Bridge Weigh-in-motion Research Articles

Virtual Axle Detector Based on Analysis of Bridge Acceleration Measurements by Fully Convolutional Network.

In the practical application of the Bridge Weigh-In-Motion (BWIM) methods, the position of the wheels or axles during the passage of a vehicle is a prerequisite in most cases. To avoid the use of conventional axle detectors and bridge type-specific methods, we propose a novel method for axle detection using accelerometers placed arbitrarily on a bridge. In order to develop a model that is as simple and comprehensible as possible, the axle detection task is implemented as a binary classification problem instead of a regression problem. The model is implemented as a Fully Convolutional Network to process signals in the form of Continuous Wavelet Transforms. This allows passages of any length to be processed in a single step with maximum efficiency while utilising multiple scales in a single evaluation. This allows our method to use acceleration signals from any location on the bridge structure and act as Virtual Axle Detectors (VADs) without being limited to specific structural types of bridges. To test the proposed method, we analysed 3787 train passages recorded on a steel trough railway bridge of a long-distance traffic line. Results of the measurement data show that our model detects 95% of the axles, which means that 128,599 out of 134,800 previously unseen axles were correctly detected. In total, 90% of the axles were detected with a maximum spatial error of 20 cm, at a maximum velocity of vmax=56.3m/s. The analysis shows that our developed model can use accelerometers as VADs even under real operating conditions.

A bridge weigh-in-motion algorithm for fast-passing railway freight vehicles considering bridge-vehicle interaction

Accurate weight information of freight vehicles is of great value for the assessment and maintenance of railway infrastructures. Existing bridge weigh-in-motion (BWIM) methods can introduce errors when the vehicle speed is high so that the dynamic effect cannot be filtered by signal processing. This paper proposes a novel BWIM algorithm for fast-passing railway freight vehicles, in which the dynamic coupling of the VB system is utilized. An extended state-space model with unknown mass parameters for the time-dependent VB system is established. Based on the model, an SPKF algorithm with a high convergence rate for identifying the mass parameters is proposed. Numerical examples of freight vehicles passing a 40 m railway bridge with a speed up to 120 km/h are presented to validate the effectiveness of the proposed algorithm.

Monitoring gross vehicle weight with a probabilistic and influence line-free bridge weight-in-motion scheme based on a transmissibility-like index

A new transmissibility-like index defined as the ratio of the frequency responses of the same monitoring location under two different loading conditions was proposed for Gross Vehicle Weights (GVWs) monitoring in this study. Based on the theoretical finding that the displacements of a beam subjected to moving loads were the convolution of the load and the influence line, the equivalence between the transmissibility-like index at the zero frequency and the ratio of two GVWs under two-moving-load scenarios was theoretically revealed. Given the reference responses for known moving loads, an influence line-free algorithm was proposed to estimate the GVW of an arbitrary vehicle by making full use of the unique property of the new transmissibility-like index. To accommodate various uncertainties and fuse the measurements of different channels simultaneously, the problem of Bridge Weigh-In-Motion (B-WIM) was formulated in the framework of Bayesian inference with the aid of a complex Gaussian ratio probabilistic model of transmissibility function. The posterior distribution of the GVW was derived analytically. By applying the proposed transmissibility-like index, this method possessed an obvious advantage in achieving robust GVWs without the requirement of any knowledge of the bridge model such as the influence line. Two applications, including a numerical example and an experimental verification, were used to demonstrate the efficiency and accuracy of the statistical and influence line-free B-WIM scheme.

