Truck load roadside vibration response simulation modeling method based on pair working condition difference

Abstract

This invention provides a simulation modeling method for roadside vibration response of truck loads based on paired condition differentials, comprising: combining acquired vehicle parameters, target load, center of gravity, vehicle speed, road surface roughness, and candidate measurement point set to generate a simulation task table; constructing a reference load task with the same vehicle speed and road surface for each task, forming a paired condition index table; calculating the wheel end basic load vector corresponding to the cargo; generating a moving loading data package by combining vehicle speed, wheelbase, contact patch, and road surface grade; substituting the data package and vehicle-road parameters into the simulation for solution, and outputting the road structure response; extracting and outputting the multi-measurement point vibration event matrix of the target and reference loads; aligning and differencing the two matrices to output the load differential event matrix; extracting the differential frequency band feature vector through frequency domain transformation; calculating the load observability score based on this, and selecting recommended measurement point frequency band combinations; and finally integrating the task table, conditions, feature vectors, and recommended combinations to construct a truck load roadside vibration response fingerprint database.

Description

Technical Field

[0001] This invention relates to vehicle load identification technology, and in particular to a simulation modeling method for roadside vibration response of truck loads based on paired operating condition differentials. Background Technology

[0002] Overloading of trucks increases road structure fatigue damage, bridge structural safety risks, and vehicle operation safety hazards. Current methods for detecting truck loads mainly include fixed weighing stations, dynamic weighing systems, vehicle-mounted weighing devices, and manual enforcement spot checks. While these methods can directly or indirectly obtain information such as total vehicle weight and axle load, practical deployment still faces challenges such as high construction costs, heavy maintenance workload, limited construction conditions, limited detection range, and difficulty in covering detour roads, low-grade roads, and source roads.

[0003] With the development of intelligent transportation sensing technology, the indirect identification of truck load status using non-contact data such as roadside vibration, acoustics, vision, and radar is gradually becoming a new direction in the management of truck overloading. Among these, roadside vibration sensing has advantages such as concealed deployment, relatively low cost, and minimal interference with vehicle traffic. However, truck load does not directly manifest as a simple roadside vibration amplitude. Vehicle load first affects the cargo center of gravity, axle load distribution, suspension compression, tire ground contact pressure, and vehicle dynamic response, and then is transmitted to the road structure through tire-road contact, ultimately manifesting as a complex vibration signal at the roadside measuring point. This transmission process is simultaneously affected by factors such as vehicle structural parameters, vehicle speed, road surface roughness, road structural parameters, measuring point location, and environmental noise.

[0004] Existing vehicle dynamics simulation methods typically focus on vehicle ride comfort, handling stability, or component load analysis; existing vehicle-road coupling simulation methods are mostly used for road structure response analysis, bridge dynamic response analysis, or vehicle-road interaction studies; existing non-contact load identification methods rely more on measured data to train models, which are easily limited by the number of samples, annotation quality, and operating condition coverage. Although the above methods can describe the dynamic response of vehicles or roads from different perspectives, they still have the following shortcomings: (1) Existing simulation methods focus on the dynamic response of vehicles or roads themselves and lack simulation organization methods for roadside non-contact load sensing tasks; (2) Existing methods usually directly compare vibration responses under different loads, but do not pair the target load condition with the benchmark load condition under the same vehicle speed, road surface, and measurement point conditions, which makes the response caused by load changes easily masked by changes in vehicle speed, road surface roughness, and propagation path; (3) Existing vehicle load simulations often simplify the load to the whole vehicle mass or axle load, which does not fully reflect the influence of longitudinal center of gravity offset and lateral off-center load on the force at each wheel end and the roadside vibration response; (4) Existing simulation outputs often remain at the level of physical quantities such as vehicle acceleration, road displacement, stress, or tire contact force, lacking a response fingerprint database construction method for further non-contact recognition algorithms, and also lacking a unified index for quantitatively evaluating the load sensitivity of measurement points and frequency bands. Therefore, it is necessary to propose a new simulation modeling method for roadside vibration response of truck loads, so that vehicle-road coupled simulation can not only obtain the vehicle and road response, but also further serve the measurement point layout, frequency band selection, feature construction and sample expansion in non-contact load sensing. Summary of the Invention

[0005] This invention provides a simulation modeling method for roadside vibration response of trucks based on pairwise differential loads, comprising: Step S1: Obtain vehicle structural parameters, target load level, cargo center of gravity parameters, vehicle speed level, road surface roughness level, and candidate roadside measurement point set, and combine them to generate a simulation task table; Step S2: For each target load task, construct a benchmark load task with the same vehicle speed, the same road surface roughness, and the same candidate roadside measuring point layout scheme as the target load task, and form a target benchmark paired working condition index table. Step S3: Calculate and output the basic load vectors at each wheel end based on the cargo mass, cargo center of gravity parameters, and vehicle structural parameters in the target baseline paired working conditions. Step S4: Generate and output a moving loading data packet based on the wheel end base load vector, vehicle speed level, axle spacing, tire contact patch size, and road surface roughness level; Step S5: Substitute the moving loading data package, vehicle dynamics parameters, and road structure parameters into the vehicle-road coupling simulation model and solve it step by step over time, while simultaneously recording and outputting the road structure response. Step S6: Based on the candidate roadside measuring point set, extract and output the multi-measuring point vibration event matrix corresponding to the target load and the reference load from the road structure response; Step S7: Read the corresponding multi-point vibration event matrix according to the paired working condition index table and perform time alignment and conditional difference to output the multi-point load difference event matrix. Step S8: Perform frequency domain transformation on the multi-measurement point load differential event matrix and divide the frequency bands, extract and output the differential frequency band feature vectors under each pair of working conditions; Step S9: Calculate the load observability score of each measurement point frequency band based on the differential frequency band feature vector, and filter and output the recommended measurement point frequency band combination set; Step S10: Integrate the simulation task table, paired working conditions, differential frequency band feature vectors, and recommended measurement point frequency band combination set to construct a truck load roadside vibration response fingerprint database.

[0006] Furthermore, the cargo center of gravity parameters mentioned in step S1 include the longitudinal center of gravity coordinates and the lateral center of gravity coordinates of the cargo; the cargo center of gravity parameters corresponding to each target load level include multiple combinations of loading states with the center of gravity centered, forward, backward, left-biased, and right-biased, which are used to simulate the influence of different loading positions on the wheel end load distribution and roadside vibration response under the same total vehicle mass.

