Test bench method and system for adaptive calibration of model parameters of active suspension system

Abstract

The application discloses a kind of active suspension system model parameter adaptive calibration test bench test method and system, belong to vehicle active suspension test technical field, sensitive system parameters are filtered out by sensitivity analysis function, then sensitive system parameters are identified and calibrated, greatly reduce the amount of calculation in calibration process, improve the test implementation efficiency;Through time delay boundary calculation, the time delay feedback coefficient and time delay are optimized, the optimal time delay feedback coefficient and optimal time delay are obtained, and through adaptive adjustment, time delay compensation calibration is carried out using Smith prediction compensation model, which significantly improves the adaptive ability and compensation accuracy of time-varying time delay;In addition, standard test signals are applied to the target active suspension test bench test system, and time domain response data are collected to determine the tolerance according to the time domain response data to complete the parameter adaptive calibration of the active suspension system dynamics model.

Description

Technical Field

[0001] This invention belongs to the field of vehicle active suspension testing technology, specifically relating to a bench test method and system for adaptive calibration of active suspension system model parameters. Background Technology

[0002] Active suspension systems can actively adjust output force according to road conditions and vehicle posture, thereby effectively damping vehicle vibration and improving ride comfort and handling stability. Electro-hydraulic (EHA) active suspension and electro-hydraulic active suspension are widely used in emergency rescue vehicles and heavy vehicles due to their advantages such as high power density, strong load-bearing capacity, and wide adjustment range.

[0003] However, active suspension systems face numerous technical challenges in practical applications. Firstly, during operation, the measurement and transmission of sensor signals, the processing and computation of controller strategies, and the generation of active force by hydraulic components all involve time delays. These time delays are time-varying, severely impacting the dynamic characteristics of the suspension and potentially leading to system instability. Furthermore, electro-hydraulic active suspension systems suffer from common hydraulic system problems such as strong nonlinearity and time-varying parameters. In addition, the load spectrum acting on the suspension unit is complex and variable, with changes in key parameters such as flow gain, effective bulk modulus, actuator effective area, and total flow pressure coefficient directly affecting the system's dynamic characteristics. Moreover, existing sensitivity analyses primarily focus on position control systems, with insufficient research on sensitivity analysis of active suspension output force control, making it difficult to effectively guide parameter calibration.

[0004] As mentioned above, how to provide a bench test method that can solve the problem of active suspension time lag, effectively cope with the time-varying characteristics of parameters, and accurately adaptively calibrate the parameters of the active suspension system model has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] The purpose of this invention is to provide a bench test method and system for adaptive calibration of active suspension system model parameters, in order to solve the above-mentioned problems existing in the prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a bench test method for adaptive calibration of active suspension system model parameters, comprising: Based on kinematic theorems, a dynamic model of an active suspension system including time delay is established, and a system parameter sensitivity analysis function is constructed for the dynamic model of the active suspension system. The sensitivity index of each system parameter is solved from the system parameter sensitivity analysis function, and the sensitive system parameters of the active suspension system dynamic model are selected. Simulated road surface excitation is applied to the target active suspension bench test system to collect the time-domain response data of the target active suspension bench test system, and the time-delay stability working boundary of the target active suspension bench test system is calculated through the time-domain response data; Based on the sensitive system parameters of the active suspension system dynamic model, the working conditions and input signals of the target active suspension bench test system are adjusted. Multiple sets of response data are collected from the target active suspension bench test system. Based on each set of response data, the real-time estimated values ​​of the sensitive system parameters are calculated. The active suspension system dynamic model is updated using the real-time estimated values ​​of the sensitive system parameters to obtain a preliminary calibrated active suspension system dynamic model. Using the time-delay stable working boundary of the target active suspension bench test system as the boundary condition, the objective function is to maximize the vibration reduction performance index of the active suspension. The time-delay feedback coefficient and time delay are the optimization variables. In the preliminary calibration of the dynamic model of the active suspension system, the optimization algorithm is used to optimize the objective function to obtain the optimal time-delay feedback coefficient and the optimal time delay. A Smith prediction compensation model is constructed based on the optimal time delay feedback coefficient and the optimal time delay amount. The Smith prediction compensation model is used to perform compensation prediction on the preliminary calibration active suspension system dynamic model to obtain the predicted compensation output. The optimal time delay feedback coefficient and the optimal time delay amount are used as compensation references. The predicted compensation output is compared with the actual system output of the target active suspension bench test system to obtain the comparison error. The time delay compensation amount is generated based on the comparison error. The time delay compensation amount is used to perform time delay compensation calibration on the preliminary calibration active suspension system dynamic model to obtain the time delay calibration active suspension system dynamic model. The actuator control force output from the time-delay calibration active suspension system dynamic model is input into the target active suspension bench test system. A standard test signal is applied to the target active suspension bench test system and time-domain response data is collected. Tolerance determination is performed based on the time-domain response data to complete the parameter adaptive calibration of the active suspension system dynamic model.

[0007] In one possible design, based on kinematic theorems, a dynamic model of the active suspension system including time delay is established. A system parameter sensitivity analysis function is constructed for the dynamic model of the active suspension system. The sensitivity indices of each system parameter are solved from the system parameter sensitivity analysis function, and the sensitive system parameters of the active suspension system dynamic model are selected, including: The suspension mechanical topology of the target active suspension bench test system is obtained, and the road surface roughness excitation is used as the model input, and the sprung mass displacement and unsprung mass displacement are used as the model output. Based on the kinematic theorem, a dynamic model of the active suspension system including time delay is constructed. Based on the dynamic model of the active suspension system, a driving force output sub-model between the control voltage signal and the actuator driving force is established for the active actuator in the target active suspension bench test system. Using the driving force output sub-model, the dynamic model of the active suspension system is organized into a first-order state-space equation, and the undetermined physical parameters are extracted from the first-order state-space equation as the system parameters of the dynamic model of the active suspension system. For the first-order state-space equation, the parameter vector partial derivatives of each system parameter are calculated to form a system parameter sensitivity analysis function, and the sensitivity of the system parameter sensitivity analysis function is solved to obtain the sensitivity index of each system parameter. A preset sensitivity threshold is obtained. Based on the sensitivity threshold, the sensitivity indices of each system parameter are compared, and each system parameter whose sensitivity index exceeds the sensitivity threshold is selected as the sensitive system parameter of the active suspension system dynamic model.

[0008] In one possible design, simulated road surface excitation is applied to the target active suspension bench test system to collect its time-domain response data. The time-delay stable operating boundary of the target active suspension bench test system is then calculated using this time-domain response data, including: A sinusoidal sweep frequency excitation signal is applied to the target active suspension bench test system as a simulated road surface excitation, and the time-domain response data output by the target active suspension bench test system is collected in real time. The time-domain response data is preprocessed, and the corresponding amplitude-frequency response features and phase-frequency response features are extracted from the preprocessed time-domain response data. The amplitude frequency response characteristics and the phase frequency response characteristics are input into the dynamic model of the active suspension system to obtain the theoretical critical time delay value under the current controllable damping conditions; Adjust the equivalent time delay in the target active suspension bench test system, and repeatedly perform simulated road excitation input to obtain the time delay corresponding to the first occurrence of continuous constant amplitude oscillation in the target active suspension bench test system. Use the time delay corresponding to the first occurrence of continuous constant amplitude oscillation in the target active suspension bench test system as the measured critical time delay value. A preset time delay deviation threshold is obtained, the time delay deviation between the theoretical critical time delay value and the measured critical time delay value is calculated, and the time delay deviation threshold and the time delay deviation are compared to determine the reliability of the active suspension system dynamic model. If the time delay deviation does not exceed the time delay deviation threshold, the dynamic model of the active suspension system is determined to be reliable, and the theoretical critical time delay value is used as the time delay stability working boundary of the target active suspension bench test system. If the time delay deviation exceeds the time delay deviation threshold, the active suspension system dynamic model is determined to be unreliable. The parameters of the active suspension system dynamic model are adjusted using the measured critical time delay value, and the theoretical critical time delay value is calculated again for the active suspension system dynamic model after parameter adjustment. The time delay deviation is calculated and the model reliability is determined for the recalculated theoretical critical time delay value until the active suspension system dynamic model is determined to be reliable.

