A Method for Establishing a Dynamic Inertia Suspension Model of a Wheel-Hub-Driven Vehicle
Through the hub-driven vehicle dynamic inertial suspension model optimized by inertial container and floor shed control combined with genetic algorithm, the problems of low-frequency vibration and tire dynamic load of the hub-driven vehicle are solved, and performance improvement in the wide frequency range and maximum utilization of inertial container performance are achieved.
Patent Information
- Application Number
- CN202210573453.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-05-25
- Publication Date
- 2025-07-04
- Estimated Expiration
- 2042-05-25
AI Technical Summary
The prior art has failed to effectively suppress the low-frequency vibration and tire dynamic load of the hub-driven automobile, and has not fully utilized the dynamic performance of the inertial container. Especially in hub-driven automobiles, traditional methods have failed to effectively solve the dynamic boundary problem under the influence of the motor.
The inertial container element and switching reluctance motor are combined with ground shed control, and the impedance transfer function of the second-order ground shed network is optimized through the genetic algorithm, and the dynamic inertial suspension model of the hub-driven vehicle is established. The inertial container is used to suppress low-frequency vibration, and the ground shed control is used to suppress high-frequency vibration, and the suspension model parameters are determined in combination with the genetic algorithm.
It significantly improves the tire dynamic load of the hub-driven car in the wide frequency range, improves the grounding performance of the car, reduces the influence of the vertical electromagnetic excitation force of the motor, maximizes the performance advantages of the inertia container, and optimizes the suspension structure.
Smart Images

Figure CN114896702B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of modeling of the dynamic inertia suspension system of a wheel-driven vehicle, and particularly relates to a method for establishing a dynamic inertia suspension model of a wheel-driven vehicle. Background Art
[0002] Professor Smith of the University of Cambridge in the UK first proposed the concept of an inertor in 2002. As a two-terminal mass element, its appearance makes up for the lack of mass impedance in the suspension system. Subsequently, the second type of electromechanical analogy theory emerged, thus realizing the strict correspondence between the electrical network and the mechanical network, and promoting the development of the network synthesis theory of the suspension system. The dynamic inertia suspension network structure can significantly suppress the low-frequency vibration of the vehicle body and improve the vehicle performance.
[0003] In the second type of electromechanical analogy theory, the inertor corresponds to the capacitor, the damper corresponds to the resistor, and the spring corresponds to the inductor. Compared with the traditional passive suspension, the emergence of the inertor enables the application of the mature electrical network structure design knowledge to the mechanical network, resulting in a series of dynamic inertia suspensions with excellent performance, which greatly promotes the improvement of the suspension performance.
[0004] Chinese Patent CN110001337B discloses a method for establishing a second-order ideal model of a vehicle dynamic inertia suspension based on the optimization of the ADD positive real network. However, it only conducts structural design for ordinary traditional vehicles, does not study the wheel-driven vehicle affected by the coupled motor, and does not determine the dynamic boundary under the interaction of ADD control and the positive real network. The dynamic performance of the inertor has not been fully exerted. Summary of the Invention
[0005] For the above reasons, the present invention provides a method for establishing a dynamic inertia suspension model of a wheel-driven vehicle. By using the characteristics of the positive real network of the inertor element to effectively suppress low-frequency vibration and the characteristics of the skyhook control to effectively suppress high-frequency vibration of the tire dynamic load, the grounding performance of the wheel-driven vehicle is improved. Since there are many second-order network elements and it is not easy to arrange them in engineering, it can be used as a reference model and realized by simulating with an electromechanical inertor.