[0007] Furthermore, the specific process of calculating and outputting the foundation load vector for each wheel end in step S3 includes: The vehicle weight is calculated based on the cargo weight and the vehicle's curb weight. Based on the distance of the cargo's longitudinal center of gravity coordinate from the reference load's longitudinal center of gravity coordinate, the vehicle's wheelbase reference length, and the proportion of unloaded axle loads for each axle, calculate the axle load distribution coefficient for each axle under the target load level. The static axle load of each axle is calculated by multiplying the axle load distribution factor by the vehicle mass. , , In the formula, Indicates the first The root axle is in the Axle load distribution factor for each load level Indicates the first The ratio of no-load or reference load axle load to the total axle load. Indicates the first Sensitivity coefficient of the root axle to longitudinal center of gravity shift of the cargo. This represents the longitudinal coordinate of the cargo's center of gravity under the target load level. This indicates the longitudinal coordinate of the cargo's center of gravity under the reference load level. Indicates the reference length of the vehicle's wheelbase. Indicates the first The root axle is in the Static axle loads at various load levels For the overall vehicle quality, Represents gravitational acceleration; Based on the lateral center of gravity coordinates of the cargo and the vehicle wheelbase, the left and right wheel end distribution coefficients of each axle are calculated, and then multiplied by the corresponding static axle loads to output the left and right wheel end basic loads of each axle, thus reconstructing the wheel end basic load vector; whereby... , , , , , In the formula, Indicates the first The left wheel end of the root axle is at the Load distribution factor for each load level Indicates the first The right wheel end of the root axle is at the first Load distribution factor for each load level Indicates the first Sensitivity coefficient of the root axle to lateral eccentric load. Indicates the lateral coordinate of the cargo's center of gravity. Indicates the vehicle's track width. Indicates the first The base load on the left wheel end of the root axle. Indicates the first The base load on the right wheel end of the axle. Indicates the first The root axle is in the Static axle loads at various load levels Indicates the first Wheel end base load vectors under each load level.

[0008] Furthermore, the tire contact patch discretization step in step S4 includes: dividing the tire contact patch at each wheel end into J contact units, and assigning a corresponding position offset to each contact unit. , and load weight , Let be the vertical offset of the j-th discrete unit relative to the center. Let be the offset of the j-th discrete element in the lateral direction relative to the center, and the sum of the load weights of all discrete contact elements satisfies 1.

[0009] Furthermore, the specific process of performing vehicle-road coupling simulation in each simulation time step in step S5 includes: The vehicle reference point and the longitudinal position of each axle are updated according to the vehicle speed level. If it is determined that the contact patch at a certain wheel end has entered the effective loading area of ​​the road structure model, the contact force calculation for the current time step is activated. The equivalent tire compression is calculated based on the vertical displacement of the unsprung mass at the corresponding wheel end, the value of the road longitudinal roughness function, and the vertical displacement of the road structure at the contact point at the wheel end; whereby... , In the formula, Indicates the first root axle Side wheel end at time Tire compression This represents the vertical displacement of the unsprung mass at the corresponding wheel end. This indicates the road surface unevenness at the longitudinal position of the wheel end. This indicates the vertical displacement of the road structure at the wheel end contact point. Indicates the first At time, the corresponding wheel end of the root axle The vertical position, Indicates the lateral position of the wheel end; The wheel-end dynamic contact force is updated by combining the equivalent tire compression, tire compression speed, tire stiffness, and tire damping; wherein the first... The contact force borne by each contact patch element is: , In the formula, Indicates the first root axle Side wheel end Each contact patch unit at time Contact force, For the first root axle Side wheel end at time Dynamic wheel-end loading; Based on the load weights of each contact unit, the dynamic wheel-end loading force is discretized and distributed to adjacent nodes of the road structure model. A time-varying mapping matrix is ​​established and substituted into the road structure dynamics equation for solution. The vehicle-side response, wheel-end contact response, and road structure response are output and recorded simultaneously. The road structure dynamics equation is as follows: , In the formula, Represents the road structure quality matrix. This represents the road structure damping matrix. Represents the road structure stiffness matrix. This represents the displacement vector of a road structure node. This represents a vector composed of all dynamic contact forces at the wheel ends. This represents the time-varying mapping matrix from wheel-end contact force to road structure nodes. and They are respectively The first and second derivatives.

[0010] Furthermore, when performing the time alignment in step S7, the time of passage of the first axis, the time of maximum peak value, or the time of peak value of the comprehensive energy of multiple measurement points is selected as the time alignment reference point in the target load event and the reference load event. The corresponding target load vibration event segment and reference load vibration event segment are translated to a unified time reference to eliminate the time delay difference in the numerical simulation or event interception process before differential calculation is performed.

[0011] Furthermore, the frequency domain transformation in step S8 is to obtain the corresponding frequency domain representation by performing Fourier transform, short-time Fourier transform or wavelet transform on the differential vibration response of the load; to integrate the square of the modulus of the frequency domain representation within each divided frequency band and output the differential frequency band energy; and to simultaneously calculate and output the differential root mean square value, differential peak value and peak frequency within the corresponding frequency band as the main frequency feature.

[0012] Further, in step S9, each virtual roadside measuring point is calculated. m In each frequency band b Load observability score The calculation formula is: In the formula, Indicates the first The first virtual roadside measuring point, the first Load observability score for each frequency band Indicates the first The first virtual roadside measuring point, the first Separability of load levels across frequency bands Indicates the first The first virtual roadside measuring point, the first Stability index of the same load in each frequency band Indicates the first The first virtual roadside measuring point, the first Vehicle speed disturbance influence coefficient for each frequency band Indicates the first The first virtual roadside measuring point, the first The influence coefficient of road surface roughness disturbance in each frequency band This indicates a small positive number that prevents the denominator from being zero.

[0013] Furthermore, when constructing the truck load roadside vibration response fingerprint database in step S10, when the corresponding load observability score is greater than or equal to the observability threshold, a recommended use tag is assigned to that record. Conversely, assign a "not recommended" label. The recommended sensor deployment scheme is an actual sensor deployment scheme generated based on the spatial coordinate interpolation of recommended measurement points with recommended usage markers.