[0009] In one possible design, based on the sensitive system parameters of the active suspension system dynamic model, the operating conditions and input signals of the target active suspension bench test system are adjusted. Multiple sets of response data are collected from the target active suspension bench test system. Real-time estimates of the sensitive system parameters are calculated based on each set of response data. The active suspension system dynamic model is then updated using these real-time estimates to obtain a preliminary calibrated active suspension system dynamic model, including: For each sensitive system parameter of the active suspension system dynamic model, corresponding identification conditions are generated. Based on the identification conditions corresponding to each sensitive system parameter, the working conditions and input signals of the target active suspension bench test system are adjusted to complete multiple tests of the target active suspension bench test system. The control voltage signal and system output signal of each test are collected as response data. The control voltage signal in each group of response data is used as an input quantity and input to the active suspension system dynamic model to obtain the driving force output by the active suspension system dynamic model as the model output force; From the system output signal in each group of response data, the actuator driving force corresponding to each group of response data is extracted as the actual output force. The square integral of the deviation between the model output force and the actual output force corresponding to each group of response data is calculated. The square integral of the deviation between the model output force and the actual output force is used as the identification target function. A preset convergence threshold is obtained, and the identification objective function corresponding to each sensitive system parameter is iteratively solved until the identification objective function value corresponding to each sensitive system parameter is lower than the convergence threshold, thereby obtaining the sensitive system parameter identification value corresponding to each sensitive system parameter. The sensitive system parameter identification values ​​corresponding to each sensitive system parameter are used as preliminary calibration parameter values ​​and input into the active suspension system dynamic model. The parameters of each sensitive system parameter in the active suspension system dynamic model are replaced to complete the update of the active suspension system dynamic model and obtain the preliminary calibration active suspension system dynamic model.

[0010] In one possible design, the time-delay stable working boundary of the target active suspension bench test system is used as the boundary condition. The objective function is to maximize the vibration reduction performance of the active suspension. The time-delay feedback coefficient and time delay are the optimization variables. An optimization algorithm is used to find the optimal time-delay feedback coefficient and optimal time delay in the initial calibration of the active suspension system's dynamic model. These include: Using the time-delay stability working boundary of the target active suspension bench test system as the upper bound constraint of the time delay, the time-delay feedback coefficient and the time delay as optimization variables, and the amplitude-frequency characteristics of the sprung mass acceleration on the road surface excitation as the vibration reduction performance index, an objective function that maximizes the vibration reduction performance index of the active suspension is constructed. From the initial calibration of the active suspension system dynamic model, a pair of time delay feedback coefficients and time delay quantities are extracted as individuals. All individuals are integrated into an initial population. The corresponding objective function value is calculated for each individual in the initial population. Based on the genetic algorithm, iterative genetic and mutation operations are performed on each individual to update the initial population until the objective function value of the updated population converges, thus obtaining the optimal individual. Extract the corresponding time delay feedback coefficient and time delay amount from the optimal individual, and use the time delay feedback coefficient and time delay amount extracted from the optimal individual as the optimal time delay feedback coefficient and optimal time delay amount.

[0011] In one possible design, a Smith prediction compensation model is constructed based on the optimal time delay feedback coefficient and the optimal time delay amount. The Smith prediction compensation model is used to perform compensation prediction, obtaining the predicted compensation output. Using the optimal time delay feedback coefficient and the optimal time delay amount as compensation references, the predicted compensation output is compared with the actual system output of the target active suspension bench test system to obtain a comparison error. A time delay compensation amount is generated based on the comparison error. This time delay compensation amount is then used to perform time delay compensation calibration on the preliminary calibrated active suspension system dynamic model, resulting in a time delay calibrated active suspension system dynamic model, including: Using the optimal time delay feedback coefficient and the optimal time delay amount as ideal time delay benchmarks, an optimal time delay feedback benchmark model is constructed, and a Smith prediction compensation model for the optimal time delay feedback benchmark model is constructed based on the time delay stability working boundary of the target active suspension bench test system and the dynamic model of the preliminary calibration active suspension system. Acquire the real-time control voltage signal and input the real-time control voltage signal into the Smith prediction compensation model, so as to use the Smith prediction compensation model to perform compensation prediction on the dynamic model of the preliminary calibration active suspension system and obtain the predicted compensation output. The real-time control voltage signal is input to the target active suspension bench test system to obtain the actual system output of the target active suspension bench test system; The estimated compensation output is compared with the system output and the difference is calculated to obtain the comparison error. The time delay compensation is generated based on the comparison error. The time delay compensation amount is input into the preliminary calibration active suspension system dynamic model, and the time delay parameters of the preliminary calibration active suspension system dynamic model are calibrated with time delay compensation to update the preliminary calibration active suspension system dynamic model to a time delay calibration active suspension system dynamic model.

[0012] In one possible design, the actuator control force output from the time-delay calibration active suspension system dynamic model is input to the target active suspension bench test system. A standard test signal is applied to the target active suspension bench test system, and time-domain response data is collected. Tolerance determination is performed based on the time-domain response data to complete the adaptive calibration of the active suspension system dynamic model parameters, including: The actuator control force output from the dynamic model of the time-delay calibrated active suspension system is input to the target active suspension bench test system, a standard test signal is applied to the target active suspension bench test system, and the time-domain response data of the target active suspension bench test system is collected synchronously. The actuator control force output from the dynamic model of the time-delay calibrated active suspension system is input to the optimal time-delay feedback reference model, a standard test signal is applied to the optimal time-delay feedback reference model, and reference response data of the optimal time-delay feedback reference model is collected synchronously. The time-domain response data is compared point by point with the reference response data to calculate the tracking deviation sequence, and the average absolute value of the tracking error is calculated based on the tracking deviation sequence. Obtain a preset tolerance range, and use the tolerance range to determine the tolerance of the absolute value of the tracking error; If the average absolute value of the tracking error belongs to the tolerance range, then the time-delay calibration active suspension system dynamic model is determined to be the final active suspension system dynamic model, and the parameter adaptive calibration of the active suspension system dynamic model is completed. If the average absolute value of the tracking error does not fall within the tolerance range, the gradient descent method is used to incrementally adjust the sensitive system parameters of the time-delay calibration active suspension system dynamic model, and a standard test signal is applied to the adjusted time-delay calibration active suspension system dynamic model to complete the tracking error sequence update until the average absolute value of the tracking error of the updated tracking error sequence converges to the preset tolerance range.