[0006] To achieve the above invention purposes, the technical solution adopted by the present invention is: A method for establishing a dynamic inertia suspension model of a wheel-driven vehicle, comprising:
[0007] Step (1): Establish a quarter model of the dynamic inertia suspension of the wheel-driven vehicle:
[0008]
[0009] Wherein, m s is the sprung mass, m u + m e is the unsprung mass, m uis the wheel mass, m e is the motor mass, k is the suspension spring stiffness, c g is the semi-active damping coefficient controlled by the skyhook, k t is the equivalent stiffness of the tire, z s is the vertical displacement of the sprung mass, is the vertical velocity of the sprung mass, is the vertical acceleration of the sprung mass, z u is the vertical displacement of the unsprung mass, is the vertical velocity of the unsprung mass, is the vertical acceleration of the unsprung mass, z r is the vertical input displacement of the road surface unevenness, T(s) is the velocity-type impedance transfer function of the second-order skyhook positive real network, F r_z is the vertical electromagnetic excitation force;
[0010] Performing Laplace transform on it gives:
[0011]
[0012] where, Z s is the Laplace transform form of the vertical displacement of the sprung mass, Z u is the Laplace transform form of the vertical displacement of the unsprung mass, Z r is the Laplace transform form of the vertical input displacement of the road surface unevenness, s is the Laplace variable;
[0013] Step (2): Establish the vertical electromagnetic excitation force model of the switched reluctance motor:
[0014]
[0015] where, is the vertical component of the unbalanced radial electromagnetic force, is the unbalanced radial electromagnetic force, i is the current, L is the winding inductance, k m is the phase of the motor, is the motor k m phase current, θ is the relative angular displacement between the stator and rotor of the motor, g m is the motor air gap when the motor is not eccentric, ε is the relative eccentricity of the motor air gap; because when the motor is eccentric, the motor air gap will also change. When the number of poles of the stator in the opposite direction is m, the motor air gap is g m (n), the motor k m phase winding inductance is When the number of poles of the stator in the opposite direction is n, the motor air gap is g m (m), the motor k m phase winding inductance is
[0016] Step (3): Determine the velocity-type impedance transfer function of the second-order floor positive-real network:
[0017]
[0018] where A, B, C, D, E, and F are coefficients, all of which are greater than or equal to 0, and D, E, and F are not all 0, and satisfy the positive-real constraint conditions;
[0019] Step (4): Determine the floor control algorithm: When the vertical velocity of the unsprung mass and the relative vertical velocity of the sprung mass and the unsprung mass are in the same direction, the damping coefficient of the control input is c max ; conversely, when the vertical velocity of the unsprung mass is in the opposite direction to the relative vertical velocity of the sprung mass and the unsprung mass, the damping coefficient of the control input is c min ; the floor damping coefficient c g needs to satisfy the following formula:
[0020]
[0021] where c max , c min are two damping coefficient values and satisfy the following formula:
[0022] c max > c min > 0;
[0023] Step (5): Select the vertical input displacement z of the road surface unevenness r ;
[0024] Step (6): Use the genetic algorithm to determine the response index constraint conditions of the second-order floor positive-real network:
[0025] BA(z) < 1.08×BA pass , SWS(z) < SWS pass , DTL(z) < DTL pass ,
[0026] where BA(z) is the root mean square value of the body acceleration of the second-order floor positive-real network, SWS(z) is the root mean square value of the suspension dynamic stroke of the second-order floor positive-real network, DTL(z) is the root mean square value of the tire dynamic load of the second-order floor positive-real network, BA pass is the root mean square value of the body acceleration of the traditional passive suspension, SWS pass is the root mean square value of the suspension dynamic stroke of the traditional passive suspension, DTL pass is the root mean square value of the tire dynamic load of the traditional passive suspension;
[0027] Step (7): Use the genetic algorithm to obtain the dynamic inertia suspension model parameters A, B, C, D, E, F, cmax , c min ;
[0028] Step (8): Substitute the suspension model parameters obtained in step (7) into the velocity-type impedance transfer function T(s) of the second-order skyhook positive real network and the skyhook damping to obtain the dynamic inertia suspension model of the wheel-driven vehicle.
[0029] Further, in step (2), g m (n) = (1 - ε)g m , g m (m) = (1 + ε)g m , |m - n| = 4; ε is:
[0030]
[0031] where Δg is the absolute eccentricity of the motor rotor in the radial direction.
