Knowledge and data fusion based multi-objective performance calculation method for vehicle suspension damping system

Abstract

The application discloses a knowledge and data fusion vehicle suspension damping system multi-target performance calculation method, S1, acquiring vehicle suspension damping system characteristic parameters; S2, establishing a key structure characteristic model according to the system characteristic parameters; S3, calculating a knowledge and data fusion vehicle suspension damping system multi-target performance simulation model integrated construction and multi-target performance simulation according to the key structure characteristic model; S4, establishing a rapid analysis system according to the knowledge and data fusion vehicle suspension damping system multi-target performance simulation model integrated construction and multi-target performance simulation; S5, testing and verifying a vehicle suspension damping system according to the rapid analysis system; the application considers the damping performance of the vehicle suspension damping system, and simultaneously considers the temperature rise performance and the anti-abnormal sound performance, so that the vehicle suspension damping system can meet various target performance requirements.

Description

Technical Field

[0001] This invention belongs to the field of automotive noise, vibration and acoustic comfort technology, specifically involving a multi-objective performance calculation method for vehicle suspension damping systems based on knowledge and data fusion. Background Technology

[0002] Shock absorbers, as a crucial component of automotive suspension, suppress the oscillation effect of springs after damping vibrations and absorb road impacts during vehicle operation. They play a vital role in the ride comfort and handling stability of a vehicle. After design, shock absorbers require tuning to achieve optimal performance matching the vehicle. Currently, manufacturers primarily rely on the experience of tuning personnel for shock absorber tuning. This results in low utilization of tuning data, a lack of tuning direction and data guidance, and numerous problems such as high tuning difficulty, long cycles, and high costs. Furthermore, it is difficult to achieve real-time performance control during the shock absorber production process.

[0003] Among the publicly available research on vibration damper calibration, Lai Fugang proposed using strain gauge sensors to measure the vibration damper's velocity under different working conditions. This allows for rapid damping force matching during vibration damper calibration, reducing the significant time required by traditional methods that rely on empirically predicting vibration damper velocity and then matching damping force. However, his research is limited to adjusting the damper's damping characteristics through experimental means to meet performance requirements (DOI:10.3969 / j.issn.1672-9668.2019.12.055). Wang Yanhui established a fluid-solid coupling model for the vibration damper based on its hydraulic, frictional, gas, and damping characteristics. He set the initial parameters of the AMEPilot model using disassembly test data from a test bench and the vibration damper, and combined this with Amesim-iSight full-factor experimental analysis to analyze the sensitivity of valve system parameters. He also studied the influence of displacement and velocity on the vibration damper's characteristics and the valve system's setting rules. Simultaneously, a thermo-liquid-solid coupled vibration damper model was established to analyze the internal heat conduction and convection mechanisms of the vibration damper, simulating related thermodynamic phenomena such as frictional heat generation, heat conduction, and heat convection within the system, and studying the influence of temperature on the vibration damper characteristics. While it comprehensively considers the speed and temperature characteristics of the vibration damper, the established thermo-liquid-solid coupled model is not only computationally cumbersome and inefficient, but also cannot evaluate the vibration damper's anti-noise performance. Since the vibration damper generates impact vibrations during the tension-reversal process, which in turn generates noise and affects the in-vehicle sound quality, the proportion of defective parts caused by this is relatively high and cannot be ignored. Du Xitao and Yuan Shimei's published invention (CN202110253063.8) proposes a vibration damper valve system adjustment method, introducing adjustment coefficients for the vibration damper valve system, namely the low-speed travel coefficient and the high-speed throttling coefficient. By introducing these adjustment coefficients, the abnormal noise factors during the vibration damper adjustment process are considered, improving the adjustment efficiency. However, the valve system structure needs to be adjusted based on the operator's subjective evaluation, and the adjustment threshold needs to be preset accordingly. This model can only evaluate the vibration damper's anti-noise performance and cannot consider the various effects of damping characteristics and temperature characteristics on the vibration damper's performance.