[0013] Secondly, the present invention provides a bench test system for adaptive calibration of active suspension system model parameters, comprising: The sensitive parameter extraction unit is used to establish a dynamic model of the active suspension system including time delay based on kinematic theorems, construct a system parameter sensitivity analysis function for the active suspension system dynamic model, solve the sensitivity index of each system parameter from the system parameter sensitivity analysis function, and select the sensitive system parameters of the active suspension system dynamic model. The time-delay boundary calculation unit is used to apply simulated road surface excitation to the target active suspension bench test system in order to collect the time-domain response data of the target active suspension bench test system and calculate the time-delay stable working boundary of the target active suspension bench test system through the time-domain response data. The preliminary calibration unit is used to adjust the working conditions and input signals of the target active suspension bench test system according to the sensitive system parameters of the active suspension system dynamic model. It collects multiple sets of response data from the target active suspension bench test system, calculates the real-time estimated values ​​of the sensitive system parameters based on each set of response data, and updates the active suspension system dynamic model using the real-time estimated values ​​of the sensitive system parameters to obtain the preliminary calibrated active suspension system dynamic model. The optimal time-delay optimization unit is used to optimize the objective function in the initial calibration of the active suspension system dynamic model by taking the time-delay stable working boundary of the target active suspension bench test system as the boundary condition, maximizing the active suspension vibration reduction performance index as the objective function, and using the time-delay feedback coefficient and time delay as optimization variables. The time delay compensation unit is used to construct a Smith prediction compensation model based on the optimal time delay feedback coefficient and the optimal time delay amount. The Smith prediction compensation model is used to perform compensation prediction on the preliminary calibration active suspension system dynamic model to obtain the predicted compensation output. The optimal time delay feedback coefficient and the optimal time delay amount are used as compensation references. The predicted compensation output is compared with the actual system output of the target active suspension bench test system to obtain the comparison error. The time delay compensation amount is generated based on the comparison error. The time delay compensation amount is used to perform time delay compensation calibration on the preliminary calibration active suspension system dynamic model to obtain the time delay calibration active suspension system dynamic model. The model tolerance determination unit is used to input the actuator control force output by the time-delay calibration active suspension system dynamic model into the target active suspension bench test system, apply standard test signals to the target active suspension bench test system and collect time-domain response data, so as to determine the tolerance based on the time-domain response data, so as to complete the parameter adaptive calibration of the active suspension system dynamic model.

[0014] Thirdly, the present invention provides an electronic device comprising a memory, a processor, and a transceiver connected in sequence and communication, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute a bench test method for adaptive calibration of active suspension system model parameters as described in the first aspect or any possible design of the first aspect.

[0015] Fourthly, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, perform a bench test method for adaptive calibration of active suspension system model parameters as described in the first aspect or any possible design of the first aspect.

[0016] Fifthly, the present invention provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform a bench test method for adaptive calibration of active suspension system model parameters as described in the first aspect or any possible design of the first aspect.

[0017] Beneficial Effects: This invention provides a bench test method and system for adaptive calibration of active suspension system model parameters. It uses a sensitivity analysis function to screen out sensitive system parameters that significantly affect output force, and then performs online identification and calibration of these sensitive parameters. This eliminates the need to identify all system parameters, greatly reducing the computational load of the calibration process and significantly improving test efficiency. Through time-delay boundary calculation, with the objective function of maximizing the active suspension's vibration reduction performance, the time-delay feedback coefficient and time-delay amount are optimized to obtain the optimal time-delay feedback coefficient and optimal time-delay amount under ideal conditions. Through adaptive adjustment, a Smith prediction compensation model is used for time-delay compensation calibration, resulting in a time-delay calibrated active suspension system dynamic model, significantly improving the adaptability to time-varying time delays and compensation accuracy. Furthermore, a standard test signal is applied to the target active suspension bench test system, and time-domain response data is collected. Tolerance judgment is performed based on the time-domain response data to complete the adaptive calibration of the active suspension system dynamic model parameters. This avoids pure simulation calibration deviating from physical reality, forming a closed-loop calibration data chain and achieving accurate adaptive calibration of the active suspension system model parameters. Attached Figure Description

[0018] Figure 1 A flowchart illustrating the bench test method for adaptive calibration of active suspension system model parameters provided in this embodiment of the invention; Figure 2 This is a functional structure diagram of a bench test system for adaptive calibration of active suspension system model parameters provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0020] It should be understood that although the terms first, second, etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit, without departing from the scope of the exemplary embodiments of the invention.

[0021] It should be understood that the term "and / or" that may appear in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist simultaneously. The term " / and" that may appear in this document describes another relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " that may appear in this document generally indicates that the related objects before and after it are in an "or" relationship.

[0022] Example: like Figure 1 As shown, the first aspect of this embodiment provides a bench test method for adaptive calibration of active suspension system model parameters, which may include, but is not limited to, the following steps: S1. Based on kinematic theorems, establish a dynamic model of an active suspension system including time delay, and construct a system parameter sensitivity analysis function for the active suspension system dynamic model. Solve the sensitivity index of each system parameter from the system parameter sensitivity analysis function, and select the sensitive system parameters of the active suspension system dynamic model. In one possible implementation, in step S1, based on kinematic theorems, a dynamic model of the active suspension system including time delay is established, and a system parameter sensitivity analysis function is constructed for the active suspension system dynamic model. The sensitivity indices of each system parameter are solved from the system parameter sensitivity analysis function, and the sensitive system parameters of the active suspension system dynamic model are selected. This can be decomposed into, but is not limited to, the following steps S11-S15, specifically including: S11. Obtain the suspension mechanical topology in the target active suspension bench test system, and take the road surface roughness excitation as the model input, and the sprung mass displacement and unsprung mass displacement as the model output. Based on the kinematic theorem, construct a dynamic model of the active suspension system including time delay. S12. Based on the dynamic model of the active suspension system, establish a driving force output sub-model between the control voltage signal and the actuator driving force in the active actuator of the target active suspension bench test system; S13. Using the driving force output sub-model, the dynamic model of the active suspension system is organized into a first-order state-space equation, and each undetermined physical parameter is extracted from the first-order state-space equation as the system parameter of the dynamic model of the active suspension system. S14. For the first-order state-space equation, calculate the parameter vector partial derivatives of each system parameter to form a system parameter sensitivity analysis function, and solve the sensitivity of the system parameter sensitivity analysis function to obtain the sensitivity index of each system parameter. S15. Obtain a preset sensitivity threshold, and based on the sensitivity threshold, compare the sensitivity indices of each system parameter to select each system parameter whose sensitivity index exceeds the sensitivity threshold as the sensitive system parameter of the active suspension system dynamic model.

[0023] In practical applications, it is necessary to determine the physical composition of a 1 / 4 vehicle two-degree-of-freedom active suspension as the suspension mechanical topology in the target active suspension bench test system. This includes the connection relationships of sprung mass, unsprung mass, suspension springs, base damping elements, and active actuators. Using the sprung mass displacement and unsprung mass displacement as the model output coordinates, and road roughness excitation as the external disturbance input to the model, force balance equations are established on the sprung mass and unsprung mass based on Newton's second law of motion, forming an initial set of dynamic equations for the active suspension system. A time delay is introduced into the active control force term in this initial set of dynamic equations, characterizing the actual output force of the active actuator as a function of the ideal control force and the time delay. This transforms the initial set of dynamic equations for the active suspension system into a set of constant-coefficient linear differential equations including time delay. This set of constant-coefficient linear differential equations serves as the dynamic model of the active suspension system including time delay, completing the model construction and characterizing the time delay effects caused by signal measurement and transmission, controller operation, and hydraulic response.

[0024] Based on the type of active actuator in the target active suspension bench test system, the flow equations for the servo valve (reflecting the relationship between valve core displacement and load flow), the flow continuity equations for the two chambers of the actuator (reflecting pressure dynamic characteristics), and the force balance equation for the piston rod are established. Combining these three equations forms the driving force output sub-model of the active actuator, which fully reflects the physical mapping relationship between the input control voltage signal and the output active driving force. By combining the active suspension system dynamics model with the driving force output sub-model, key undetermined physical quantities in the active suspension system dynamics model (such as sprung mass velocity, unsprung mass velocity, actuator two-chamber pressure, piston rod displacement, etc.) are selected as system state variables. This transforms the active suspension system dynamics model into a standard first-order state-space equation with the control voltage signal as input and the actuator driving force as output. All undetermined physical parameters in the active suspension system dynamics model are then used as system parameters (vectors).