[0014] Compared with the prior art, the significant advantages of this invention are: (1) This invention focuses on the non-contact load sensing task and organizes the vehicle-road coupling simulation into target benchmark paired working conditions, so that the simulation process can directly serve the discovery of load-sensitive features; (2) This invention constructs a benchmark load task with the same vehicle speed, the same road surface roughness, and the same measurement point layout conditions under each target load task, and weakens the influence of background factors such as vehicle speed, road surface and propagation path through the same condition difference, highlighting the vibration response increment caused by load change; (3) This invention converts cargo mass, longitudinal center of gravity and lateral off-center load into basic loads of each axle and each wheel end, so that the simulation loading object is refined from the total weight of the vehicle to the wheel end loading path, which is more in line with the actual physical process of road vibration being excited by the tire contact point; (4) This invention generates a six-axle truck wheel end moving loading sequence according to vehicle speed and wheelbase. It can express the process of different axles being loaded sequentially, the axle group spacing effect, and the wheel end pressure changing with time when a multi-axle vehicle passes through the measuring point area; (5) The present invention sets up virtual roadside measuring points in the road structure model, converting the overall road response into vibration event segments that can be observed by actual sensors, so that the simulation output can directly correspond to the roadside sensor deployment location and signal processing flow; (6) The present invention proposes a load observability score, while considering the separability between different load levels, the stability under the same load level, the influence of vehicle speed disturbance and the influence of road surface roughness disturbance, and can quantitatively screen the measuring points and frequency bands suitable for non-contact load identification; (7) The present invention can generate a simulation fingerprint library covering multiple loads, multiple speeds, and multiple road surfaces when there are insufficient real vehicle samples, providing simulation samples, feature priors, and physical constraints for non-contact overload identification algorithms.

[0015] The present invention will now be described in further detail with reference to the accompanying drawings. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method of the present invention.

[0017] Figure 2 This is a schematic diagram illustrating the paired construction of the target payload task and the reference payload task of the present invention.

[0018] Figure 3This is a schematic diagram illustrating the conversion from cargo loading state to wheel end foundation load in this invention.

[0019] Figure 4 This is a schematic diagram of the wheel-end moving loading sequence of the six-axle truck of the present invention.

[0020] Figure 5 This is a flowchart of the time step execution of the vehicle-road coupling simulation of the present invention.

[0021] Figure 6 This is a schematic diagram of the virtual roadside measuring point layout and vibration event acquisition of the present invention.

[0022] Figure 7 This is a schematic diagram illustrating the alignment difference between the target load response and the reference load response of the present invention.

[0023] Figure 8 This is a schematic diagram of the load observability score and load vibration response fingerprint database construction of the present invention. Detailed Implementation

[0024] A simulation modeling method for roadside vibration response of truck loads based on pairwise differential simulation is proposed. This method integrates truck load state, wheel-end loading path, vehicle-road coupling response, and virtual roadside measuring point signals into a unified simulation process, establishing an interpretable mapping relationship between truck load changes and roadside vibration response. The method first establishes a baseline load condition with the same vehicle speed, road surface roughness, and measuring point layout for each target load condition. Then, it extracts the vibration response increment caused by load changes individually through pairwise simulation and conditional differential simulation. Finally, it selects suitable measuring points, channels, and frequency bands for non-contact load identification based on load observability scores. The method includes the following steps: Step S1, generating a simulation task table: Obtain the vehicle structural parameters, load level, cargo center of gravity position, vehicle speed level, road surface roughness level, and candidate roadside measuring point layout scheme for the truck to be simulated, and combine the above parameters to generate a simulation task table. Further, step S1 specifically includes: Step S101, Obtain vehicle structural parameters: Obtain the curb weight of the six-axle truck. Number of vehicle axles, longitudinal distance of each axle relative to the vehicle reference point Wheelbase Vehicle wheelbase reference length Unloaded axle load ratio of each axle Suspension stiffness Suspension damping Tire stiffness Tire damping And road structure parameters. These parameters describe the basic mechanical properties of vehicles and roads, among which... Indicates the number of a six-axle truck Root axle.

[0025] Step S102, establish load levels and cargo center of gravity parameters: set multiple target load levels. Each target load level corresponds to a cargo mass Longitudinal center of gravity coordinates of the cargo and the lateral center of gravity coordinates of the cargo ,in, , This indicates the number of target load levels. The longitudinal center of gravity of the cargo describes whether the cargo is loaded towards the front or rear, while the lateral center of gravity describes whether the cargo is loaded to the left or right.

[0026] Step S103, Establish vehicle speed levels: Set multiple vehicle speed levels ,in , This indicates the number of speed levels. Speed ​​levels are used to control the time it takes for a vehicle to pass through a simulated road segment and a virtual measurement point area. Since vehicles with the same wheelbase will produce different axle passage times and frequency responses at different speeds, speed levels need to be treated as independent operating condition parameters.

[0027] Step S104, Establish road surface roughness levels: Set multiple road surface roughness levels. ,in , This indicates the number of road surface roughness grades. Road surface roughness grades are used to simulate the impact of roads with different smoothness on vehicle vibration and road response.

[0028] Step S105, Establish a candidate roadside measurement point set: Pre-set a candidate roadside measurement point set in the road structure model: , In the formula, This represents the set of candidate roadside measurement points. Indicates the first One candidate virtual roadside measurement point This indicates the number of candidate virtual roadside monitoring points. Each virtual roadside monitoring point... Corresponding to the lateral position, longitudinal position, burial depth, or roadside installation height of the road.

[0029] Step S106, Generate simulation task table: Combine the target load level, vehicle speed level, road surface roughness level, and candidate roadside measuring point set to form the simulation task table: , Among them, the Each target payload task is represented as: , In the formula, This represents the simulation task table. Indicates the first One target payload mission, Indicates the total number of target payload missions. Indicates the first Each target load level, Indicates the first The longitudinal coordinate of the cargo's center of gravity under each target load level. Indicates the first The lateral coordinate of the cargo's center of gravity under each target load level. Indicates the first Each speed level, Indicates the first Each road surface roughness grade This represents the set of candidate roadside measurement points. This indicates the target load task number. This task table allows subsequent simulations to be performed task by task, rather than setting operating parameters piecemeal.

[0030] Step S2, constructing target and baseline paired load conditions: For each target load task in the simulation task table, an automatic baseline load task is constructed. The target load task and the baseline load task maintain consistency in vehicle speed level, road surface roughness level, and candidate roadside measuring point set, differing only in load level, cargo mass, and cargo center of gravity parameters. Further, step S2 specifically includes: Step S201, Determine the reference load level: Set the reference load level The reference load level can be an unloaded load, a standard unloaded load, a legal standard load, or a reference load state determined by actual vehicle weighing. The reference load level corresponds to the cargo mass. Longitudinal center of gravity coordinates of the cargo and the lateral center of gravity coordinates of the cargo .

[0031] Step S202, Generate the baseline payload task: For the target payload task Generate the corresponding baseline payload task: , In the formula, Indicates the target payload mission The corresponding baseline payload task, Indicates the reference payload mission number. and In , and Totally consistent.

[0032] Step S203, forming paired operating conditions: The target load task and the reference load task are paired operating conditions. , In the formula, Indicates the first The system is configured to operate under specific conditions. In subsequent simulations, the system will successively test the operating conditions. and The same vehicle-road coupling simulation process was executed, and the corresponding vibration event segments were extracted at the same measurement point location and the same time reference.