[0004] The difference between this patent and the aforementioned papers and patents lies in the establishment of a vehicle suspension damping system model that integrates knowledge and data. This model considers not only the most basic damping characteristics of the damping system but also its temperature characteristics and anti-noise characteristics. Furthermore, a rapid calibration system was built based on AppDesigner and the fused model to achieve co-simulation, facilitating rapid and efficient prediction of the vehicle suspension damping system performance. Summary of the Invention

[0005] The purpose of this invention is to address the problem that existing technologies offer relatively singular performance predictions for vehicle suspension damping systems, covering damping performance, temperature rise performance, and noise reduction performance. Furthermore, it proposes a multi-objective performance calculation method for vehicle suspension damping systems that integrates knowledge and data, thereby improving the calculation efficiency and accuracy of the prediction model.

[0006] The technical solution of this invention is as follows:

[0007] A multi-objective performance calculation method for vehicle suspension damping systems based on knowledge and data fusion includes the following steps:

[0008] S1. Obtain the characteristic parameters of the vehicle suspension damping system;

[0009] S2. Establish a key structural characteristic model based on the system characteristic parameters;

[0010] S3. Based on the key structural characteristic model, integrate and build a multi-objective performance simulation model of the vehicle suspension damping system by fusion of computational knowledge and data, and perform multi-objective performance simulation.

[0011] S4. Based on the integration of knowledge and data to build a multi-objective performance simulation model of the vehicle suspension damping system and perform multi-objective performance simulation, establish a rapid analysis system;

[0012] S5. Based on the rapid analysis system, multi-objective performance tests are conducted to verify the vehicle's suspension damping system.

[0013] Preferably, the characteristic parameters of the vehicle suspension damping system in step S1 include: the physicochemical properties of hydraulic oil, excitation characteristics, gas characteristics of the compensation chamber, environmental characteristics, and structural dimensions.

[0014] Preferably, step S2 includes the following sub-steps:

[0015] S21. Establish a knowledge model of the deflection characteristics of the throttle valve plate assembly;

[0016] S22. Establish a data-driven model for the throttling characteristics of the normally open orifice of the throttling valve plate;

[0017] S23. Establish an AMEsim parametric model of the vehicle suspension damping system.

[0018] Preferably, in step S21, the stiffness characteristics of the throttling valve plate assembly are characterized by establishing a knowledge model of the valve plate deformation process.

[0019] Preferably, in step S22, the throttling characteristics of the normally open orifice of the throttling valve plate are predicted by establishing a data-driven model.

[0020] Preferably, step S23 includes the following sub-steps:

[0021] Sub-step S231: The AMEsim parametric model is established using AMEsim parametric software;

[0022] Sub-step S232: Select a component model corresponding to the structure for characterization.

[0023] Preferably, step S3 includes the following sub-steps:

[0024] Sub-step S31: Integrate and build a vehicle suspension damping system model that combines knowledge and data;

[0025] Sub-step S32: Simulate and calculate the multi-objective performance of the vehicle suspension damping system based on the fusion model.

[0026] Preferably, the multi-objective performance in step S32 includes: damping performance, temperature rise performance, and noise resistance performance.

[0027] Preferably, in step S4, the rapid analysis system is established using AppDesigner and associated with the fusion model in step S3 to achieve rapid prediction of the multi-objective performance of the vehicle suspension damping system.

[0028] Preferably, in step S5, the fusion model is simulated using the rapid analysis system in step S4, and the damping performance, temperature rise performance, and anti-noise performance of the vehicle suspension damping system are tested and verified to verify the confidence level of the fusion model in step S3.

[0029] The beneficial effects of the knowledge and data fusion method for multi-objective performance calculation of vehicle suspension damping systems in this invention are as follows:

[0030] 1. This invention introduces mathematical models and knowledge-driven models to establish a knowledge- and data-integrated vehicle suspension damping system fusion model, which reasonably avoids the problems of insufficient equivalent accuracy of traditional parametric models and low efficiency of finite element simulation.

[0031] 2. Based on the damping performance of the vehicle suspension damping system, this invention also takes into account the temperature rise performance and the anti-noise performance, so that it can meet the requirements of multiple performances. Attached Figure Description

[0032] Figure 1 The flowchart of the multi-objective performance calculation method and analysis system for vehicle suspension damping system based on knowledge and data fusion provided by this invention is shown.