[0025] It should be noted that in the bench test method provided in this embodiment, the system parameter sensitivity analysis function is actually the partial derivative of the system state variable with respect to the system parameters. Specifically, the partial derivatives of the state vector and the parameter vector are obtained respectively for the established first-order state space equation to form a first-order trajectory sensitivity differential equation. The first-order trajectory sensitivity differential equation is used as the system parameter sensitivity analysis function. The coefficient term of the system parameter sensitivity analysis function reflects the influence of the state change of the active suspension system dynamic model on the sensitivity, while the free term reflects the influence of the system parameter change on the state of the active suspension system dynamic model. When solving the sensitivity analysis function of the system parameters, a standard step excitation signal is applied to the dynamic model of the active suspension system. The sensitivity equation is solved using the numerical integration method to obtain the response curve of the sensitivity of each system parameter over time. Based on the response curve of the sensitivity of each system parameter over time, the change in output caused by each system parameter is extracted, and its percentage relative to the steady-state output of the system is used as the sensitivity index of each system parameter. The sensitivity index of each system parameter is compared with a preset sensitivity threshold. System parameters that exceed the sensitivity threshold are identified as sensitive system parameters that significantly affect the dynamic characteristics of the dynamic model of the active suspension system, and a list of sensitive system parameters is output as the basis for effective parameters in subsequent calibration.

[0026] S2. Apply simulated road surface excitation to the target active suspension bench test system to collect the time-domain response data of the target active suspension bench test system, and calculate the time-delay stable working boundary of the target active suspension bench test system through the time-domain response data; In one possible implementation, step S2, applying simulated road surface excitation to the target active suspension bench test system to collect time-domain response data of the target active suspension bench test system, and calculating the time-delay stable operating boundary of the target active suspension bench test system using the time-domain response data, can be decomposed into, but is not limited to, the following steps S21-S27, specifically including: S21. Apply a sinusoidal sweep frequency excitation signal to the target active suspension bench test system as a simulated road surface excitation, and collect the time-domain response data output by the target active suspension bench test system in real time; S22. Preprocess the time-domain response data, and extract the corresponding amplitude-frequency response features and phase-frequency response features from the preprocessed time-domain response data; S23. Input the amplitude frequency response characteristics and the phase frequency response characteristics into the dynamic model of the active suspension system to obtain the theoretical critical time delay value under the current controllable damping conditions; S24. Adjust the equivalent time delay in the target active suspension bench test system, and repeatedly perform simulated road excitation input to obtain the time delay corresponding to the first occurrence of continuous constant amplitude oscillation in the target active suspension bench test system. Use the time delay corresponding to the first occurrence of continuous constant amplitude oscillation in the target active suspension bench test system as the measured critical time delay value. S25. Obtain a preset time delay deviation threshold, calculate the time delay deviation between the theoretical critical time delay value and the measured critical time delay value, and compare the time delay deviation threshold with the time delay deviation to determine the reliability of the active suspension system dynamic model. S26. If the time delay deviation does not exceed the time delay deviation threshold, the dynamic model of the active suspension system is determined to be reliable, and the theoretical critical time delay value is used as the time delay stability working boundary of the target active suspension bench test system. S27. If the time delay deviation exceeds the time delay deviation threshold, the active suspension system dynamic model is determined to be unreliable. The parameters of the active suspension system dynamic model are adjusted using the measured critical time delay value, and the theoretical critical time delay value is calculated again for the active suspension system dynamic model after parameter adjustment. The time delay deviation is calculated and the model reliability is determined for the recalculated theoretical critical time delay value until the active suspension system dynamic model is determined to be reliable.

[0027] In specific application scenarios, before proceeding to step S21, the mechanical assembly and hydraulic pipeline connection of the target active suspension bench test system need to be completed according to preset structural parameters. This target active suspension bench test system may include, but is not limited to, an electric vibration table, upper and lower crossbeams, a slider guide mechanism, a sprung mass counterweight, coil springs, shock absorbers, an electro-hydraulic actuator, and various sensors. Static calibration is then performed on the force sensor, displacement sensor, and acceleration sensor to ensure that the measurement accuracy of each sensor meets the bench test requirements. The hydraulic system needs to be vented and preheated to stabilize the oil temperature to the operating temperature, ensuring the consistency of hydraulic parameters.

[0028] In addition, when applying simulated road surface excitation to the target active suspension bench test system, it is necessary to control the electric vibration table in the target active suspension bench test system to apply a sinusoidal sweep frequency excitation signal with constant amplitude and continuously changing frequency at a preset rate to the target active suspension bench test system within a preset sweep frequency range, so as to simulate road surface input at different frequencies. At this time, it is necessary to provide various sensors to synchronously collect the time-domain response signals of sprung mass acceleration, suspension dynamic deflection, tire dynamic load and actuator output force in the target active suspension bench test system. The sampling frequency is set to 5-10 times the highest operating frequency of the target active suspension bench test system to ensure that key dynamic information is not lost.

[0029] The time-domain response data is preprocessed (by filtering and denoising to remove high-frequency noise and DC bias), and the amplitude-frequency response characteristics and phase-frequency response characteristics at different excitation frequencies are extracted from the preprocessed time-domain response data using spectral analysis. These characteristics are then substituted into the dynamic model of the active suspension system to obtain the theoretical critical time delay value under the current controllable damping conditions. Simultaneously, in the frequency sweep test, the equivalent time delay of the control loop of the target active suspension bench test system is progressively increased (simulated through a controller delay element). The evolution of the target active suspension bench test system response from convergence (stability) to constant-amplitude oscillation (critical stability) and then to divergence (instability) is observed. The time delay value corresponding to the first occurrence of sustained constant-amplitude oscillation in the target active suspension bench test system is determined as the measured critical time delay value.

[0030] It should be noted that when determining the reliability of the active suspension system dynamic model, if the time delay deviation exceeds the preset time delay deviation threshold, the active suspension system dynamic model is determined to have an error and is unreliable. It is necessary to use the measured critical time delay value as a benchmark to reverse-correct the key parameters (such as equivalent damping and hydraulic volume) in the active suspension system dynamic model, so that the active suspension system dynamic model after parameter adjustment is closer to the response characteristics of the actual target active suspension bench test system, that is, to complete the original parameter calibration of the target active suspension bench test system.

[0031] S3. Based on the sensitive system parameters of the active suspension system dynamic model, adjust the working conditions and input signals of the target active suspension bench test system, collect multiple sets of response data from the target active suspension bench test system, calculate the real-time estimated values ​​of the sensitive system parameters based on each set of response data, and update the active suspension system dynamic model using the real-time estimated values ​​of the sensitive system parameters to obtain a preliminary calibrated active suspension system dynamic model; In one possible implementation, in step S3, based on the sensitive system parameters of the active suspension system dynamic model, the operating conditions and input signals of the target active suspension bench test system are adjusted. Multiple sets of response data are collected from the target active suspension bench test system. Based on each set of response data, real-time estimates of the sensitive system parameters are calculated. The active suspension system dynamic model is updated using the real-time estimates of the sensitive system parameters to obtain a preliminary calibrated active suspension system dynamic model. This can be, but is not limited to, decomposed into the following steps S31-S35, specifically including: S31. Generate corresponding identification conditions for each sensitive system parameter of the active suspension system dynamic model, and adjust the working conditions and input signals of the target active suspension bench test system based on the identification conditions corresponding to each sensitive system parameter, so as to complete multiple tests of the target active suspension bench test system, and collect the control voltage signal and system output signal of each test as response data; S32. The control voltage signal in each group of response data is used as an input quantity and input to the active suspension system dynamic model to obtain the driving force output by the active suspension system dynamic model as the model output force; S33. Extract the actuator driving force corresponding to each group of response data from the system output signal in each group of response data as the actual output force, calculate the square integral of the deviation between the model output force and the actual output force corresponding to each group of response data, and use the square integral of the deviation between the model output force and the actual output force as the identification target function. S34. Obtain a preset convergence threshold, iteratively solve the identification objective function corresponding to each sensitive system parameter until the identification objective function value corresponding to each sensitive system parameter is lower than the convergence threshold, and obtain the sensitive system parameter identification value corresponding to each sensitive system parameter; S35. The sensitive system parameter identification values ​​corresponding to each sensitive system parameter are used as preliminary calibration parameter values ​​and input into the active suspension system dynamic model. The parameters of each sensitive system parameter in the active suspension system dynamic model are replaced to complete the update of the active suspension system dynamic model and obtain the preliminary calibration active suspension system dynamic model.