[0032] Further, in step (2), when the number of poles of the stator in the opposite direction is m, the motor air gap is g m (n), and the winding inductance of the k m phase of the motor is:
[0033] where N r is the number of poles of the motor rotor, a n , b n are the polynomial fitting coefficients, L u is the inductance at the pole-slot alignment position, and L n is the Fourier coefficient of the winding inductance of each phase;
[0034] When the number of poles of the stator in the opposite direction is n, the motor air gap is g m (m), and the winding inductance of the k m phase of the motor is:
[0035]
[0036] Further, in step (3), the positive real constraint on the velocity-type impedance transfer function T(s) of the second-order skyhook positive real network is:
[0037]
[0038] Further, step 5) is specifically:
[0039]
[0040] where v represents the driving speed, n0 represents the reference spatial frequency, Gq (n0) represents the road surface unevenness coefficient, w(t) represents Gaussian white noise with a mean of 0, and z r is the vertical input displacement of the road surface unevenness, and is the vertical input velocity of the road surface unevenness.
[0041] Furthermore, among them, the step (7) is specifically as follows: The genetic algorithm mainly includes generating an initial population, calculating fitness values, selection, crossover, mutation, and iteration, specifically as follows:
[0042] Step (7.1): Generate an initial population as a randomly generated population size;
[0043] Step (7.2): Taking the tire dynamic load as the optimization objective, the calculation formula for the single-objective optimization fitness function of the genetic algorithm is:
[0044]
[0045] where Punishment is the penalty number;
[0046] Step (7.3): Selection means directly inheriting the optimized individuals to the next generation or generating new individuals through crossover and then inheriting them to the next generation;
[0047] Step (7.4): Crossover means replacing and recombining some genes of two individuals in the population to generate new individuals;
[0048] Step (7.5): Mutation means changing the gene values at certain gene loci of the individuals in the population;
[0049] Step (7.6): Iteration means calculating whether the termination condition is satisfied. If not, continue the above operations until the maximum number of iterations is reached; if satisfied, output the optimal solution.
[0050] Furthermore, in step (7.2), the value-taking rule of the penalty number Punishment is: As long as the root mean square value of the body acceleration BA(z) of the second-order floor positive real network, the root mean square value of the suspension dynamic stroke SWS(z) of the second-order floor positive real network, and the root mean square value of the tire dynamic load DTL(z) of the second-order floor positive real network, as long as one of them is greater than 1.08 times the root mean square value of the body acceleration of the traditional passive suspension 1.08×BA pass , the root mean square value of the suspension dynamic stroke SWS of the traditional passive suspension pass , the root mean square value of the tire dynamic load DTL of the traditional passive suspension pass , then the penalty number Punishment takes the value of 1000, otherwise it takes the value of 0.
[0051] The beneficial implementation effects of adopting the present invention are as follows: Based on the ground shed control, the present invention has a good inhibitory effect on the medium and high frequencies of the tire dynamic load. Combining the characteristics of the positive real network system of the dynamic inertia suspension, which can significantly improve the low-frequency vibration of the vehicle, a new type of efficient dynamic inertia suspension is designed to suppress the tire dynamic load of the hub-driven vehicle in a wide frequency range, improve the vehicle's grounding performance, and effectively reduce the influence of the vertical electromagnetic excitation force of the motor during the operation of the hub-driven vehicle. Moreover, through the genetic algorithm, the dynamic boundary under the interaction of the positive real network of the dynamic inertia suspension and the ground shed control is obtained, and the optimization constraint conditions are determined. While ensuring the vehicle performance, the performance advantages of the inertance are maximally exerted, the optimal parameters of the dynamic inertia suspension structure are obtained, and using the network synthesis theory, the specific structure of the dynamic inertia suspension with excellent performance in the frequency domain and time domain is obtained. Brief Description of the Drawings
[0052] The present invention will be further described below in conjunction with the drawings and embodiments.
[0053] Figure 1 It is a design flow chart of the second-order ground shed positive real network of the dynamic inertia suspension for the hub-driven vehicle.
[0054] Figure 2 It is a schematic diagram of a quarter model of the dynamic inertia suspension for the hub-driven vehicle.
[0055] Figure 3 It is a schematic diagram of a 4-phase 8 / 6-pole outer rotor switched reluctance motor model.
[0056] Figure 4 It is a flow chart of the genetic algorithm.