Non-contact vehicle weighing method based on tire-road contact model and computer vision techniques

Vehicle overloading is a very common phenomenon. The overloaded vehicles not only cause severe damage to the road/bridge and shorten its service life, but also likely lead to traffic accidents. Therefore, the identification of vehicle load is of great importance to the management of overload vehicles and can also provide valuable information for the design and maintenance of transportation infrastructures. The commonly used vehicle weighing methods including static weighing, pavement weigh-in-motion (PWIM), and bridge weigh-in-motion (BWIM) are contact technologies with weakness such as cumbersome installation, poor durability, short service life, and high maintenance cost. It is desirable to have a non-contact vehicle weight identification method without any sensor or scales on/under the road and bridge. Thus, the present study proposed a non-contact method based on the tire-road contact model and computer vision techniques. Firstly, the tire-road contact model is theoretically derived based on the Hertz contact theory, and the relationship between the tire vertical force and the tire deformation is obtained. Then, computer vision techniques such as the image segmentation and character recognition are adopted for the identification of the tire deformation and inflation pressure that are combined with the theoretical model to determine the tire vertical force and then the vehicle weight. Finally, the field test on passenger cars and trucks are conducted to verify the proposed method. The results indicate that the proposed method has a better performance than those methods in the literature, and shows a high accuracy and robust performance in vehicle weight identification under different conditions of inflation pressure, vehicle weight, motion state, and tire type.

Bridge-weigh-in-motion approach for simultaneous multiple vehicles on concrete-box-girder bridges

Bridge-weigh-in-motion (BWIM) is a robust system that can accurately obtain traffic information to facilitate bridge design, maintenance, and management. However, the main issue with the existing BWIM approaches is their limited suitability for simultaneous multiple-vehicle cases on multiple-lane bridges. To address this limitation, in this study, a novel BWIM approach is proposed. The approach is built around the removal of the non-localized portion of the strain response. Keeping the localized portion of the strain response, which is not sensitive to nearby loads, allows for enhanced detection. The superiority of this approach stems from its capability to handle multiple-vehicle cases. These may present with an arbitrary number of trucks and light-weight vehicles, simultaneously passing the bridge in any arbitrary pattern or configuration. To show the applicability of the approach, a finite element (FE) model of a long-span concrete-box-girder bridge was simulated. The model was validated against the experimental data collected under known large events. The FE model was then used to consider single-truck events (for proof-of-concept) as well as complex multiple-truck traffic cases. These were including in-one-row trucks, zigzag patterns, side-by-side trucks, and a combination of several trucks with several light-weight vehicles present. The results demonstrated that the proposed BWIM approach is capable of obtaining the axle weights and gross vehicle weight (GVW) of the traversing trucks. Based on all complex multiple-truck cases, the overall mean absolute errors for GVW and axle weight estimations were 4.5% and 11.3%, respectively.

Data Loss Reconstruction Method for a Bridge Weigh-in-Motion System Using Generative Adversarial Networks.

In the application of a bridge weigh-in-motion (WIM) system, the collected data may be temporarily or permanently lost due to sensor failure or system transmission failure. The high data loss rate weakens the distribution characteristics of the collected data and the ability of the monitoring system to conduct assessments on bridge condition. A deep learning-based model, or generative adversarial network (GAN), is proposed to reconstruct the missing data in the bridge WIM systems. The proposed GAN in this study can model the collected dataset and predict the missing data. Firstly, the data from stable measurements before the data loss are provided, and then the generator is trained to extract the retained features from the dataset and the data lost in the process are collected by using only the responses of the remaining functional sensors. The discriminator feeds back the recognition results to the generator in order to improve its reconstruction accuracy. In the model training, two loss functions, generation loss and confrontation loss, are used, and the general outline and potential distribution characteristics of the signal are well processed by the model. Finally, by applying the engineering data of the Hangzhou Jiangdong Bridge to the GAN model, this paper verifies the effectiveness of the proposed method. The results show that the final reconstructed dataset is in good agreement with the actual dataset in terms of total vehicle weight and axle weight. Furthermore, the approximate contour and potential distribution characteristics of the original dataset are reproduced. It is suggested that the proposed method can be used in real-life applications. This research can provide a promising method for the data reconstruction of bridge monitoring systems.