[0033] Figure 2 This is the structure of the throttle valve plate assembly of the present invention.

[0034] Figure 3 This is the AMEsim parametric element model of the throttle valve system of the present invention.

[0035] Figure 4 This invention provides a data-driven prediction model architecture for the throttling characteristics of the normally open orifice of the throttling valve plate.

[0036] Figure 5 This invention provides a knowledge and data fusion model for vehicle suspension damping systems.

[0037] Figure 6 This invention provides a rapid analysis system interface. Detailed Implementation

[0038] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0039] This invention provides a multi-objective performance calculation method for vehicle suspension damping systems that integrates knowledge and data, such as... Figure 1 As shown, it includes the following steps:

[0040] S1. Obtaining characteristic parameters of the vehicle suspension damping system;

[0041] The necessary characteristic parameters for building a multi-objective performance model based on the vehicle suspension and damping system include: excitation parameters, physicochemical properties parameters, environmental properties parameters, and structural dimensions (working cylinder parameters, compensation chamber parameters, piston connecting rod parameters, and valve system parameters).

[0042] S2. Establishment of key structural characteristic models;

[0043] The throttle valve system in a vehicle suspension damping system is generally a disc-shaped valve, and the uppermost valve typically has a non-structural normally open orifice. To effectively avoid the problem of low efficiency in calculating the deflection characteristics of the throttle valve assembly through finite element simulation, a knowledge model is introduced to describe the deflection characteristics of the throttle valve assembly. To effectively solve the problem of insufficient accuracy in throttling characteristics equivalent to non-structural normally open orifices using purely parametric element models, a data-driven model is introduced to describe the throttling characteristics of the normally open orifices of the throttle valve.

[0044] This step specifically includes the following sub-steps:

[0045] S21. Establishment of a knowledge model for the deflection characteristics of throttle valve plate assembly;

[0046] Disc valve structure as Figure 2 As shown, the stiffness element model in the AMEsim parametric element model is only applicable to the equivalent spring stiffness of a push valve, and its calculation kernel has a large error when calculating the deflection characteristics of a disc valve plate with deflection. This patent introduces a knowledge model, based on the theory of elasticity, to build a knowledge model to represent the deformation process of the valve plate. The original stiffness element model in AMEsim is as follows: Figure 3 As shown, a knowledge model is used to replace the stiffness element model in AMEsim.

[0047] The throttle valve assembly in a vehicle suspension damping system consists of a throttle valve, a washer valve, and a nut. The nut secures the washer valve to the throttle valve with a specific torque, and the throttle valve has a certain degree of deflection.

[0048] The structure establishes a polar coordinate system with the center of the valve plate as the pole. Since the structure and the load are both z-axis symmetric, the differential equations of the bending deformation surface of the elastic valve plate can be obtained according to the principles of elasticity, as shown in formulas (1) to (2):

[0049]

[0050]

[0051] In the formula: E is the elastic modulus of the valve plate; μ is Poisson's ratio; r a r is the inner radius of the valve plate. b r is the outer radius of the valve plate; r is the polar diameter, r∈[r a ,r b f is the bending deformation of the valve plate at the extreme diameter r; δ is the thickness of the valve plate. When multiple valve plates are stacked, δ needs to be calculated based on the equivalent thickness.

[0052] The general solution of equation (1) is expressed by formula (3):

[0053]

[0054] In the formula: C1, C2, C3 and C4 are undetermined coefficients, which are determined by the boundary conditions of the valve plate.

[0055] S22. Data-driven model establishment of throttling characteristics of the normally open orifice of the throttling valve plate;

[0056] To ensure hydraulic oil flow in the shock absorber at low speeds, the throttle valve plate has a non-structural normally open orifice. The normally open orifice element model in AMEsim can only calculate the throttling characteristics of orifices with regular cross-sectional shapes, and its calculation kernel has a large error when calculating the non-structural normally open orifice of the throttle valve plate in this patent. The finite element analysis method is usually used to calculate the flow-throttling pressure curve of the non-structural normally open orifice, but this equivalent method is inefficient. This patent introduces a data-driven model to predict the flow-throttling pressure curve of the non-structural normally open orifice of the throttle valve plate to characterize its throttling characteristics. The original normally open orifice throttling characteristic element model in AMEsim is as follows: Figure 2 As shown, a data-driven model is used to replace the normal through-hole throttling characteristic element model in AMEsim.