[0057] Figure 5 It is a graph showing the variation of the root mean square value of the tire dynamic load of the second-order ground shed positive real network with the root mean square value of the vehicle body acceleration.
[0058] Figure 6 It is a schematic diagram of the structure of the velocity-type impedance transfer function T(s) of the second-order ground shed positive real network.
[0059] Figure 7 It is a performance index graph of the second-order ground shed positive real network of the dynamic inertia suspension for the hub-driven vehicle, where (a) is the vehicle body acceleration response graph, (b) is the suspension dynamic stroke response graph, and (c) is the tire dynamic load response graph. Detailed Embodiment
[0060] The present invention will be further described below in conjunction with the drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.
[0061] As Figure 1As shown in the figure, a method for establishing a dynamic inertia suspension model of a wheel hub drive vehicle according to the present invention includes the following steps: Step (1): Establish a quarter model of the dynamic inertia suspension of the wheel hub drive vehicle; Step (2): Establish a vertical electromagnetic excitation force model of the switched reluctance motor; Step (3): Determine the velocity-type impedance transfer function of the second-order terrain positive real network; Step (4): Determine the terrain control algorithm; Step (5): Select the vertical input displacement of the road surface unevenness; Step (6): Use the genetic algorithm to determine the response index constraint conditions of the second-order terrain positive real network; Step (7): Use the genetic algorithm to obtain the suspension model parameters; Step (8): Substitute the suspension model parameters obtained in Step (7) into the velocity-type impedance transfer function T(s) of the second-order terrain positive real network and the terrain damping to obtain the dynamic inertia suspension model of the wheel hub drive vehicle. After the model is established, it is verified through the following steps: Simulate the dynamic inertia suspension model of the wheel hub drive vehicle based on the second-order terrain positive real network.
[0062] Among them, Step (1) is specifically as follows: As Figure 2 shown, establish a quarter model of the dynamic inertia suspension of the wheel hub drive vehicle:
[0063]
[0064] Among them, m s is the sprung mass, m u +m e is the unsprung mass, m u is the wheel mass, m e is the motor mass, k is the suspension spring stiffness, c g is the semi-active damping coefficient of the terrain control, k t is the equivalent stiffness of the tire, z s is the vertical displacement of the sprung mass, is the vertical velocity of the sprung mass, is the vertical acceleration of the sprung mass, z u is the vertical displacement of the unsprung mass, is the vertical velocity of the unsprung mass, is the vertical acceleration of the unsprung mass, z r is the vertical input displacement of the road surface unevenness, T(s) is the velocity-type impedance transfer function of the second-order terrain positive real network, F r_z is the vertical electromagnetic excitation force.
[0065] Furthermore, (z s -z u ), k t (z u -z rThey are respectively the vertical acceleration of the sprung mass, the dynamic suspension travel, and the dynamic tire load, which are three dynamic performance indicators of the suspension system.
[0066] Furthermore, the Laplace transform of the above quarter-car model of the dynamic inertia suspension is obtained as follows:
[0067]
[0068] where Z s is the Laplace transform form of the vertical displacement of the sprung mass, Z u is the Laplace transform form of the vertical displacement of the unsprung mass, Z r is the Laplace transform form of the vertical input displacement of the road surface unevenness, and s is the Laplace variable.
[0069] Among them, step (2) is specifically as follows: Figure 3 is a schematic diagram of a 4-phase 8 / 6-pole outer rotor switched reluctance motor model, which consists of a rotor, a stator, windings, and a motor shaft. The rotor has 6 poles, the stator has 8 poles, and every two opposite stator poles form a phase. Establish its vertical electromagnetic excitation force model:
[0070]
[0071] where is the vertical component of the unbalanced radial electromagnetic force, is the unbalanced radial electromagnetic force, i is the current, L is the winding inductance, k m is the phase of the motor, is the current of the k m phase of the motor, θ is the relative angular displacement between the stator and rotor of the motor, g m is the motor air gap when the motor is not eccentric, and ε is the relative eccentricity of the motor air gap; because when the motor is eccentric, the motor air gap will also change. When the number of poles of the stator in the opposite direction is m, the motor air gap is g m (n), and the winding inductance of the k m phase of the motor is When the number of poles of the stator in the opposite direction is n, the motor air gap is g m (m), and the winding inductance of the k m phase of the motor is g m (n) = (1 - ε)g m g m (m) = (1 + ε)g m , |m - n| = 4.