[0033] Step S204, establish a paired load case index: An index table is created for each paired load case. The index table records the target load task number, the reference load task number, the vehicle speed level, the road surface roughness level, the candidate measurement point set, and the simulation result storage path. This index table is used for subsequent automatic retrieval of the target response and reference response under the same conditions, avoiding mismatches between different load cases.

[0034] Step S3, converting the cargo loading state into wheel-end basic load: Based on the cargo mass, longitudinal center of gravity position, and lateral center of gravity position in the target load task and the reference load task, as well as the vehicle wheelbase, track width, and unloaded axle load ratio, calculate the static axle load of each axle and the basic load of each wheel end. Further, step S3 specifically includes: Step S301, calculate the total vehicle load mass: for the first The target load level and the total vehicle mass are: , In the formula, Indicates the first Vehicle mass under a target load level; Indicates the vehicle's curb weight; Indicates the first The cargo mass corresponding to each target load level. This mass is used for subsequent axle load distribution calculations.

[0035] Step S302, calculate the axle load distribution factor: The root axle is in the The axle load distribution factor for each load level is: , In the formula, Indicates the first The root axle is in the Axle load distribution factor for each load level Indicates the first The ratio of no-load or reference load axle load to the total axle load. Indicates the first Sensitivity coefficient of the root axle to longitudinal center of gravity shift of the cargo. This represents the longitudinal coordinate of the cargo's center of gravity under the target load level. This indicates the longitudinal coordinate of the cargo's center of gravity under the reference load level. This indicates the reference length of the vehicle's wheelbase.

[0036] This formula describes how the load distribution ratio of each axle changes when the cargo is positioned forward or backward. The denominator is used for normalization, ensuring that the sum of the axle load distribution coefficients for the six axles is 1.

[0037] Step S303, calculate the static shaft load of each shaft: The root axle is in the The static shaft load at each load level is: , In the formula, Indicates the first The root axle is in the Static axle loads at various load levels For the overall vehicle quality, This represents the acceleration due to gravity. The static axle load represents the vertical load borne by each axle under static or quasi-static conditions of the vehicle, and is the basis for subsequent calculations of wheel end foundation loads and dynamic contact forces.

[0038] Step S304, calculate the lateral wheel end distribution coefficient: considering the lateral eccentric loading of the cargo, the first... The distribution coefficients for the left and right wheels of the root axle are as follows: , , In the formula, Indicates the first The left wheel end of the root axle is at the Load distribution factor for each load level Indicates the first The right wheel end of the root axle is at the first Load distribution factor for each load level Indicates the first Sensitivity coefficient of the root axle to lateral eccentric load. Indicates the lateral coordinate of the cargo's center of gravity. This indicates the vehicle's wheelbase. When the cargo's center of gravity is near the vehicle's centerline, the load on the left and right wheel ends is nearly evenly distributed; when the cargo shifts to one side, the load on the corresponding wheel end increases.

[0039] Step S305, calculate the foundation load at each wheel end: The foundation loads at the left and right wheel ends of the axle are as follows: , , In the formula, Indicates the first The base load on the left wheel end of the root axle. Indicates the first The basic load on the right wheel end of the root axle. Through this step, the system converts the total mass of the vehicle and the center of gravity of the cargo into basic loads at twelve wheel end loading points.

[0040] Step S306, forming the wheel end foundation load vector: The wheel end foundation load vector under each load level is: , In the formula, Indicates the first The wheel-end base load vector under each load level. This vector serves as the input for generating the wheel-end moving load sequence in step S4.

[0041] Step S4 generates the wheel-end movement loading sequence for the six-axle truck: Based on the wheel-end base load, vehicle speed, axle spacing, tire contact patch size, and road surface roughness, the loading process of each wheel end of the six-axle truck entering, passing through, and leaving the simulated road segment is calculated. Further, step S4 specifically includes: Step S401, Generate road surface roughness function: based on road surface roughness level Generate road longitudinal roughness function In one implementation, road surface roughness can be obtained by synthesizing the road surface power spectrum: , In the formula, Indicates the road surface roughness level Lower edge of the road longitudinal position Road surface unevenness, Represents the number of discrete spatial frequencies. Indicates the first Each road surface roughness level in spatial frequency Power spectral density at that point Indicates spatial frequency interval. This represents a random phase. This function is used to simulate the road surface excitation experienced by the vehicle's wheels as they pass through different road locations.

[0042] Step S402, calculate the arrival time of each axle at the reference measuring point: The time when the root axle reaches the longitudinal position of the reference measuring point is: , In the formula, Indicates the vehicle speed level Next The moment when the axle reaches the longitudinal position of the reference measuring point. Indicates the longitudinal coordinates of the reference measuring point. This represents the initial longitudinal coordinates of the vehicle's reference point when the vehicle enters the simulated road section. Indicates the first The longitudinal distance of each axle relative to a vehicle reference point. This calculation determines the time sequence in which the six axles pass through the measurement point area.

[0043] Step S403, calculate the instantaneous position of the wheel end: At time, the left and right wheel ends of the root axle The vertical position is: , In the formula, Indicates the first At time, the corresponding wheel end of the root axle The longitudinal position. Since the longitudinal positions of the left and right wheel ends of the same axle are consistent, the differences between the left and right wheel ends are mainly reflected in the lateral position and the load distribution at the wheel ends.

[0044] Step S404, establish tire contact patch discrete elements: divide the tire contact patch of each wheel end into... Each contact unit has a position offset. , and load weight , Let be the vertical offset of the j-th discrete unit relative to the center. Let be the horizontal offset of the j-th discrete unit relative to the center, and satisfy: , In the formula, This represents the number of discrete cells in a single tire contact patch. Indicates the first The contact load weight borne by each contact unit. Through contact patch discretization, the effect of the tire on the road is no longer simplified as a single concentrated force, but is represented as multiple loading points with spatial distribution characteristics.

[0045] Step S405, generate wheel-end movement loading sequence: for the first root axle, first Side wheel end, at time The dynamic wheel-end loading can be represented as: , In the formula, , Indicates the left wheel end. Indicates the right wheel end; Indicates the first root axle The side wheel end is at the Dynamic wheel-end loading at various load levels This indicates the corresponding wheel end foundation load. It represents the dynamic increment caused by suspension dynamic deformation, tire compression, and road surface unevenness.

[0046] Step S406, forming a moving load data package: combining the position of each wheel end within the simulation time range, contact patch elements, base load, and dynamic increment into a moving load data package: , In the formula, Indicates the first The target payload mission at time... The mobile loading data packet.