[0057] This patent introduces a data-driven model, constructing a multi-output random forest model to predict the flow-throttling pressure curve of a normally open orifice. Twenty cross-sections are taken at equal intervals along the length of the orifice using a plane, and the areas of the cross-sections are calculated as: A1, A2, A3…A 19 A20 In the multi-output random forest model, the length of the constant through-hole is L, and the cross-sectional area is A1~A. 20 The model is structured as follows: The input parameter is the flow rate-throttling pressure curve of the normally open orifice, and the output parameter is the flow rate-throttling pressure curve of the normally open orifice. Figure 3 As shown. Here, assume the dataset W has N samples, i.e., D = {(x...} (1) ,y (1) ),...,(x (N) ,y (N) The feature set has p dimensions, x (l) ={x1 (l) ,...,x j (l) ,...x p (l) The target set has q dimensions, y (l) =(y1) (l) ,...,y i (l) ,...,y q (l) The optimal splitting criterion, namely the sum of the squared errors of multiple variables, is selected to divide the dataset into two subsets. Assuming that the value s corresponding to the j-th feature vector is chosen as the splitting variable and splitting point, the training set is divided into two subsets as expressed by formulas (4) to (5):

[0058] R1(j,s)={x∣x(j)≤s} (4)

[0059] R2(j,s)={x∣x(j)>s} (5)

[0060] Formulas (6) to (8) must be satisfied:

[0061]

[0062]

[0063]

[0064] Iterate through the variable j to find the j and s that minimize the above expression, i.e., the optimal partitioning variable and the optimal partitioning point, and obtain the output value expressed by formula (9):

[0065]

[0066] Repeat the partitioning process until the stopping condition is met. The input set is divided into M regions, and the generated decision tree is represented by formula (10):

[0067]

[0068] S23. Establishment of AMEsim parametric component model for vehicle suspension damping system (excluding knowledge model and data-driven model in steps S21 and S22).

[0069] The vehicle suspension damping system model is built using AMEsin parametric elements (excluding the knowledge model and data-driven model in steps S21 and S22), which includes the excitation element model, cylinder element model, and compensation cavity element model.

[0070] S3. Integration and construction of vehicle suspension damping system model based on knowledge and data fusion, and multi-objective performance simulation calculation;

[0071] The models established in steps S21 to S23 are integrated to form a vehicle suspension damping system model that combines knowledge and data. Based on this, damping performance, temperature performance, and noise resistance performance are calculated.

[0072] This step specifically includes the following sub-steps:

[0073] S31. Integration and construction of a vehicle suspension damping system model that combines knowledge and data.

[0074] The knowledge-driven model of throttle valve plate deflection characteristics and the data-driven model of throttle valve plate normal through-hole throttle characteristics established in steps S21-S22 are integrated with the AMEsim parametric element model in step S23 to form a knowledge- and data-integrated vehicle suspension damping system model, such as... Figure 5 As shown.

[0075] S32. Multi-objective performance simulation calculation of vehicle suspension damping system based on fusion model.

[0076] Step S32 involves multi-objective performance simulation calculations of the vehicle suspension damping system, including: damping performance, temperature rise performance, and noise resistance performance. Damping performance refers to the damping force of the damping system at different operating speeds; temperature rise performance refers to the change in cylinder outer wall temperature over time during operation; and noise resistance performance refers to the vibration acceleration at the piston connecting rod tip under different operating conditions.

[0077] During the reciprocating motion of the vehicle suspension damping system in the recovery and compression processes, hydraulic oil flows through various throttling components to form damping force. The damping performance simulation analysis evaluates the damping force at different operating speeds. In formulas (11) to (12), F fy and F ys The damping forces represent the recovery stroke and compression stroke, respectively; the anti-noise performance simulation analysis is evaluated by the vibration acceleration at the end of the connecting rod, where a in formulas (11) to (12) represents the damping force. fy and a ys These represent the vibration acceleration at the end of the connecting rod during the recovery stroke and compression stroke, respectively.