Trajectory Tracking and Load Monitoring for Moving Vehicles on Bridge Based on Axle Position and Dual Camera Vision

Monitoring traffic loads is vital for ensuring bridge safety and overload controlling. Bridge weigh-in-motion (BWIM) technology, which uses an instrumented bridge as a scale platform, has been proven as an efficient and durable vehicle weight identification method. However, there are still challenges with traditional BWIM methods in solving the inverse problem under certain circumstances, such as vehicles running at a non-constant speed, or multiple vehicle presence. For conventional BWIM systems, the velocity of a moving vehicle is usually assumed to be constant. Thus, the positions of loads, which are vital in the identification process, is predicted from the acquired speed and axle spacing by utilizing dedicated axle detectors (installed on the bridge surface or under the bridge soffit). In reality, vehicles may change speed. It is therefore difficult or even impossible for axle detectors to accurately monitor the true position of a moving vehicle. If this happens, the axle loads and bridge response cannot be properly matched, and remarkable errors can be induced to the influence line calibration process and the axle weight identification results. To overcome this problem, a new BWIM method was proposed in this study. This approach estimated the bridge influence line and axle weight by associating the bridge response and axle loads with their accurate positions. Binocular vision technology was used to continuously track the spatial position of the vehicle while it traveled over the bridge. Based on the obtained time–spatial information of the vehicle axles, the ordinate of influence line, axle load, and bridge response were correctly matched in the objective function of the BWIM algorithm. The influence line of the bridge, axle, and gross weight of the vehicle could then be reliably determined. Laboratory experiments were conducted to evaluate the performance of the proposed method. The negative effect of non-constant velocity on the identification result of traditional BWIM methods and the reason were also studied. Results showed that the proposed method predicted bridge influence line and vehicle weight with a much better accuracy than conventional methods under the considered adverse situations, and the stability of BWIM technique also was effectively improved. The proposed method provides a competitive alternative for future traffic load monitoring.

A Bayesian algorithm with second order autoregressive errors for B-WIM weight estimation

Bridge weigh-in-motion (B-WIM) systems are employed for estimating axle weights of vehicles traveling over the bridge structure, providing useful information for many applications regarding structural health monitoring. In this regard, the improvement of single axle weight estimates is a major concern, since B-WIM systems in general have more difficulties with such quantities when compared to the prediction of the total weight, mainly for closely spaced axles. The main goal of the present work is to develop a weigh strategy for B-WIM systems that prevents the occurrence of spurious values, improving the overall accuracy of estimates for single axle weights. For reaching this goal, prior beliefs regarding axle weights, such as their order of magnitude and similarity for closely spaced axles, are employed. Bayesian modeling is well suited for the present problem, since it allows the suitable combination of prior beliefs and experimental data for providing proper weight estimates. In addition, a covariance matrix based on a second order autoregressive process is employed for modeling the error between theoretical and measured responses aiming to overcome the negative effects due to the presence of serial correlation in such errors. Both simulated signals and an example of B-WIM system calibration data are employed for assessing the suitability of the proposed approach. Moreover, sensitivity analyses are conducted aiming to check the robustness of the strategy to its own model parameters. For all analyses, the overall accuracy of the proposed approach, when considering both single axle as well as gross vehicle weight (GVW) estimates, outperforms the baseline algorithms. Furthermore, the sensitivity analyses indicate that the conclusions are the same for distinct prior distributions based on the same prior information.

Axle Configuration and Weight Sensing for Moving Vehicles on Bridges Based on the Clustering and Gradient Method

Traffic information, including vehicle weight and axle spacing, is vital for bridge safety. The bridge weigh-in-motion (BWIM) system remotely estimates the axle weights of moving vehicles using the response measured from instrumented bridges. It has been proved more accurate and durable than the traditional pavement-based method. However, the main drawback of conventional BWIM algorithms is that they can only identify the axle weight and the information of axle configuration (the number of axles and axle spacing) is required to be determined using an extra device in advance of the weight identification procedure. Namely, dedicated sensors (pressure-sensitive sensors placed on the deck surface or under the soffit of a bridge) in addition to weighing sensors must be adopted for identifying the axle configuration, which significantly decreases the utility, feasibility, and economic efficiency of BWIM technology. In this study, a new iterative procedure simultaneously identifying axle spacing as well as axle weights and gross weights of vehicles is proposed. The novel method is based on k-means clustering and the gradient descent method. In this method, both the axle weight and the axle location are obtained by using the same global response of bridges; thus the axle detectors are no longer required, which makes it economical and easier to be implemented. Furthermore, the proposed optimization method has good computational efficiency and thus is practical for real-time application. Comprehensive numerical simulations and laboratory experiments based on scaled vehicle and bridge models were conducted to verify the proposed method. The identification results show that the proposed method has good accuracy and high computational efficiency in axle spacing and axle weight identification.