[0032] It should be noted that in the bench test method provided in this embodiment, the sensitive system parameters include at least flow gain, effective bulk modulus, actuator effective area, and total flow pressure coefficient. For each sensitive system parameter, differentiated identification conditions need to be designed. Specifically: for flow gain and total flow pressure coefficient, multiple sets of step excitation tests under different oil supply pressures and different valve core openings are designed; for effective bulk modulus, sinusoidal sweep frequency excitation tests under different load forces and different piston positions are designed; for actuator effective area, step response tests are designed when the piston rod is at different positions in its stroke. This ensures that the amplitude and frequency range of each excitation signal are sufficient to fully excite the dynamic characteristics of the corresponding sensitive system parameter. Based on the identification conditions corresponding to each sensitive system parameter, each set of bench tests is sequentially executed on the target active suspension bench test system. For each set of bench tests, control voltage signals (such as servo valve control voltage or valve core displacement commands) and system output signals (such as actuator output force, two-chamber pressure, and piston rod displacement) are synchronously acquired, and timestamps are recorded to ensure strict time alignment of input and output data. Each test group was repeated at least three times, and the average value was taken as the final response data to reduce the impact of random noise on the bench test.

[0033] The active suspension system dynamics model is used as the identification framework. The actual control voltage signal collected in the response data is used as the model input, and the values ​​to be identified for each sensitive system parameter (sensitive system parameter value) are used as variables to drive the active suspension system dynamics model to obtain the active actuator driving force output by the model. This force is used as the model output force, and the square integral of the deviation between the model output force and the actual collected output force is calculated as the identification objective function. That is, the smaller the value of the identification objective function, the closer the sensitive system parameter to be identified is to the true value. In addition, based on the positive and negative directions and relative magnitudes of the influence of each sensitive parameter on the output force (extracted from the response curve of sensitivity changing with time in step S15), the initial identification values ​​and search boundaries of each sensitive system parameter are set (the search boundary is set to within ±30% of the design value) to ensure that the identification results have clear upper and lower limit protection.

[0034] In specific application scenarios, the identification objective function is solved iteratively using a numerical optimization algorithm. In each iteration, the sensitive system parameters to be identified are updated, and the corresponding identification objective function value is recalculated until the identification objective function value is less than a preset convergence threshold (or the number of iterations reaches the upper limit). During this process, if the identification result of a certain sensitive system parameter shows significant dispersion in multiple repeated experiments (standard deviation exceeding 10% of its mean), the parameter is marked as a low-identifiability parameter, and its theoretical design value is used instead of the identified value in subsequent step S35.

[0035] The identified values ​​of each sensitive parameter obtained under various identification conditions are comprehensively processed: for sensitive system parameters with multiple sets of identification results, their arithmetic mean is taken as the final preliminary calibration parameter value; for sensitive system parameters marked as low identification parameters, the weighted average of the theoretical design value and the identification mean (the weight ratio is determined according to the accuracy of the bench sensor) is taken as the final preliminary calibration parameter value. The preliminary calibration parameter values ​​of all sensitive system parameters are substituted into the active suspension system dynamic model, replacing the original system parameters, to complete the preliminary calibration of the active suspension system dynamic model, resulting in the preliminary calibrated active suspension system dynamic model.

[0036] S4. Using the time-delay stable working boundary of the target active suspension bench test system as the boundary condition, the objective function is to maximize the vibration reduction performance index of the active suspension. The time-delay feedback coefficient and the time delay are the optimization variables. In the preliminary calibration of the dynamic model of the active suspension system, the optimization algorithm is used to optimize the objective function to obtain the optimal time-delay feedback coefficient and the optimal time delay. In one possible implementation, step S4 uses the time-delay stable working boundary of the target active suspension bench test system as the boundary condition, maximizes the vibration reduction performance index of the active suspension as the objective function, and uses the time-delay feedback coefficient and time delay as optimization variables. An optimization algorithm is used to optimize the objective function in the preliminary calibration of the active suspension system dynamic model to obtain the optimal time-delay feedback coefficient and optimal time delay. This can be decomposed into, but is not limited to, the following steps S41-S43, specifically including: S41. The time-delay stable working boundary of the target active suspension bench test system is used as the upper bound constraint of the time delay, the time delay feedback coefficient and the time delay are used as optimization variables, and the amplitude-frequency characteristics of the sprung mass acceleration on the road surface excitation are used as the vibration reduction performance index to construct the objective function that maximizes the vibration reduction performance index of the active suspension. S42. From the initial calibration of the active suspension system dynamic model, extract a pair of time delay feedback coefficients and time delay as an individual, integrate all individuals into an initial population, calculate the corresponding objective function value for each individual in the initial population, and perform iterative genetic and mutation operations on each individual based on a genetic algorithm to update the initial population until the updated population objective function value converges, thus obtaining the optimal individual; S43. Extract the corresponding time delay feedback coefficient and time delay amount from the optimal individual, and use the time delay feedback coefficient and time delay amount extracted from the optimal individual as the optimal time delay feedback coefficient and optimal time delay amount.

[0037] It should be noted that the preliminary calibration of the active suspension system dynamic model serves as a constraint platform for objective function optimization. A smaller objective function value indicates a better vibration reduction effect of the preliminary calibration of the active suspension system dynamic model. An initial population containing multiple individuals is initialized within the constraint space (state space formed by each individual) created by the preliminary calibration of the active suspension system dynamic model. Each individual consists of a set of time-delay feedback coefficients and time-delay parameters. The objective function value for each individual is calculated. Based on the objective function value, selection, crossover, and mutation operations are performed to generate a new generation of population. This process is repeated iteratively until the objective function value converges (or the preset number of iterations is reached), at which point the process stops. The parameter combination with the optimal objective function value is output as the optimal time-delay feedback coefficient and the optimal time-delay.

[0038] In a possible implementation, the obtained optimal time-delay feedback coefficient and optimal time-delay amount are substituted into the dynamic model of the initial calibration active suspension system. Under typical road excitation, the key performance indicators of the suspension (including sprung mass acceleration, suspension dynamic deflection, and tire dynamic load) are simulated and calculated, and compared with the uncontrolled state. If the vibration reduction effect meets the preset vibration reduction requirements, the optimal solution (optimal time-delay feedback coefficient and optimal time-delay amount) is confirmed to be effective. If the preset vibration reduction requirements are not met, the parameters of the genetic algorithm are adjusted and steps S42-S43 are re-executed until the vibration reduction effect meets the preset vibration reduction requirements.