[0078] m d a fy =F f +F hydrd +F air -F fy (11)

[0079] m d a ys =F f -F hydrd -F air -F ys (12)

[0080] Where: m d x is the mass of the piston connecting rod; d For piston rod displacement; F f F is the frictional force acting on the piston connecting rod. hydrv F is the hydraulic pressure exerted on the piston connecting rod. hydrd =p rec A rec -p compr A compr p rec To ensure uniform pressure distribution of the oil in the shock absorber's recovery chamber, A rec p is the area of ​​the piston upper surface where the oil acts. compr A is the pressure of the oil in the compression chamber of the shock absorber that is evenly distributed. compr F represents the area of ​​the piston's lower surface where the oil acts; air F is the atmospheric force acting on the connecting rod. air =p air A rod p air For standard atmospheric pressure, A rod Let be the cross-sectional area of ​​the connecting rod.

[0081] For the simulation calculation of the temperature rise performance of the vehicle suspension damping system, based on the analysis of damping performance, ignoring other forms of energy dissipation, the work done by the damping force will eventually be converted into heat energy, part of which causes the temperature of the damper system to rise, and the other part is lost through heat exchange with the outside, as expressed by formula (13):

[0082] W=c s m s ΔT s +c o m o ΔT o +Q (13)

[0083] In the formula: c s m s c o m oT represents the specific heat capacity and mass of the internal structure and the specific heat capacity and mass of the oil, respectively; s and T o These are the average temperature of the internal structure and the average temperature of the oil, respectively.

[0084] The heat transfer inside the damper is equivalent to one-dimensional radial heat transfer, and the heat transfer modes include conduction, convection and thermal radiation. The heat transfer between the oil in the working cylinder of the damping system and the inner wall of the working cylinder is forced convection heat transfer, and the heat flow rate is calculated using formula (14).

[0085]

[0086] In the formula: l is the length of the circular tube; r1 and r2 represent the inner and outer diameters of the circular tube, respectively; λ is the thermal conductivity of the material; and T1 and T2 represent the temperatures of the inner and outer walls, respectively.

[0087] The thermal convection process includes convection between the oil in the working cylinder and the inner wall of the working cylinder, convection between the oil in the storage cylinder and the outer wall of the working cylinder, convection between the oil in the storage cylinder and the inner wall of the storage cylinder, and convection between the outer wall of the storage cylinder and the environment. Among them, the convection between the inner wall of the working cylinder and the inner wall of the storage cylinder and the oil is a forced convection within the pipe, and its Nusselt number is calculated using formulas (15) to (16).

[0088]

[0089]

[0090] Where: N ul N represents the laminar Nusselt number; ut For denoted as Nusselt number; Re is the Reynolds number; P is the flow-Nucler number. r d is the Prandtl number; d is the pipe diameter; l is the length of the heat exchange zone pipe; μ is the absolute viscosity of the oil at the fluid temperature; μ s This represents the absolute viscosity of the oil at the pipe wall temperature.

[0091] The convection between the oil in the reservoir and the outer wall of the working cylinder, as well as the convection between the outer wall of the reservoir and the environment, all belong to forced convection outside the pipe. The Nusselt number is calculated using formula (17).

[0092] Nu = c(Re) n Pr 0.33 (17)

[0093] In the formula: the coefficient c is a function of n and Re.

[0094] After calculating the Nusselt number, the convective heat transfer coefficient is expressed by formula (18):

[0095]

[0096] In the formula: λ is the thermal conductivity of the fluid; cd The characteristic dimension is the dimension that reflects the geometric characteristics of the flow field. For fluid inside a pipe, the characteristic dimension refers to the inner diameter of the pipe perpendicular to the flow direction; when the fluid flows around a cylinder, the characteristic dimension refers to the outer diameter of the pipe.

[0097] The radiative heat exchange between the outer wall of the oil reservoir of the vibration damping system and the environment per unit time is expressed by formula (19):

[0098] q=εSσ(T w -T g (19)

[0099] In the formula: ε is the radiation factor, which is related to the coating material of the outer wall; S is the outer wall area of ​​the oil reservoir; σ is the Boltzmann constant; T w and T g These are the temperatures of the outer wall of the oil reservoir and the ambient temperature, respectively.