[0047] Step S5: Perform vehicle-road coupled simulation in time steps: Based on the moving loading sequence, vehicle dynamics parameters, and road structure parameters, the simulation progresses according to preset time steps. Within each time step, the system determines whether each wheel has entered the simulated road segment based on the vehicle's current position, updates the tire compression state and wheel-end dynamic contact force, maps the wheel-end contact force to the road structure model, and solves for the road structure vibration response. This step keeps the vehicle response, tire contact response, and road response synchronized on the time axis. Further, step S5 specifically includes: Step S501, Set the simulation time step: Set the simulation start time. End time and time step The simulated time series is obtained: , In the formula, Represents the simulated time series, with time step. The parameters are determined based on vehicle speed, wheelbase, target frequency band upper limit, and simulation accuracy requirements.

[0048] Step S502, Update vehicle location: at each time step According to vehicle speed level Update the vehicle reference point position and the positions of each axle, and determine whether each wheel end contact patch is within the effective loading area of ​​the road structure model. Wheel ends that are not within the effective loading area do not participate in the road loading at the current time step; wheel ends that are within the effective loading area participate in the contact force calculation at the current time step.

[0049] Step S503, calculate tire compression: For the wheel end entering the effective loading area, the system calculates the tire compression based on the unsprung mass displacement, road surface unevenness, and road structure displacement; root axle The equivalent tire compression at the side wheel end can be expressed as: , In the formula, Indicates the first root axle Side wheel end at time Tire compression This represents the vertical displacement of the unsprung mass at the corresponding wheel end. This indicates the road surface unevenness at the longitudinal position of the wheel end. This indicates the vertical displacement of the road structure at the wheel end contact point. Indicates the first At time, the corresponding wheel end of the root axle The vertical position, Indicates the lateral position of the wheel end.

[0050] Step S504, update wheel end dynamic contact force: update wheel end dynamic contact force based on tire compression, tire compression speed, tire stiffness and tire damping; wheel end dynamic contact force is used to represent the instantaneous vertical force that the vehicle actually acts on the road structure during operation.

[0051] Step S505, Mapping wheel-end contact force to road structure: Based on the position and load weight of the discrete elements of the tire contact patch, the dynamic contact force at the wheel end is distributed to adjacent nodes or mesh elements of the road structure model; The contact force borne by each contact patch element is: , In the formula, Indicates the first root axle Side wheel end Each contact patch unit at time Contact force, For the first root axle Side wheel end at time Dynamic wheel-end loading. Through this mapping, tire contact force is applied to the road structure in the form of distributed forces.

[0052] Step S506, Solve for the road structure response: The road structure response can be solved using a finite element model, an equivalent elastic foundation model, or a multi-layer road dynamics model. In one embodiment, the road structure dynamics equation is: , In the formula, Represents the road structure quality matrix. This represents the road structure damping matrix. Represents the road structure stiffness matrix. This represents the displacement vector of a road structure node. This represents a vector composed of all dynamic contact forces at the wheel ends. This represents the time-varying mapping matrix from wheel-end contact force to road structure nodes. and They are respectively The first and second derivatives.

[0053] Step S507, synchronously record multi-source responses: Within each time step, synchronously record the vehicle side response, wheel-end contact response, and road structure response. The vehicle side response includes the vehicle frame vertical acceleration, vehicle frame pitch angle, roll angle, suspension deformation, and unsprung mass displacement; the wheel-end contact response includes dynamic wheel-end force, contact patch pressure, and wheel-end position; the road structure response includes road node displacement, velocity, and acceleration.

[0054] Step S6, Virtual Roadside Measuring Point Vibration Event Acquisition: Multiple virtual roadside measuring points are set in the road structure model. Acceleration, velocity, or displacement time series at each virtual roadside measuring point are extracted from the road structure response. Vibration event segments are then extracted based on the times when vehicles enter, pass through, and leave the measuring point area. This step converts the overall road structure response into a signal form that can be observed by actual roadside sensors. Further, step S6 specifically includes: Step S601, Set the location of virtual roadside measuring points: Set virtual roadside measuring points near the lane centerline, wheel track line, lane edge, shoulder, roadside surface, or inside the road structure. Each virtual roadside measuring point has spatial coordinates. x m , y m , z m ),in Represents the vertical coordinate. Represents the horizontal coordinate. Represents vertical or depth coordinates.

[0055] Step S602, extract the vibration time series of the measuring point: The acceleration response of each virtual roadside measuring point is: , In the formula, Indicates the first The virtual roadside measuring point at the first Acceleration response under each task This represents the interpolation vector obtained by interpolating the responses of road structure nodes to virtual roadside measurement points. This represents the acceleration vector of a road structure node.

[0056] Step S603, determine the vibration event interception window: based on the time when the first axis reaches the reference measuring point. The moment when the last axle leaves the reference measuring point area is used to determine the center moment of the vehicle passing through the event. ;by Take the time length as the center and cut it forward. truncate the time length This creates a vibration event capture window. , In the formula, Indicates the first Vibration event capture window for each task. This indicates the duration taken before the event. This indicates the duration taken after the event.

[0057] Step S604, extract vibration event segments: in the window Within this process, the response of each virtual roadside measuring point is synchronously captured to obtain vibration event segments: , In the formula, Indicates the first The virtual roadside measuring point at the first Vibration event fragments under a specific task.

[0058] Step S605, forming a multi-point event matrix: Arrange the vibration event segments of all virtual roadside measuring points according to the measuring point number to form a multi-point event matrix: , In the formula, Indicates the first A multi-point vibration event matrix corresponding to each task. This matrix preserves the differences in vibration response at different roadside locations during vehicle passage.

[0059] Step S7, Alignment and Differentiation of Target Load Response and Reference Load Response: Time alignment is performed on the target load vibration event segments and reference load vibration event segments in the same paired operating conditions, and the differential load vibration response is calculated. This step is a key step that distinguishes this invention from ordinary vehicle-road coupled simulation. Its function is to reduce the background vibration common under the same vehicle speed, the same road surface, and the same measuring point conditions, highlighting the response differences caused by changes in load level. Further, step S7 specifically includes: Step S701, Read paired event segments: Read the target payload task according to the paired operating condition index table. Corresponding multi-point vibration event matrix and reference payload mission Corresponding multi-point vibration event matrix .

[0060] Step S702, determine the alignment reference point: In the target load event and the reference load event, select the first axis passage time, the maximum peak time, or the multi-measurement point comprehensive energy peak time as the alignment reference point. The alignment reference point is used to eliminate the micro-delay differences that may occur during numerical simulation or event interception.

[0061] Step S703, Perform time alignment: Translate the target load vibration event segment and the reference load vibration event segment to a unified time reference to obtain the time-aligned target load vibration event segment. Vibration event fragments under reference load .