[0039] S5. Construct a Smith prediction compensation model based on the optimal time delay feedback coefficient and the optimal time delay amount. Use the Smith prediction compensation model to perform compensation prediction on the preliminary calibration active suspension system dynamic model to obtain the predicted compensation output. Use the optimal time delay feedback coefficient and the optimal time delay amount as compensation references. Compare the predicted compensation output with the actual system output of the target active suspension bench test system to obtain the comparison error. Generate a time delay compensation amount based on the comparison error. Use the time delay compensation amount to perform time delay compensation calibration on the preliminary calibration active suspension system dynamic model to obtain the time delay calibration active suspension system dynamic model. In one possible implementation, step S5 involves constructing a Smith prediction compensation model based on the optimal time delay feedback coefficient and the optimal time delay amount. The Smith prediction compensation model is then used to perform compensation prediction, yielding a predicted compensation output. Using the optimal time delay feedback coefficient and the optimal time delay amount as compensation references, the predicted compensation output is compared with the actual system output of the target active suspension bench test system to obtain a comparison error. A time delay compensation amount is generated based on the comparison error, and this time delay compensation amount is used to perform time delay compensation calibration on the preliminary calibrated active suspension system dynamic model, resulting in a time delay calibrated active suspension system dynamic model. This step can be, but is not limited to, decomposed into the following steps S51-S55, specifically including: S51. Using the optimal time delay feedback coefficient and the optimal time delay amount as ideal time delay benchmarks, construct an optimal time delay feedback benchmark model, and construct a Smith prediction compensation model for the optimal time delay feedback benchmark model based on the time delay stability working boundary of the target active suspension bench test system and the dynamic model of the preliminary calibration active suspension system. S52. Obtain the real-time control voltage signal and input the real-time control voltage signal into the Smith prediction compensation model to use the Smith prediction compensation model to perform compensation prediction on the preliminary calibration of the active suspension system dynamic model and obtain the predicted compensation output. S53. Input the real-time control voltage signal into the target active suspension bench test system to obtain the actual system output of the target active suspension bench test system; S54. Compare the estimated compensation output with the system output and calculate the difference to obtain the comparison error. Generate the time delay compensation based on the comparison error. S55. Input the time delay compensation amount into the preliminary calibration active suspension system dynamic model, perform time delay compensation calibration on the time delay parameters of the preliminary calibration active suspension system dynamic model, so as to update the preliminary calibration active suspension system dynamic model into a time delay calibration active suspension system dynamic model.

[0040] In specific application scenarios, the Smith prediction and compensation model pre-defines a target mathematical model describing the dynamic characteristics of the target active suspension bench test system and a time delay estimation model that estimates the time delay of the current target active suspension bench test system in real time. Based on the estimated time delay, the time delay parameters in the target mathematical model are dynamically adjusted to ensure they always match the actual target active suspension bench test system. In traditional Smith predictors, the time delay compensation stage is fixed. This embodiment uses an integrator and a multiplier to replace the fixed stage, allowing the time delay parameters to be adjusted online. The multiplier's output can adaptively adjust the compensation amount based on the error signal without manual intervention.

[0041] S6. Input the actuator control force output by the time-delay calibration active suspension system dynamic model into the target active suspension bench test system, apply a standard test signal to the target active suspension bench test system and collect time-domain response data, so as to make tolerance judgment based on the time-domain response data, so as to complete the parameter adaptive calibration of the active suspension system dynamic model.

[0042] In one possible implementation, step S6 involves inputting the actuator control force output from the time-delay calibration active suspension system dynamic model into the target active suspension bench test system, applying a standard test signal to the target active suspension bench test system, and collecting time-domain response data. Tolerance determination is then performed based on the time-domain response data to complete the adaptive calibration of the active suspension system dynamic model's parameters. This step can be broken down into, but is not limited to, the following steps S61-S66, specifically including: S61. Input the actuator control force output by the time-delay calibration active suspension system dynamic model to the target active suspension bench test system, apply a standard test signal to the target active suspension bench test system, and simultaneously collect the time-domain response data of the target active suspension bench test system; S62. Input the actuator control force output by the time-delay calibration active suspension system dynamic model to the optimal time-delay feedback reference model, apply a standard test signal to the optimal time-delay feedback reference model, and simultaneously collect the reference response data of the optimal time-delay feedback reference model; S63. The time-domain response data is compared point by point with the reference response data to calculate the tracking deviation sequence, and the average absolute value of the tracking error is calculated based on the tracking deviation sequence. S64. Obtain a preset tolerance range, and use the tolerance range to determine the tolerance of the absolute value of the tracking error; S65. If the average absolute value of the tracking error belongs to the tolerance range, then the time-delay calibration active suspension system dynamic model is determined to be the final active suspension system dynamic model, and the parameter adaptive calibration of the active suspension system dynamic model is completed. S66. If the average absolute value of the tracking error does not belong to the tolerance range, the gradient descent method is used to incrementally adjust the sensitive system parameters of the time-delay calibration active suspension system dynamic model, and a standard test signal is applied to the adjusted time-delay calibration active suspension system dynamic model to complete the tracking error sequence update until the average absolute value of the tracking error of the updated tracking error sequence converges to the preset tolerance range.

[0043] like Figure 2 As shown, the second aspect of this embodiment provides a hardware system for implementing the bench test method for adaptive calibration of active suspension system model parameters as described in the first aspect of the embodiment, including: The sensitive parameter extraction unit is used to establish a dynamic model of the active suspension system including time delay based on kinematic theorems, construct a system parameter sensitivity analysis function for the active suspension system dynamic model, solve the sensitivity index of each system parameter from the system parameter sensitivity analysis function, and select the sensitive system parameters of the active suspension system dynamic model. The time-delay boundary calculation unit is used to apply simulated road surface excitation to the target active suspension bench test system in order to collect the time-domain response data of the target active suspension bench test system and calculate the time-delay stable working boundary of the target active suspension bench test system through the time-domain response data. The preliminary calibration unit is used to adjust the working conditions and input signals of the target active suspension bench test system according to the sensitive system parameters of the active suspension system dynamic model. It collects multiple sets of response data from the target active suspension bench test system, calculates the real-time estimated values ​​of the sensitive system parameters based on each set of response data, and updates the active suspension system dynamic model using the real-time estimated values ​​of the sensitive system parameters to obtain the preliminary calibrated active suspension system dynamic model. The optimal time-delay optimization unit is used to optimize the objective function in the initial calibration of the active suspension system dynamic model by taking the time-delay stable working boundary of the target active suspension bench test system as the boundary condition, maximizing the active suspension vibration reduction performance index as the objective function, and using the time-delay feedback coefficient and time delay as optimization variables. The time delay compensation unit is used to construct a Smith prediction compensation model based on the optimal time delay feedback coefficient and the optimal time delay amount. The Smith prediction compensation model is used to perform compensation prediction on the preliminary calibration active suspension system dynamic model to obtain the predicted compensation output. The optimal time delay feedback coefficient and the optimal time delay amount are used as compensation references. The predicted compensation output is compared with the actual system output of the target active suspension bench test system to obtain the comparison error. The time delay compensation amount is generated based on the comparison error. The time delay compensation amount is used to perform time delay compensation calibration on the preliminary calibration active suspension system dynamic model to obtain the time delay calibration active suspension system dynamic model. The model tolerance determination unit is used to input the actuator control force output by the time-delay calibration active suspension system dynamic model into the target active suspension bench test system, apply standard test signals to the target active suspension bench test system and collect time-domain response data, so as to determine the tolerance based on the time-domain response data, so as to complete the parameter adaptive calibration of the active suspension system dynamic model.

[0044] The working process, working details and technical effects of the system provided in this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.

[0045] like Figure 3As shown, the third aspect of this embodiment provides an electronic device, including: a memory, a processor, and a transceiver that are sequentially and communicatively connected, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the bench test method for adaptive calibration of active suspension system model parameters as described in the first aspect of the embodiment.

[0046] For specific examples, the memory may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or first-in-last-out (FILO) memory, etc.; specifically, the processor may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor may be implemented using at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), PLA (Programmable Logic Array). The processor may also include a main processor and a coprocessor. The main processor, also known as the CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state.

[0047] In some embodiments, the processor may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. For example, the processor may not be limited to microprocessors of the STM32F105 series, reduced instruction set computer (RISC) microprocessors, x86 architecture processors, or processors with integrated neural network processing units (NPUs). The transceiver may be, but is not limited to, a Wi-Fi transceiver, a Bluetooth transceiver, a General Packet Radio Service (GPRS) transceiver, a ZigBee (a low-power LAN protocol based on the IEEE 802.15.4 standard) transceiver, a 3G transceiver, a 4G transceiver, and / or a 5G transceiver. Furthermore, the electronic device may also include, but is not limited to, a power module, a display screen, and other necessary components.

[0048] The working process, working details and technical effects of the electronic device provided in this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.