[0100] S4. Rapid analysis system established;

[0101] A rapid analysis system interface was built using the programming software AppDesigner. Figure 6 As shown. Input parameters include the hydraulic oil physicochemical properties, excitation properties, compensation chamber gas properties, environmental properties, and structural dimensions of the vehicle suspension damping system in step S1. A callback function is used to associate the parameter assignment module in the rapid analysis system with the fusion model, allowing the rapid analysis system to modify the fusion model parameters. A callback function is also used to associate the function plotting module in the rapid analysis system with the fusion model to receive the performance data generated by the fusion model. The damping performance curve, temperature rise performance curve, and vibration acceleration curve are plotted in the rapid analysis system, achieving efficient and accurate calculation of the vehicle suspension damping system performance simulation through knowledge data fusion.

[0102] S5. Multi-objective performance test verification of vehicle suspension damping system.

[0103] The vehicle suspension damping system was installed on the MTS test bench. Damping forces were collected at different operating speeds to verify damping characteristics; outer cylinder temperatures were collected at different operating speeds and ambient temperatures to verify temperature rise performance; and connecting rod end vibration accelerations were collected at different operating speeds to verify anti-noise performance. The confidence level of the vehicle suspension damping system model established in step S3, which integrates knowledge and data, was verified using the above evaluation indicators.

[0104] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A multi-objective performance calculation method for vehicle suspension damping systems based on knowledge and data fusion, characterized in that, Includes the following steps: S1. Obtain the characteristic parameters of the vehicle suspension damping system; S2. Establish a key structural characteristic model based on the system characteristic parameters; S3. Based on the key structural characteristic model, calculate and build a multi-objective performance simulation model of the vehicle suspension damping system by integrating knowledge and data, and perform multi-objective performance simulation. S4. Based on the integration of knowledge and data to build a multi-objective performance simulation model of the vehicle suspension damping system and perform multi-objective performance simulation, establish a rapid analysis system; S5. Based on the rapid analysis system, multi-objective performance tests are conducted to verify the vehicle's suspension and vibration damping system. The characteristic parameters of the vehicle suspension damping system in step S1 include: hydraulic oil physicochemical properties, excitation characteristics, compensation chamber gas characteristics, environmental characteristics, and structural dimensions; Step S2 includes the following sub-steps: S21, establishing a knowledge model of the deflection characteristics of the throttle valve plate assembly; S22, establishing a data-driven model of the throttle characteristics of the normally open orifice of the throttle valve plate; S23, establishing an AMEsim parametric model of the vehicle suspension damping system. In step S21, the stiffness characteristics of the throttling valve plate assembly are characterized by establishing a knowledge model of the valve plate deformation process; the knowledge model is used to replace the stiffness element model in AMEsim. In step S22, the throttling characteristics of the throttling valve plate through the normally open orifice are predicted by establishing a data-driven model; the data-driven model is used to replace the normally open orifice throttling characteristic element model in AMEsim. Step S23 includes the following sub-steps: Sub-step S231: The AMEsim parametric model is established using AMEsim parametric software; Sub-step S232: Component models corresponding to the structure are selected for characterization; Step S3 includes the following sub-steps: Sub-step S31: Integrating and building a vehicle suspension damping system model that fuses knowledge and data, by fusing the knowledge-driven model of throttle valve plate deflection characteristics and the data-driven model of throttle valve plate normal through-hole throttling characteristics established in steps S21-S22 with the AMEsim parametric element model in step S23 to form a vehicle suspension damping system model that fuses knowledge and data; Sub-step S32: Simulating and calculating the multi-objective performance of the vehicle suspension damping system based on the fused model; The multi-objective performance in step S32 includes: damping performance, temperature rise performance, and noise resistance performance; In step S4, the rapid analysis system is established using Appdesigner and associated with the fusion model in step S3 to achieve rapid prediction of the multi-objective performance of the vehicle suspension damping system. In step S5, the fusion model is simulated using the rapid analysis system in step S4. The damping performance, temperature rise performance, and anti-noise performance of the vehicle suspension damping system are tested and verified to verify the confidence level of the fusion model in step S3.

Images