Unburdening the Roads

How traffic management centers (TMCs) can help to protect infrastructure by using data from bridge weigh-in-motion (B-WIM) systems

An investigation of bridge influence line identification using time-domain and frequency-domain methods

The method to obtain an accurate influence line (IL) from the direct measurement is an important research topic for structural condition assessment, model correction and bridge weigh-in-motion (BWIM) system. The two most common approaches used for the identification of IL are the time-domain (TD) method and the frequency-domain (FD) method. Despite having a similar mathematical framework, the TD and the FD methods are treated as two different methods by the researchers working on this field. This paper presents a detailed theoretical demonstration to show that the two methods discussed above are nothing but the same. The two methods were compared experimentally by using field measurement data on an existing steel girder bridge which were obtained by using three calibration trucks (CTs) with different axle weights and axle configurations. Although the ILs identified by the two methods were apparently different, but a theoretical insight into the frameworks revealed that the TD and FD methods are basically the same and a seeming difference between the two methods is due to the inherent assumptions involved in the discrete Fourier transform (DFT) such as the assumption of cyclic nature of analysis interval. Finally, a method to obtain an accurate influence line has been outlined.

Effects of speeds and weights of travelling vehicles on the road surface temperature

It is important to predict the fatigue conditions of in-service roads for monitoring the state of internal pavement layers, planning the routine maintenance services and for ensuring the safety of travelling vehicles. Among various factors, the frictional heat generated by the tires of moving vehicles is considered to be an important factor causing such a distress in in-service roads. Because of this frictional heat, the temperature of wheel tracks is slightly warmer than other parts of the pavement, thereby, causing a non-uniform distribution of road surface temperature (RST) that may be responsible for developing cracks and subsequent increase in permeability of existing roads. In this study, the effect of travelling vehicles on the RST was investigated by using an infrared (IR) camera in an in-service road. In order to know the actual distribution of temperature, the RSTs were compared between the wheel track of each lane and in-between lanes. Then, the effect of speeds and weights of travelling vehicles on the RST was investigated. A bridge weigh-in-motion (BWIM) system using MEMS accelerometers was utilized to identify the speed and weight of a travelling vehicle. A strong correlation was observed between the rise of RST of wheel tracks due to the friction with the speeds and weights of travelling vehicles. The rate of increase in surface temperature was observed to be higher for the higher axle vehicles. Additionally, an influence of ambient air temperature was observed on the relationship between the rise of RST with the speeds and weights of travelling vehicles.

Acceleration-Based Deep Learning Method for Vehicle Monitoring

An automatic vehicle monitoring system can provide supports not only for intelligent transportation systems, but also for bridge weigh-in-motion (BWIM) systems, which use structural response to identify vehicle weights. In this paper, we provide a vehicle monitoring solution for acceleration-based BWIM system, utilizing deep learning and wavelet transform methods. The monitoring task is divided into three subtasks, including valid sequence detection, valid axle location, and driving lane identification. In first procedure, a shallow convolutional neural network is trained using time-frequency spectrograms to discover valuable time series. After that, an adaptive wavelet transform method is employed to locate axles from each valid sequence. Finally, the driving lane can be determined by cross-comparing vibration responses. Comparing with solutions based on all-in-one deep networks, the proposed method is computationally efficient and has improved generalization capability owing to the three-step division of the task. Evaluation is conducted on a multi-lane highway bridge located in Tokyo. Results show that 97% of vehicles can be identified correctly. For all recognized vehicles, the accuracy of driving lane detection is 100%.