[0049] The fourth aspect of this embodiment provides a storage medium that stores instructions for a bench test method for adaptive calibration of active suspension system model parameters as described in the first aspect of the embodiment. That is, the storage medium stores instructions that, when executed on a computer, perform the bench test method for adaptive calibration of active suspension system model parameters as described in the first aspect of the embodiment.

[0050] The storage medium refers to a carrier for storing data, which may include, but is not limited to, floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or memory sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0051] The working process, working details and technical effects of the storage medium provided in this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.

[0052] The fifth aspect of this embodiment provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform a bench test method for adaptive calibration of active suspension system model parameters as described in the first aspect of this embodiment, wherein the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0053] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A bench test method for adaptive calibration of active suspension system model parameters, characterized in that, include: Based on kinematic theorems, a dynamic model of an active suspension system including time delay is established, and a system parameter sensitivity analysis function is constructed for the dynamic model of the active suspension system. The sensitivity index of each system parameter is solved from the system parameter sensitivity analysis function, and the sensitive system parameters of the active suspension system dynamic model are selected. Simulated road surface excitation is applied to the target active suspension bench test system to collect the time-domain response data of the target active suspension bench test system, and the time-delay stability working boundary of the target active suspension bench test system is calculated through the time-domain response data; Based on the sensitive system parameters of the active suspension system dynamic model, the working conditions and input signals of the target active suspension bench test system are adjusted. Multiple sets of response data are collected from the target active suspension bench test system. Based on each set of response data, the real-time estimated values ​​of the sensitive system parameters are calculated. The active suspension system dynamic model is updated using the real-time estimated values ​​of the sensitive system parameters to obtain a preliminary calibrated active suspension system dynamic model. Using the time-delay stable working boundary of the target active suspension bench test system as the boundary condition, the objective function is to maximize the vibration reduction performance index of the active suspension. The time-delay feedback coefficient and time delay are the optimization variables. In the preliminary calibration of the dynamic model of the active suspension system, the optimization algorithm is used to optimize the objective function to obtain the optimal time-delay feedback coefficient and the optimal time delay. A Smith prediction compensation model is constructed based on the optimal time delay feedback coefficient and the optimal time delay amount. The Smith prediction compensation model is used to perform compensation prediction on the preliminary calibration active suspension system dynamic model to obtain the predicted compensation output. The optimal time delay feedback coefficient and the optimal time delay amount are used as compensation references. The predicted compensation output is compared with the actual system output of the target active suspension bench test system to obtain the comparison error. The time delay compensation amount is generated based on the comparison error. The time delay compensation amount is used to perform time delay compensation calibration on the preliminary calibration active suspension system dynamic model to obtain the time delay calibration active suspension system dynamic model. The actuator control force output from the time-delay calibration active suspension system dynamic model is input into the target active suspension bench test system. A standard test signal is applied to the target active suspension bench test system and time-domain response data is collected. Tolerance determination is performed based on the time-domain response data to complete the parameter adaptive calibration of the active suspension system dynamic model.

2. The bench test method for adaptive calibration of active suspension system model parameters according to claim 1, characterized in that, Based on kinematic theorems, a dynamic model of an active suspension system including time delay is established. A system parameter sensitivity analysis function is constructed for this model. Sensitivity indices for each system parameter are solved from this function. Sensitive system parameters of the active suspension system dynamic model are then selected, including: The suspension mechanical topology of the target active suspension bench test system is obtained, and the road surface roughness excitation is used as the model input, and the sprung mass displacement and unsprung mass displacement are used as the model output. Based on the kinematic theorem, a dynamic model of the active suspension system including time delay is constructed. Based on the dynamic model of the active suspension system, a driving force output sub-model between the control voltage signal and the actuator driving force is established for the active actuator in the target active suspension bench test system. Using the driving force output sub-model, the dynamic model of the active suspension system is organized into a first-order state-space equation, and the undetermined physical parameters are extracted from the first-order state-space equation as the system parameters of the dynamic model of the active suspension system. For the first-order state-space equation, the parameter vector partial derivatives of each system parameter are calculated to form a system parameter sensitivity analysis function, and the sensitivity of the system parameter sensitivity analysis function is solved to obtain the sensitivity index of each system parameter. A preset sensitivity threshold is obtained. Based on the sensitivity threshold, the sensitivity indices of each system parameter are compared, and each system parameter whose sensitivity index exceeds the sensitivity threshold is selected as the sensitive system parameter of the active suspension system dynamic model.

3. The bench test method for adaptive calibration of active suspension system model parameters according to claim 1, characterized in that, Simulated road surface excitation is applied to the target active suspension bench test system to collect its time-domain response data. The time-delay stability operating boundary of the target active suspension bench test system is then calculated using this data, including: A sinusoidal sweep frequency excitation signal is applied to the target active suspension bench test system as a simulated road surface excitation, and the time-domain response data output by the target active suspension bench test system is collected in real time. The time-domain response data is preprocessed, and the corresponding amplitude-frequency response features and phase-frequency response features are extracted from the preprocessed time-domain response data. The amplitude frequency response characteristics and the phase frequency response characteristics are input into the dynamic model of the active suspension system to obtain the theoretical critical time delay value under the current controllable damping conditions; Adjust the equivalent time delay in the target active suspension bench test system, and repeatedly perform simulated road excitation input to obtain the time delay corresponding to the first occurrence of continuous constant amplitude oscillation in the target active suspension bench test system. Use the time delay corresponding to the first occurrence of continuous constant amplitude oscillation in the target active suspension bench test system as the measured critical time delay value. A preset time delay deviation threshold is obtained, the time delay deviation between the theoretical critical time delay value and the measured critical time delay value is calculated, and the time delay deviation threshold and the time delay deviation are compared to determine the reliability of the active suspension system dynamic model. If the time delay deviation does not exceed the time delay deviation threshold, the dynamic model of the active suspension system is determined to be reliable, and the theoretical critical time delay value is used as the time delay stability working boundary of the target active suspension bench test system. If the time delay deviation exceeds the time delay deviation threshold, the active suspension system dynamic model is determined to be unreliable. The parameters of the active suspension system dynamic model are adjusted using the measured critical time delay value, and the theoretical critical time delay value is calculated again for the active suspension system dynamic model after parameter adjustment. The time delay deviation is calculated and the model reliability is determined for the recalculated theoretical critical time delay value until the active suspension system dynamic model is determined to be reliable.

4. The bench test method for adaptive calibration of active suspension system model parameters according to claim 1, characterized in that, Based on the sensitive system parameters of the active suspension system dynamic model, the operating conditions and input signals of the target active suspension bench test system are adjusted. Multiple sets of response data are collected from the target active suspension bench test system. Real-time estimates of the sensitive system parameters are calculated based on each set of response data. The active suspension system dynamic model is updated using the real-time estimates of the sensitive system parameters to obtain a preliminary calibrated active suspension system dynamic model, including: For each sensitive system parameter of the active suspension system dynamic model, corresponding identification conditions are generated. Based on the identification conditions corresponding to each sensitive system parameter, the working conditions and input signals of the target active suspension bench test system are adjusted to complete multiple tests of the target active suspension bench test system. The control voltage signal and system output signal of each test are collected as response data. The control voltage signal in each group of response data is used as an input quantity and input to the active suspension system dynamic model to obtain the driving force output by the active suspension system dynamic model as the model output force; From the system output signal in each group of response data, the actuator driving force corresponding to each group of response data is extracted as the actual output force. The square integral of the deviation between the model output force and the actual output force corresponding to each group of response data is calculated. The square integral of the deviation between the model output force and the actual output force is used as the identification target function. A preset convergence threshold is obtained, and the identification objective function corresponding to each sensitive system parameter is iteratively solved until the identification objective function value corresponding to each sensitive system parameter is lower than the convergence threshold, thereby obtaining the sensitive system parameter identification value corresponding to each sensitive system parameter. The sensitive system parameter identification values ​​corresponding to each sensitive system parameter are used as preliminary calibration parameter values ​​and input into the active suspension system dynamic model. The parameters of each sensitive system parameter in the active suspension system dynamic model are replaced to complete the update of the active suspension system dynamic model and obtain the preliminary calibration active suspension system dynamic model.