[0072] Furthermore, ε is:
[0073]
[0074] Among them, △g is the absolute eccentricity of the motor rotor in the radial direction.
[0075] Further, when the number of poles of the stator in the opposite direction is m, the motor air gap is g m (n), the motor k m The winding inductance of the phase is:
[0076]
[0077] Among them, N r is the number of poles of the motor rotor, a n 、b n are polynomial fitting coefficients, L u is the inductance at the position where the pole is aligned with the slot, L n is the Fourier coefficient of the winding inductance of each phase.
[0078] Further, when the number of poles of the stator in the opposite direction is n, the motor air gap is g m (m), the motor k m The winding inductance of the phase is:
[0079]
[0080] Among them, step (3) is specifically: The expression of the velocity-type impedance transfer function T(s) of the second-order floor positive-real network is as follows:
[0081]
[0082] Among them, A, B, C, D, E, F are coefficients, and their values are all greater than or equal to 0, and D, E, F are not all 0, and they satisfy the positive-real constraint conditions;
[0083] Further, according to the passive network synthesis theory, the velocity-type impedance transfer function T(s) of the second-order floor positive-real network is realized by using inertance, spring, and damping in series and parallel.
[0084] Further, the positive-real constraint on the velocity-type impedance transfer function T(s) of the second-order floor positive-real network is:
[0085]
[0086] Among them, step (4) is specifically: The floor control algorithm is: When the vertical velocity of the unsprung mass and the relative vertical velocity of the sprung mass and the unsprung mass are in the same direction, the damping coefficient of the control input is c max ; conversely, when the vertical velocity of the unsprung mass is in the opposite direction to the relative vertical velocity of the sprung mass and the unsprung mass, the damping coefficient of the control input is c min ; the floor damping coefficient c g needs to satisfy the following formula:
[0087]
[0088] Among them, c max and c min are two damping coefficient values and satisfy the following formula:
[0089] c max > c min > 0.
[0090] Among them, step (5) is specifically: select the vertical input displacement z r of the road surface unevenness, specifically:
[0091]
[0092] Among them, v represents the driving speed, taken as 20 m / s, n0 represents the reference spatial frequency, taken as 0.1 m -1 , G q (n0) represents the road surface unevenness coefficient, w(t) represents Gaussian white noise with a mean of 0, z r is the vertical input displacement of the road surface unevenness, and is the vertical input speed of the road surface unevenness.
[0093] Furthermore, select a Class C road surface as the random road surface input, and the road surface unevenness coefficient G q (n0) takes the geometric mean value of 2.56×10 -4 m 3 .
[0094] Table 1 Quarter suspension model parameters
[0095]
[0096] Among them, m u = 45 kg, m e = 30 kg.
[0097] Among them, step (6) is specifically: because the floor control will deteriorate the vehicle body acceleration, study the optimization effect of the second-order floor positive real network on the root mean square value of the tire dynamic load under the change of the root mean square value constraint condition of the vehicle body acceleration through the genetic algorithm. As Figure 5As shown, the vehicle body acceleration increases by 2% each time from 1.4129 to 1.6955. When the vehicle body acceleration constraint is 1.4129, the vehicle body acceleration does not deteriorate, and the tire dynamic load is optimized by 10.05%; when the vehicle body acceleration constraint is 1.5259, the vehicle body acceleration deteriorates by 8%, and the tire dynamic load is optimized by 19.40%; when the vehicle body acceleration constraint is 1.6955, the vehicle body acceleration deteriorates by 20%, and the tire dynamic load is optimized by 23.99%. And taking 1.5259 as the demarcation point, the optimization effect of the tire dynamic load is less than 2% each time afterwards, and it is no longer appropriate to sacrifice the vehicle body acceleration for the optimization of the tire dynamic load. Therefore, when the vehicle body acceleration constraint is 1.5259, a larger optimization effect of the tire dynamic load is obtained by sacrificing a smaller vehicle body acceleration deterioration. So the response index constraint conditions of the second-order ground-hitch positive-real network are determined as:
[0098] BA(z) < 1.08×BA pass , SWS(z) < SWS pass , DTL(z) < DTL pass
[0099] Among them, BA(z) is the root mean square value of the vehicle body acceleration of the second-order ground-hitch positive-real network, SWS(z) is the root mean square value of the suspension dynamic stroke of the second-order ground-hitch positive-real network, DTL(z) is the root mean square value of the tire dynamic load of the second-order ground-hitch positive-real network, BA pass is the root mean square value of the vehicle body acceleration of the traditional passive suspension, SWS pass is the root mean square value of the suspension dynamic stroke of the traditional passive suspension, DTL pass is the root mean square value of the tire dynamic load of the traditional passive suspension.