Probabilistic vehicle weight estimation using physics‐constrained generative adversarial network

AbstractTraffic information plays an important role in the design and management of civil transportation infrastructure. Bridge weigh‐in‐motion (BWIM) provides an effective tool for traffic information gathering by estimating vehicle parameters including its weight through bridge responses. Most existing BWIM algorithms rarely consider the epistemic uncertainty of vehicle weight in terms of the probabilistic distribution of estimated axle weights (AWs) of the vehicle. This paper proposes a novel methodology for probabilistic vehicle weight estimation using a physics‐constrained generative adversarial network (GAN). Generative models are introduced to describe the probabilistic distributions of estimated AWs and bridge responses. Physics constraints on the generative models are formulated and enforced by minimizing a physics‐based loss function. The generative models are then learned by training a physics‐constrained GAN using the observed bridge responses. Numerical study and field testing are conducted to demonstrate the proposed method using representative highway bridges and vehicles. The results show that the proposed method can successfully capture the uncertainty in the vehicle weight estimation and provide the probabilistic distributions of the estimated AWs for different vehicle types and loading conditions considered, which can enhance the application of BWIM for relevant tasks such as traffic data collection and truck overloading enforcement. Based on the results obtained from the numerical study and field testing, the maximum coefficient of variation obtained for the AWs and gross vehicle weight of the presented cases are 0.55 and 0.11, respectively.

Bridge Weigh-in-Motion system for the identification of train loads using fiber-optic technology

This paper presents a cost-effective Bridge Weigh-in-Motion (B-WIM) system for the identification of train loads using fiber-optic technology. The system is capable of estimating the train’s speed, geometry and static axle loads, using the algorithm proposed by Moses. This algorithm involves the resolution of an inverse identification problem where the measured structural response is known and the loading scheme is unknown. The method relies on the concept of Influence Line (IL), which is estimated from the passage of a calibration vehicle with known characteristics. A numerical validation example, based on a simply supported bridge with a train passing at different speeds, demonstrated the functionality and potential accuracy of the implemented B-WIM system. For speeds up to 120 km/h, the maximum estimated errors of the wheelbase and axle loads were 10 cm and 2.5%, respectively. A B-WIM system was installed in an existing filler-beam bridge, located on the Portuguese Railways, consisting of a minimalist layout of fiber Bragg grating sensors, which guarantee higher-quality measurements, quick installation and long-term stability. The system allowed a precise characterization of several Alfa Pendular, Urban and Regional trains. This study constitutes a step forward in the development of online B-WIM systems capable of automatically estimating traffic characteristics. The accurate estimation of the traffic loads is a valuable information for the evaluation of the structural integrity and safety of railway bridges.

Moving point load approximation from bridge response signals and its application to bridge Weigh-in-Motion

This paper presents a revised theoretical interpretation for the analysis of bridge response signals arising during a vehicle passage. The derivation shows that it is possible to approximate the forcing function from bridge responses, by means of a finite difference scheme. This Point Load Approximation (PLA) is also formulated into explicit expressions following classic beam theory. This idea is then generalized to include other load effects and their corresponding time derivatives. These findings are particularly relevant for Bridge Weigh-in-Motion (BWIM) applications. The influence line for a BWIM system with multiple sensors can be explicitly defined using the presented formulation. The theory and applications are validated numerically by means of examples and case studies. These studies include the performance evaluation of a multiple strain sensor BWIM system, as well as, a novel approach to acceleration-based BWIM.

Bridge weigh-in-motion through bidirectional Recurrent Neural Network with long short-term memory and attention mechanism

In bridge weigh-in-motion (BWIM), dynamic bridge response is measured during traffic and used to identify overloaded vehicles. Most past studies of BWIM use mechanics-based algorithms to estimate axle weights. This research instead investigates deep learning, specifically the recurrent neural network (RNN), toward BWIM. In order to acquire the large data volume to train a RNN network that uses bridge response to estimate axle weights, a finite element bridge model is built through the commercial software package LS-DYNA. To mimic everyday traffic scenarios, tens of thousands of randomized vehicle formations are simulated, with different combinations of vehicle types, spacings, speeds, axle weights, axle distances, etc. Dynamic response from each of the randomized traffic scenarios is recorded for training the RNN. In this paper we propose a 3-stage Bidirectional RNN toward BWIM. Long short-term memory (LSTM) and attention mechanism are embedded in the BRNN to further improve the network performance. Additional test data indicates that the BRNN network achieves high accuracy in estimating axle weights, in comparison with a conventional moving force identification (MFI) method.