5. The bench test method for adaptive calibration of active suspension system model parameters according to claim 1, characterized in that, Using the time-delay stability working boundary of the target active suspension bench test system as the boundary condition, maximizing the vibration reduction performance index of the active suspension as the objective function, and the time-delay feedback coefficient and time delay as optimization variables, an optimization algorithm is used to optimize the objective function in the preliminary calibration of the active suspension system dynamic model to obtain the optimal time-delay feedback coefficient and optimal time delay, including: Using the time-delay stability working boundary of the target active suspension bench test system as the upper bound constraint of the time delay, the time-delay feedback coefficient and the time delay as optimization variables, and the amplitude-frequency characteristics of the sprung mass acceleration on the road surface excitation as the vibration reduction performance index, an objective function that maximizes the vibration reduction performance index of the active suspension is constructed. From the initial calibration of the active suspension system dynamic model, a pair of time delay feedback coefficients and time delay quantities are extracted as individuals. All individuals are integrated into an initial population. The corresponding objective function value is calculated for each individual in the initial population. Based on the genetic algorithm, iterative genetic and mutation operations are performed on each individual to update the initial population until the objective function value of the updated population converges, thus obtaining the optimal individual. Extract the corresponding time delay feedback coefficient and time delay amount from the optimal individual, and use the time delay feedback coefficient and time delay amount extracted from the optimal individual as the optimal time delay feedback coefficient and optimal time delay amount.

6. The bench test method for adaptive calibration of active suspension system model parameters according to claim 1, characterized in that, A Smith-based prediction and compensation model is constructed based on the optimal time delay feedback coefficient and the optimal time delay. The Smith-based prediction and compensation model is used to predict the compensation output. Using the optimal time delay feedback coefficient and the optimal time delay as compensation references, the predicted compensation output is compared with the actual system output of the target active suspension bench test system to obtain the comparison error. A time delay compensation amount is generated based on the comparison error. This time delay compensation amount is then used to perform time delay compensation calibration on the preliminary calibrated active suspension system dynamic model, resulting in a time delay calibrated active suspension system dynamic model, including: Using the optimal time delay feedback coefficient and the optimal time delay amount as ideal time delay benchmarks, an optimal time delay feedback benchmark model is constructed, and a Smith prediction compensation model for the optimal time delay feedback benchmark model is constructed based on the time delay stability working boundary of the target active suspension bench test system and the dynamic model of the preliminary calibration active suspension system. Acquire the real-time control voltage signal and input the real-time control voltage signal into the Smith prediction compensation model, so as to use the Smith prediction compensation model to perform compensation prediction on the dynamic model of the preliminary calibration active suspension system and obtain the predicted compensation output. The real-time control voltage signal is input to the target active suspension bench test system to obtain the actual system output of the target active suspension bench test system; The estimated compensation output is compared with the system output and the difference is calculated to obtain the comparison error. The time delay compensation is generated based on the comparison error. The time delay compensation amount is input into the preliminary calibration active suspension system dynamic model, and the time delay parameters of the preliminary calibration active suspension system dynamic model are calibrated with time delay compensation to update the preliminary calibration active suspension system dynamic model to a time delay calibration active suspension system dynamic model.

7. The bench test method for adaptive calibration of active suspension system model parameters according to claim 6, characterized in that, The actuator control force output from the time-delay calibration active suspension system dynamic model is input into the target active suspension bench test system. A standard test signal is applied to the target active suspension bench test system, and time-domain response data is collected. Tolerance determination is performed based on the time-domain response data to complete the adaptive calibration of the active suspension system dynamic model parameters, including: The actuator control force output from the dynamic model of the time-delay calibrated active suspension system is input to the target active suspension bench test system, a standard test signal is applied to the target active suspension bench test system, and the time-domain response data of the target active suspension bench test system is collected synchronously. The actuator control force output from the dynamic model of the time-delay calibrated active suspension system is input to the optimal time-delay feedback reference model, a standard test signal is applied to the optimal time-delay feedback reference model, and reference response data of the optimal time-delay feedback reference model is collected synchronously. The time-domain response data is compared point by point with the reference response data to calculate the tracking deviation sequence, and the average absolute value of the tracking error is calculated based on the tracking deviation sequence. Obtain a preset tolerance range, and use the tolerance range to determine the tolerance of the absolute value of the tracking error; If the average absolute value of the tracking error belongs to the tolerance range, then the time-delay calibration active suspension system dynamic model is determined to be the final active suspension system dynamic model, and the parameter adaptive calibration of the active suspension system dynamic model is completed. If the average absolute value of the tracking error does not fall within the tolerance range, the gradient descent method is used to incrementally adjust the sensitive system parameters of the time-delay calibration active suspension system dynamic model, and a standard test signal is applied to the adjusted time-delay calibration active suspension system dynamic model to complete the tracking error sequence update until the average absolute value of the tracking error of the updated tracking error sequence converges to the preset tolerance range.

8. A bench test system for adaptive calibration of active suspension system model parameters, characterized in that, The bench test method for adaptive calibration of model parameters of the active suspension system as described in any one of claims 1 to 7 includes: The sensitive parameter extraction unit is used to establish a dynamic model of the active suspension system including time delay based on kinematic theorems, construct a system parameter sensitivity analysis function for the active suspension system dynamic model, solve the sensitivity index of each system parameter from the system parameter sensitivity analysis function, and select the sensitive system parameters of the active suspension system dynamic model. The time-delay boundary calculation unit is used to apply simulated road surface excitation to the target active suspension bench test system in order to collect the time-domain response data of the target active suspension bench test system and calculate the time-delay stable working boundary of the target active suspension bench test system through the time-domain response data. The preliminary calibration unit is used to adjust the working conditions and input signals of the target active suspension bench test system according to the sensitive system parameters of the active suspension system dynamic model. It collects multiple sets of response data from the target active suspension bench test system, calculates the real-time estimated values ​​of the sensitive system parameters based on each set of response data, and updates the active suspension system dynamic model using the real-time estimated values ​​of the sensitive system parameters to obtain the preliminary calibrated active suspension system dynamic model. The optimal time-delay optimization unit is used to optimize the objective function in the initial calibration of the active suspension system dynamic model by taking the time-delay stable working boundary of the target active suspension bench test system as the boundary condition, maximizing the active suspension vibration reduction performance index as the objective function, and using the time-delay feedback coefficient and time delay as optimization variables. The time delay compensation unit is used to construct a Smith prediction compensation model based on the optimal time delay feedback coefficient and the optimal time delay amount. The Smith prediction compensation model is used to perform compensation prediction on the preliminary calibration active suspension system dynamic model to obtain the predicted compensation output. The optimal time delay feedback coefficient and the optimal time delay amount are used as compensation references. The predicted compensation output is compared with the actual system output of the target active suspension bench test system to obtain the comparison error. The time delay compensation amount is generated based on the comparison error. The time delay compensation amount is used to perform time delay compensation calibration on the preliminary calibration active suspension system dynamic model to obtain the time delay calibration active suspension system dynamic model. The model tolerance determination unit is used to input the actuator control force output by the time-delay calibration active suspension system dynamic model into the target active suspension bench test system, apply standard test signals to the target active suspension bench test system and collect time-domain response data, so as to determine the tolerance based on the time-domain response data, so as to complete the parameter adaptive calibration of the active suspension system dynamic model.

9. An electronic device, characterized in that, The system includes a memory, a processor, and a transceiver that are sequentially connected in communication. The memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the bench test method for adaptive calibration of active suspension system model parameters as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or the instructions are executed by the computer, they implement the bench test method for adaptive calibration of active suspension system model parameters as described in any one of claims 1 to 7.

Images