[0062] Step S704, calculate the differential vibration response: differentiate the target load response and the reference load response under the same measuring point, the same vehicle speed, and the same road surface roughness. , In the formula, Indicates the first The virtual roadside measuring point at the first The differential vibration response to the load under the working condition is formed. This represents a time-aligned segment of the target load vibration event. This represents a time-aligned segment of a reference load vibration event. Because... and For the same vehicle speed, the same road surface roughness, and the same measuring point location, the difference between the two can offset the common background response to a certain extent, making the response increment caused by load change more prominent.

[0063] Step S705, Construct the differential event matrix: Combine the load differential vibration responses of all virtual roadside measuring points into a differential event matrix: , In the formula, Indicates the first This forms a differential event matrix of loads at multiple measurement points under the same working condition.

[0064] Step S8, Extracting Differential Frequency Band Features: The differential vibration response of the load is divided into frequency bands and features are extracted to obtain the differential frequency band features of each virtual roadside measuring point within each frequency band. This step is not simply a spectral analysis, but rather used to identify which frequency ranges are primarily affected by load changes, and to transform the continuous differential vibration waveform into comparable, filterable, and database-ready features. Further, step S8 specifically includes: Step S801, Establish frequency band set: Based on the sampling frequency of the roadside vibration sensor, the vehicle speed, and the natural frequency range of the road structure, the analysis frequency range is divided into multiple frequency bands: , In the formula, Represents a set of frequency bands; Indicates the first One frequency band, Indicates the number of frequency bands. Indicates the first The lower limit frequency of each frequency band Indicates the first The upper limit frequency of each frequency band.

[0065] Step S802, calculate the differential vibration response spectrum: for the load differential vibration response Perform a frequency domain transformation to obtain the frequency domain representation. The frequency domain transformation can be performed using Fourier transform, short-time Fourier transform, or wavelet transform.

[0066] Step S803, calculate the differential frequency band energy: The virtual roadside measuring point at the first The energy of the differential frequency band within each frequency band is: , In the formula, Indicates the first The virtual roadside measuring point at the first Composition of working conditions, first Differential band energy within each frequency band This represents the frequency domain representation of the differential vibration response under load. The differential frequency band energy is used to characterize the change in vibration intensity caused by load variations at the measurement point and within that frequency band.

[0067] Step S804: Calculate the differential peak value and root mean square characteristic: In addition to the differential frequency band energy, calculate the differential peak value, differential root mean square value, and dominant frequency characteristic for each measurement point and each frequency band. The differential peak value is used to describe the maximum instantaneous response caused by load changes, the differential root mean square value is used to describe the overall energy level, and the dominant frequency characteristic is used to describe the frequency position where the load change mainly affects.

[0068] Step S805, forming differential frequency band feature vectors: The virtual roadside measuring point at the first The frequency band, the first The differential frequency band feature vectors under the same operating conditions are as follows: , In the formula, This represents the differential frequency band feature vector. Represents the root mean square difference value. Indicates the differential peak value. Indicates the differential response at the th Peak frequency within a frequency band.

[0069] Step S9: Calculate the load observability score and select measurement point frequency bands: Based on the differential frequency band characteristics, evaluate the sensitivity of each virtual roadside measurement point and each frequency band to load changes. A suitable combination of measurement point frequency bands for non-contact load identification should meet the following requirements: significant differences between different load levels, stable response under the same load level, minimal susceptibility to vehicle speed disturbances, and minimal susceptibility to road surface roughness disturbances. This step outputs the load-sensitive measurement points, load-sensitive channels, and load-sensitive frequency bands. Further, step S9 specifically includes: Step S901, calculate the separability between load levels: for the first The virtual roadside measuring point and the first For each frequency band, the characteristic mean distance between different load levels is calculated to obtain the separability between load levels. In one implementation, the separability between load classes is expressed as: , In the formula, Indicates the first The first virtual roadside measuring point, the first Separability of load levels across frequency bands; Indicates load rating The mean value of the lower differential frequency band characteristics; This represents the L2 norm. The larger the value, the easier it is to distinguish different load levels at that measurement point and within that frequency band.

[0070] Step S902, calculate the stability under the same load: For the differential frequency band characteristics generated by different vehicle speeds, different road surface roughness, or different random road surface samples under the same load level, calculate its dispersion: In the formula, Indicates the first The first virtual roadside measuring point, the first Stability index of the same load in each frequency band; Indicates load rating The standard deviation of the lower differential frequency band characteristics. The smaller the value, the more stable the response under the same load level.

[0071] Step S903, calculate the vehicle speed disturbance influence coefficient: The vehicle speed disturbance influence coefficient is: In the formula, Indicates the first The first virtual roadside measuring point, the first The vehicle speed disturbance influence coefficient of each frequency band; This indicates that the variance is calculated along the vehicle speed class direction; This indicates that the variance is calculated along the load level direction; This indicates a small positive number that prevents the denominator from being zero. If The large value indicates that the characteristics of this measuring point and frequency band are easily affected by changes in vehicle speed.

[0072] Step S904, calculate the road surface roughness disturbance influence coefficient: The road surface roughness disturbance influence coefficient is: In the formula, Indicates the first The first virtual roadside measuring point, the first The influence coefficient of road surface roughness disturbance in each frequency band; This indicates the variance calculated along the direction of the road surface roughness grade. If The large value indicates that the characteristics of this measuring point and frequency band are easily affected by changes in road conditions.

[0073] Step S905, calculate the load observability score: The load observability score is: In the formula, Indicates the first The first virtual roadside measuring point, the first The load observability score for each frequency band. The higher the score, the more stably the frequency band combination of the measurement point can reflect load changes; the lower the score, the more susceptible the frequency band combination of the measurement point is to vehicle speed, road surface or random disturbances, and is not suitable as a load identification feature.

[0074] Step S906, Filter recommended measurement points and recommended frequency bands: Set observability threshold The frequency band combination of the measurement points that meets the following conditions will be selected as the recommended combination: In the formula, This represents the recommended set of frequency band combinations for measurement points; This indicates the preset observability threshold. Recommended combinations can be used to guide the actual deployment of roadside sensors and the selection of signal processing frequency bands.

[0075] Step S10: Construct a fingerprint database of truck load roadside vibration response. Based on the target load task, the baseline load task, the differential vibration response of the load, the differential frequency band characteristics, and the load observability score, a fingerprint database of truck load roadside vibration response is constructed. This fingerprint database is not a typical simulation result table, but rather a load-sensitive feature database formed after target-baseline pairing, event alignment, conditional differencing, and load observability filtering. Further, step S10 specifically includes: Step S1001, establish fingerprint recording units: Each fingerprint recording unit includes target load task number, reference load task number, load level, cargo center of gravity position, vehicle speed level, road surface roughness level, virtual roadside measuring point number, frequency band number, load differential vibration response, differential frequency band characteristics and load observability score.