[0100] Among them, step (7) is specifically: using the genetic algorithm to obtain the suspension model parameters. The genetic algorithm mainly includes generating the initial population, calculating the fitness value, selection, crossover, mutation, and iteration. The process is as Figure 4 shown.
[0101] Furthermore, before the genetic algorithm is optimized, it is first necessary to determine the parameters X to be optimized of the suspension model, X = (A, B, C, D, E, F, c max , c min ).
[0102] Furthermore, the parameters to be optimized of the suspension model need to be optimized within a certain range. Therefore, its constraint range is set as:
[0103] LB = (10, 10, 10, 10, 10, 10, 10, 10)
[0104] UB = (100000, 5000000, 5000000, 1000, 10000, 100000, 10000, 10000)
[0105] LB ≤ X ≤ UB
[0106] Where LB is the lower limit of the parameter to be optimized, and UB is the upper limit of the parameter to be optimized.
[0107] Furthermore, the initial population is randomly generated with a population size.
[0108] Furthermore, with the tire dynamic load as the optimization objective, the calculation formula for the single-objective optimization fitness function of the genetic algorithm is designed as:
[0109]
[0110] The value-taking rule of the penalty number Punishment in the above calculation formula for the single-objective optimization fitness function of the genetic algorithm is as follows: As long as the root mean square value of the body acceleration BA(z) of the second-order floor positive-real network, the root mean square value of the suspension dynamic stroke SWS(z) of the second-order floor positive-real network, or the root mean square value of the tire dynamic load DTL(z) of the second-order floor positive-real network is greater than 1.08 times the root mean square value of the body acceleration 1.08×BA of the traditional passive suspension pass of the traditional passive suspension, the root mean square value of the suspension dynamic stroke SWS pass of the traditional passive suspension, and the root mean square value of the tire dynamic load DTL pass of the traditional passive suspension, then the penalty number Punishment takes the value of 1000; otherwise, it takes the value of 0.
[0111] Furthermore, the selection is to directly inherit the individuals to be optimized to the next generation or generate new individuals through crossover and then inherit them to the next generation.
[0112] Furthermore, the crossover is to replace and recombine some genes of two individuals in the population to generate new individuals.
[0113] Furthermore, the mutation is to change the gene values at some gene loci of the individuals in the population.
[0114] Furthermore, the iteration is to calculate whether the termination condition is satisfied. If not, continue the above operations until the maximum number of iterations is reached; if satisfied, output the optimal solution.
[0115] Furthermore, in the genetic algorithm, the initial population is set to 100, the crossover probability is 0.8, the mutation probability is 0.05, and the number of iterations is 500.
[0116] Furthermore, the optimization result is to obtain the optimal values of the 8 suspension model parameters after optimization.
[0117] Furthermore, Table 2 shows the results of the parameter optimization:
[0118] Table 2 Results of Parameter Optimization
[0119]
[0120] Furthermore, according to the optimization results, the expression of the velocity-type impedance transfer function T(s) of the second-order floor positive-real network is obtained as follows:
[0121] Furthermore, the structure of the obtained velocity-type impedance transfer function T(s) of the second-order floor positive-real network is verified. Using the passive network synthesis theory, such as Figure 6 The schematic diagram of the five-element structure of the velocity-type impedance transfer function T(s) of the second-order floor positive-real network is shown as follows, and Table 3 shows the parameters of the corresponding components.