Bayesian Bridge Weigh-in-Motion and Uncertainty Estimation

Many researchers have developed bridge weigh-in-motion (BWIM) technology, mainly focusing on the representative value of the estimated axle weights. However, the estimation of the probabilistic distribution of axle weights is also important for understanding the ill conditioning of BWIM formulations and the uncertainty of estimation. Bayesian updating provides a coherent framework for assimilating data into models. Here, Bayesian bridge weigh-in-motion (BBWIM), which combines Bayesian updating and BWIM, is proposed. BBWIM can estimate not only the representative value of axle weights but also the uncertainty of the estimated value and the correlation among estimates. Uncertainties in estimated axle weight are quantitatively discussed with a simple two-axle problem. It is shown that the estimated weights of closely spaced axles have large uncertainty. BBWIM is applied to the measured data for an actual bridge. It is shown that additional information, in the form of a weak constraint on axle weight, namely, that closely spaced axles have similar weights, can reduce the uncertainty of estimated axle weights.

Study on bridge weigh in motion (BWIM) system for measuring the vehicle parameters based on strain measurement using FBG sensors

A Weigh in Motion (WIM) system is proposed, employing two fiber Bragg grating (FBG) sensors to evaluate the strain at a point on a prototype bridge due to a moving vehicle. The FBG's are fixed at two different longitudinal girders of bridge along the line of motion at different orientations. The parameters of investigation include strain developed in the proximity of FBG’s and speed of the vehicle. The response has been studied by arranging system of FBG sensors in the form of a rosette. It can be deduced from results that the suggested WIM structure could efficiently monitor variations in the load and speed of advancing vehicles over the bridge. The ideal arrangement of FBG’s in the sensor is found to be that oriented along the bridge axis due to a high sensitivity of 0.36μɛ/gram and linearity with a good coefficient of correlation of about 0.99 than other orientations of FBG. The principal strains obtained from the FBG rosette are in close correlation with the values simulated using numerical tool MATLAB.

BwimNet: A Novel Method for Identifying Moving Vehicles Utilizing a Modified Encoder-Decoder Architecture.

Traffic loading monitoring plays an important role in bridge structural health monitoring, which is helpful in overloading detection, transportation management, and safety evaluation of transportation infrastructures. Bridge weigh-in-motion (BWIM) is a method that treats traffic loading monitoring as an inverse problem, which identifies the traffic loads of the target bridge by analyzing its dynamic strain responses. To achieve accurate prediction of vehicle loads, the configuration of axles and vehicle velocity must be obtained in advance, which is conventionally acquired via additional axle-detecting sensors. However, problems arise from additional sensors such as fragile stability or expensive maintenance costs, which might plague the implementation of BWIM systems in practice. Although data-driven methods such as neural networks can estimate traffic loadings using only strain sensors, the weight data of vehicles crossing the bridge is difficult to obtain. In order to overcome these limitations, a modified encoder-decoder architecture grafted with signal-reconstruction layer is proposed in this paper to identify the properties of moving vehicles (i.e., velocity, wheelbase, and axle weight) using merely the bridge dynamic response. Encoder-decoder is an unsupervised method extracting higher features from original data. The numerical bridge model based on vehicle-bridge coupling vibration theory is established to illustrate the applicability of this new encoder-decoder method. The identification results demonstrate that the proposed approach can predict traffic loadings without using additional sensors and without requiring vehicle weight labels. Parametric studies also show that this new approach achieves better stability and reliability in identifying the properties of moving vehicles, even under the circumstances of large data pollution.