[0076] Step S1002, forming a fingerprint database: The fingerprint database of truck load roadside vibration response is represented as follows: In the formula, This represents a fingerprint database representing the roadside vibration response of trucks under load. This indicates that the tag is recommended. When hour, This indicates that the frequency band combination for this measurement point is recommended for use; when hour, This indicates that the frequency band combination for this measurement point is not recommended for use.

[0077] Step S1003, output recommended sensor deployment scheme: based on The system recommends the spatial locations of measurement points and outputs a sensor deployment scheme suitable for roadside vibration sensing. The scheme includes recommended measurement point locations, recommended sampling directions, recommended installation depth or height, and recommended signal processing frequency bands.

[0078] Step S1004, Output physical constraint features: Based on the differential frequency band features and load observability score, output a set of physical constraint features that can be used in non-contact load identification algorithms. In the formula, This represents the set of physical constraint features. This set can be used as input for training machine learning models, as a basis for expanding simulation samples, or as a basis for evaluating the credibility of load identification results.

[0079] Step S1005, Output calibration interface: When a small amount of actual weighing vehicle data is obtained later, the actual vibration response can be compared with the corresponding working condition in the fingerprint database to calibrate the vehicle suspension parameters, tire stiffness damping, road damping and measuring point propagation attenuation parameters, so that the fingerprint database is closer to the actual road environment.

[0080] In step S1, the load level can be set to multiple levels such as no load, light load, standard load, near the limit load, and over the limit load. For each load level, multiple loading states can also be set, such as the center of gravity of the cargo being in the center, the center of gravity of the cargo being in front, the center of gravity of the cargo being in the back, the cargo being skewed to the left, and the cargo being skewed to the right, to simulate the influence of different loading positions on the wheel end load and roadside vibration response under the same total mass.

[0081] In step S2, the reference load level The baseline load level can be determined based on the actual application scenario. When this method is used to identify overloaded vehicles, the baseline load level can be set to the legal load limit state; when this method is used to analyze load variation patterns, the baseline load level can be set to the no-load or light-load state. Once the baseline load level is determined, it should be kept consistent in the same batch of simulation tasks to ensure the comparability of the differential results.

[0082] In step S3, if the actual static weighing data of the vehicle can be obtained, the actual axle load data can be used for correction. , and If measured axle load data is temporarily unavailable, initial values ​​can be set based on the vehicle wheelbase, cargo box position, axle assembly structure, and manufacturer parameters, and then calibrated using subsequent measured data.

[0083] In step S4, the tire contact patch can be divided using a one-dimensional strip along the driving direction, or a two-dimensional rectangular or elliptical mesh. When the focus is on the propagation of roadside vibration, multi-point discretization along the driving direction can be preferred; when the focus is on the distribution of tire ground pressure, two-dimensional contact patch mesh discretization can be used.

[0084] In step S5, the vehicle-road coupling simulation can be implemented using a finite element road model, an equivalent elastic foundation model, or a multi-layer road structure dynamics model. When improved computational efficiency is required, an equivalent elastic foundation model can be used; when it is necessary to analyze the response of the internal structural layers of the road, a finite element road model can be used.

[0085] In step S6, virtual roadside measuring points can be set near wheel tracks, lane edges, shoulders, roadside pole foundations, or inside the road structure. By comparing the load observability scores of different virtual roadside measuring points, it can be determined whether the actual sensors are more suitable for placement near wheel tracks, roadside edges, or inside the road.

[0086] In step S7, time alignment can be achieved by aligning the first axle passage time, peak time alignment, or cross-correlation maximum value alignment. When the simulation event has a clear axle group passage peak value, the first axle passage time alignment is preferred; when the peak value of the measurement point signal is obvious but the axle peak value is difficult to distinguish, the cross-correlation maximum value alignment can be used.

[0087] In step S8, the frequency band division can be determined based on the sensor sampling frequency, vehicle speed, and the frequency range of road structure response. For road structures with significant low-frequency response, a denser low-frequency band can be set; for scenarios with significant tire contact impact, mid-to-high frequency band analysis can be added.

[0088] In step S9, the load observability score can have its weighting coefficient increased according to actual needs. For example, when the vehicle speed changes significantly in the application scenario, the penalty weight of the vehicle speed disturbance influence coefficient can be increased; when the road surface conditions change significantly in the application scenario, the penalty weight of the road surface roughness disturbance influence coefficient can be increased.

[0089] In step S10, the truck load roadside vibration response fingerprint database can serve as a training sample library for the non-contact load identification model, and also as a sensor deployment simulation and evaluation tool before the actual roadside sensing system goes online. When the actual system collects new weighing vehicle data, the new data can be written back to the fingerprint database to iteratively update the simulation model and the recommended measurement point frequency band combination.

Claims

1. A simulation modeling method for roadside vibration response of trucks under load based on pairwise differential load conditions, characterized in that, include: Step S1: Obtain vehicle structural parameters, target load level, cargo center of gravity parameters, vehicle speed level, road surface roughness level, and candidate roadside measurement point set, and combine them to generate a simulation task table; Step S2: For each target load task, construct a benchmark load task with the same vehicle speed, the same road surface roughness, and the same candidate roadside measuring point layout scheme as the target load task, and form a target benchmark paired working condition index table. Step S3: Calculate and output the basic load vectors at each wheel end based on the cargo mass, cargo center of gravity parameters, and vehicle structural parameters in the target baseline paired working conditions. Step S4: Generate and output a moving loading data packet based on the wheel end base load vector, vehicle speed level, axle spacing, tire contact patch size, and road surface roughness level; Step S5: Substitute the moving loading data package, vehicle dynamics parameters, and road structure parameters into the vehicle-road coupling simulation model and solve it step by step over time, while simultaneously recording and outputting the road structure response. Step S6: Based on the candidate roadside measuring point set, extract and output the multi-measuring point vibration event matrix corresponding to the target load and the reference load from the road structure response; Step S7: Read the corresponding multi-point vibration event matrix according to the paired working condition index table and perform time alignment and conditional difference to output the multi-point load difference event matrix. Step S8: Perform frequency domain transformation on the multi-measurement point load differential event matrix and divide the frequency bands, extract and output the differential frequency band feature vectors under each pair of working conditions; Step S9: Calculate the load observability score of each measurement point frequency band based on the differential frequency band feature vector, and filter and output the recommended measurement point frequency band combination set; Step S10: Integrate the simulation task table, paired working conditions, differential frequency band feature vectors, and recommended measurement point frequency band combination set to construct a truck load roadside vibration response fingerprint database.

