) is an emerging field in civil engineering in recent years. The main objectives of the SHM are to identify structural integrity issues at early stage and improve the structural safety through measuring and analyzing structural behaviors. Sensing systems for SHM can be used to identify applied vehicle loads for bridge structures. Bridge weigh-in-motion (BWIM) is one type of such vehicle load identification. As a tool to monitor the vehicle weight moving on the bridges, BWIM uses the structural responses induced by moving vehicle on the bridge to back-calculate vehicle information. In this thesis, optimized BWIM systems using multimetric measurements will be investigated. In Chapter 1, the concept and background of BWIM systems will be introduced. The objective of this research will be also demonstrated in this chapter. Chapter 2 is the literature review section. In Chapter 3, the finite element bridge model adopted for this study will be described. In this section, the moving-load time history analysis, sectional properties for bridge members, and other structural parameters of bridge model will be introduced. The methodology of BWIM systems used in this study will be demonstrated in Chapter 4. In Chapter 5, optimized sensor locations for BWIM using normal and shear strain measurements and acceleration measurement will be discussed for the case without measurement noise. In Chapter 6, sensor location optimization for the case considering measurement noises will be investigated. A new acceleration-based BWIM method is proposed in this section. Non-drift displacement reconstruction technique using acceleration measurement and FIR filtering is applied for BWIM. Finally, Chapter 7 is the conclusion part of this thesis.

Research in the area of bridge design has been and still isconcentrated on the study of the strength of materials andrelatively few studies have been performed on traffic loads andtheir effects. Traffic loads have usually been assumed to begiven in codes. This is mainly because it is very difficult tomodel traffic loads in an accurate manner because of theirrandomness. In this work, statistical evaluations of traffic loadeffects, obtained from real as well as Monte Carlo (MC)simulated vehicle data, are presented. As the dynamiccontribution of the vehicle load was filtered by the systemused for measuring vehicle weight, no attention was paid in thepresent study to the dynamic effects or the impact factor. Thedynamic contribution of the traffic load models from codes wasdeducted wherever they were compared with the result from theevaluation of the real data. First, the accuracy of thecollected data was investigated. This was done to examine theinfluence of what was most probably unreasonable data on thefinal evaluated results. Subsequently, the MC simulationtechnique, using a limited amount of the collected data, wasused to generate fictitious vehicle data that could representresults from field measurements which would otherwise have tobe recorded under a long period. Afterwards, the characteristictotal traffic loads for bridges with large spans weredetermined by probabilistic analysis. This was done using realas well as simulated data and the two were compared. Theseresults were also compared with the corresponding valuescalculated using the traffic load model from the Swedish bridgedesign code. Furthermore, using traffic data, different load effects onbridges (girder distribution factor of slab-on-girder bridgesand the mid-span deflection as well as the longitudinal stressat critical locations on box-girder bridges) were investigated.The main task was to obtain a more accurate knowledge oftraffic load distributions on bridges as well as their effectsfor infrastructure design. The results showed that the trafficload models from codes gave considerably higher load effectscompared to the current actual traffic load effects. Theseinvestigations were based on the available data for the actualposition of the vehicles on a single bridge and might not coverall possible traffic scenarios. The results showed only how thereal traffic loads, under”normal”conditions andtheir transverse positions relate to the load model accordingto the codes. <b>KEYWORDS:</b>bridge, traffic load, load effect, transversedistribution, characteristic value,weigh in motion, MonteCarlo simulation, Rice’s formula, level crossinghistogram, vehicle queue.

Transportation by trucks plays a major role in North America’s economy. The growth of this industry will increase the loads on existing roads and highways and raises the possibility of overloaded vehicles, which causes significant damage to the pavement and consequently will reduce the lifespan of the roads. Weigh-in-motion (WIM) systems technology helps to address the challenge of overloaded vehicles. This technology provides traffic monitoring, collects data for pavement research and design, and improves the capacity of static weigh station operations. However, there is still a lack of knowledge about the behaviour of WIM sensors installed in different environments, which affects reliable and precise data gathering. More knowledge is required on proper installation procedures, pavement design for WIM systems, choice of sensor type for location, and calibration processes. This research is intended to explore the behaviour of WIM piezoelectric sensors under different loads and environmental conditions. Specifically, the effects of air and pavement temperature, and weight and speed of trucks are examined with respect to the estimation accuracy of WIM sensors. To accomplish this, three WIM systems composed of different piezoelectric transducers were installed at the CPATT test site at the Waste Management facility of the Region of Waterloo in 2007, and two WIM systems were installed between exits 238 and 250 on Highway 401 eastbound near Woodstock, Ontario. It was concluded that the output of the polymer piezoelectric sensor is influenced by temperature and weight factors but not by normally observed vehicle speed differences. While temperature can be compensated for, not enough information has been gathered yet does the same for weight factor. It should be noted that very low speeds (e.g. < 50 km/hr) result in significant errors for all the sensors, so that in congested sections WIM results should be interpreted accordingly. These results will be useful for investigating the effects of environmental conditions on other WIM systems and for predicting the responses of sensors in actual installation environments. This will assist in the recommendation of: (1) alternative and transparent calibration procedures for the WIM sensor systems, (2) and improved benefits of least expensive technology.

<p>Research in the area of bridge design has been and still isconcentrated on the study of the strength of materials andrelatively few studies have been performed on traffic loads andtheir effects. Traffic loads have usually been assumed to begiven in codes. This is mainly because it is very difficult tomodel traffic loads in an accurate manner because of theirrandomness.</p><p>In this work, statistical evaluations of traffic loadeffects, obtained from real as well as Monte Carlo (MC)simulated vehicle data, are presented. As the dynamiccontribution of the vehicle load was filtered by the systemused for measuring vehicle weight, no attention was paid in thepresent study to the dynamic effects or the impact factor. Thedynamic contribution of the traffic load models from codes wasdeducted wherever they were compared with the result from theevaluation of the real data. First, the accuracy of thecollected data was investigated. This was done to examine theinfluence of what was most probably unreasonable data on thefinal evaluated results. Subsequently, the MC simulationtechnique, using a limited amount of the collected data, wasused to generate fictitious vehicle data that could representresults from field measurements which would otherwise have tobe recorded under a long period. Afterwards, the characteristictotal traffic loads for bridges with large spans weredetermined by probabilistic analysis. This was done using realas well as simulated data and the two were compared. Theseresults were also compared with the corresponding valuescalculated using the traffic load model from the Swedish bridgedesign code.</p><p>Furthermore, using traffic data, different load effects onbridges (girder distribution factor of slab-on-girder bridgesand the mid-span deflection as well as the longitudinal stressat critical locations on box-girder bridges) were investigated.The main task was to obtain a more accurate knowledge oftraffic load distributions on bridges as well as their effectsfor infrastructure design. The results showed that the trafficload models from codes gave considerably higher load effectscompared to the current actual traffic load effects. Theseinvestigations were based on the available data for the actualposition of the vehicles on a single bridge and might not coverall possible traffic scenarios. The results showed only how thereal traffic loads, under”normal”conditions andtheir transverse positions relate to the load model accordingto the codes.</p><p><b>KEYWORDS:</b>bridge, traffic load, load effect, transversedistribution, characteristic value,weigh in motion, MonteCarlo simulation, Rice’s formula, level crossinghistogram, vehicle queue.</p>