2. The method according to claim 1, characterized in that, The cargo center of gravity parameters mentioned in step S1 include the longitudinal center of gravity coordinates and the lateral center of gravity coordinates of the cargo; the cargo center of gravity parameters corresponding to each target load level include multiple combinations of loading states with the center of gravity centered, forward, backward, left-biased, and right-biased, which are used to simulate the influence of different loading positions on the wheel end load distribution and roadside vibration response under the same total vehicle mass.

3. The method according to claim 2, characterized in that, The specific process of calculating and outputting the foundation load vectors at each wheel end in step S3 includes: The vehicle weight is calculated based on the cargo weight and the vehicle's curb weight. Based on the distance of the cargo's longitudinal center of gravity coordinate from the reference load's longitudinal center of gravity coordinate, the vehicle's wheelbase reference length, and the proportion of unloaded axle loads for each axle, calculate the axle load distribution coefficient for each axle under the target load level. The static axle load of each axle is calculated by multiplying the axle load distribution factor by the vehicle mass. ， ， In the formula, Indicates the first The root axle is in the Axle load distribution factor for each load level Indicates the first The ratio of no-load or reference load axle load to the total axle load. Indicates the first Sensitivity coefficient of the root axle to longitudinal center of gravity shift of the cargo. This represents the longitudinal coordinate of the cargo's center of gravity under the target load level. This indicates the longitudinal coordinate of the cargo's center of gravity under the reference load level. Indicates the reference length of the vehicle's wheelbase. Indicates the first The root axle is in the Static axle loads at various load levels For the overall vehicle quality, Represents gravitational acceleration; Based on the lateral center of gravity coordinates of the cargo and the vehicle wheelbase, the left and right wheel end distribution coefficients of each axle are calculated, and then multiplied by the corresponding static axle loads to output the left and right wheel end basic loads of each axle, thus reconstructing the wheel end basic load vector; whereby... ， ， ， ， ， In the formula, Indicates the first The left wheel end of the root axle is at the Load distribution factor for each load level Indicates the first The right wheel end of the root axle is at the first Load distribution factor for each load level Indicates the first Sensitivity coefficient of the root axle to lateral eccentric load. Indicates the lateral coordinate of the cargo's center of gravity. Indicates the vehicle's track width. Indicates the first The base load on the left wheel end of the root axle. Indicates the first The base load on the right wheel end of the axle. Indicates the first The root axle is in the Static axle loads at various load levels Indicates the first Wheel end base load vectors under each load level.

4. The method according to claim 3, characterized in that, The tire contact patch discretization step in step S4 includes: dividing the tire contact patch at each wheel end into J contact units, and assigning a corresponding position offset to each contact unit. , and load weight , Let be the vertical offset of the j-th discrete unit relative to the center. Let be the offset of the j-th discrete element in the lateral direction relative to the center, and the sum of the load weights of all discrete contact elements satisfies 1.

5. The method according to claim 4, characterized in that, The specific process of performing vehicle-road coupling simulation in each simulation time step in step S5 includes: The vehicle reference point and the longitudinal position of each axle are updated according to the vehicle speed level. If it is determined that the contact patch at a certain wheel end has entered the effective loading area of ​​the road structure model, the contact force calculation for the current time step is activated. The equivalent tire compression is calculated based on the vertical displacement of the unsprung mass at the corresponding wheel end, the value of the road longitudinal roughness function, and the vertical displacement of the road structure at the contact point at the wheel end; whereby... ， In the formula, Indicates the first root axle Side wheel end at time Tire compression This represents the vertical displacement of the unsprung mass at the corresponding wheel end. This indicates the road surface unevenness at the longitudinal position of the wheel end. This indicates the vertical displacement of the road structure at the wheel end contact point. Indicates the first At time, the corresponding wheel end of the root axle The vertical position, Indicates the lateral position of the wheel end; The wheel-end dynamic contact force is updated by combining the equivalent tire compression, tire compression speed, tire stiffness, and tire damping; wherein the first... The contact force borne by each contact patch element is: ， In the formula, Indicates the first root axle Side wheel end Each contact patch unit at time Contact force, For the first root axle Side wheel end at time Dynamic wheel-end loading; Based on the load weights of each contact unit, the dynamic wheel-end loading force is discretized and distributed to adjacent nodes of the road structure model. A time-varying mapping matrix is ​​established and substituted into the road structure dynamics equation for solution. The vehicle-side response, wheel-end contact response, and road structure response are output and recorded simultaneously. The road structure dynamics equation is as follows: ， In the formula, Represents the road structure quality matrix. This represents the road structure damping matrix. Represents the road structure stiffness matrix. Represents the displacement vector of road structure nodes. This represents a vector composed of all dynamic contact forces at the wheel ends. This represents the time-varying mapping matrix from wheel-end contact force to road structure nodes. and They are respectively The first and second derivatives.

6. The method according to claim 5, characterized in that, When performing the time alignment in step S7, the time of passage of the first axis, the time of maximum peak value, or the time of peak value of the comprehensive energy of multiple measurement points is selected as the time alignment reference point in the target load event and the reference load event. The corresponding target load vibration event segment and the reference load vibration event segment are translated to a unified time reference to eliminate the time delay difference in the numerical simulation or event interception process before differential calculation is performed.

7. The method according to claim 6, characterized in that, The frequency domain transformation in step S8 is to obtain the corresponding frequency domain representation by performing Fourier transform, short-time Fourier transform or wavelet transform on the differential vibration response of the load; to integrate the square of the modulus of the frequency domain representation within each divided frequency band and output the differential frequency band energy; and to simultaneously calculate and output the differential root mean square value, differential peak value and peak frequency in the corresponding frequency band as the main frequency feature.

8. The method according to claim 7, characterized in that, In step S9, each virtual roadside measuring point is calculated. m In each frequency band b Load observability score The calculation formula is: In the formula, Indicates the first The first virtual roadside measuring point, the first Load observability score for each frequency band Indicates the first The first virtual roadside measuring point, the first Separability of load levels across frequency bands Indicates the first The first virtual roadside measuring point, the first Stability index of the same load in each frequency band Indicates the first The first virtual roadside measuring point, the first Vehicle speed disturbance influence coefficient for each frequency band Indicates the first The first virtual roadside measuring point, the first The influence coefficient of road surface roughness disturbance in each frequency band It represents a tiny positive number that prevents the denominator from being zero.

9. The method according to claim 8, characterized in that, In step S10, when constructing the truck load roadside vibration response fingerprint database, if the corresponding load observability score is greater than or equal to the observability threshold, a recommended use tag is assigned to that record. Conversely, assign a "not recommended" label. The recommended sensor deployment scheme is an actual sensor deployment scheme generated based on the spatial coordinate interpolation of recommended measurement points with recommended usage markers.

Images