[0122] Table 3 Parameters of the corresponding components
[0123]
[0124] The following is the simulation verification:
[0125] A quarter model of the traditional passive suspension and the dynamic inertia suspension of the second-order floor positive-real network is built through MATLAB / Simulink respectively, and the tire dynamic load performance index is used as the evaluation index. Since it is impossible to directly analyze and solve through the frequency response function, a sine excitation is used as the frequency-domain road surface input:
[0126] z r = Asin(2πft)
[0127] where A represents the excitation amplitude, with a value of 0.01 m, and f represents the excitation frequency, with a value range of 0.01 - 15 Hz.
[0128] As Figure 7 shown, where (a) is the vehicle body acceleration response diagram, (b) is the suspension dynamic stroke response diagram, and (c) is the tire dynamic load response diagram.
[0129] Table 4 Comparison table of suspension root mean square values
[0130]
[0131] The above results show that the tire dynamic load and the suspension dynamic stroke of the dynamic inertia suspension model of the second-order floor positive-real network of the present invention have been significantly improved, effectively enhancing the grounding performance and handling stability of the wheel-driven vehicle.
[0132] The described embodiments are the preferred embodiments of the present invention, but the present invention is not limited to this embodiment. Without departing from the essence of the present invention, the modifications, deformations, and substitutions made by those skilled in the art all fall within the protection scope of the present invention.
Claims
1. A method for establishing a dynamic inertia suspension model of a wheel hub drive vehicle, characterized in that Including: Step (1): Establish a quarter model of the dynamic inertia suspension of a wheel hub drive vehicle: Among them, m s is the sprung mass, m u + m e is the unsprung mass, m u is the wheel mass, m e is the motor mass, k is the suspension spring stiffness, c g is the semi-active damping coefficient of the skyhook control, k t is the equivalent stiffness of the tire, z s is the vertical displacement of the sprung mass, is the vertical velocity of the sprung mass, is the vertical acceleration of the sprung mass, z u is the vertical displacement of the unsprung mass, is the vertical velocity of the unsprung mass, is the vertical acceleration of the unsprung mass, z r is the vertical input displacement of the road surface unevenness, T(s) is the velocity-type impedance transfer function of the second-order skyhook positive real network, F r_z is the vertical electromagnetic excitation force; Performing a Laplace transform on it to obtain: Among them, Z s is the Laplace transform form of the vertical displacement of the sprung mass, Z u is the Laplace transform form of the vertical displacement of the unsprung mass, Z r is the Laplace transform form of the vertical input displacement of the road surface unevenness, and s is the Laplace variable; Step (2): Establish a vertical electromagnetic excitation force model of a switched reluctance motor: Among them, is the vertical component of the unbalanced radial electromagnetic force, is the unbalanced radial electromagnetic force, i is the current, L is the winding inductance, k m is the phase of the motor, is the current of phase k of the motor m of the motor, θ is the relative angular displacement between the stator and the rotor of the motor, g m is the air gap of the motor when the motor is not eccentric, ε is the relative eccentricity of the air gap of the motor; when the number of poles of the stator in the opposite direction is m, the air gap of the motor is g m (n), the winding inductance of phase k of the motor m is When the number of poles of the stator in the opposite direction is n, the air gap of the motor is g m (m), the winding inductance of phase k of the motor m is Step (3): Determine the velocity-type impedance transfer function of a second-order terrain positive real network: Wherein, A, B, C, D, E, and F are coefficients, and their values are all greater than or equal to 0, and D, E, and F are not all 0, and satisfy the positive real constraint condition; Step (4): Determine the ground-hitch control algorithm: When the vertical velocity of the unsprung mass and the relative vertical velocity between the sprung mass and the unsprung mass are in the same direction, the damping coefficient of the control input is c max ; Conversely, when the vertical velocity of the unsprung mass is in the opposite direction to the relative vertical velocity between the sprung mass and the unsprung mass, the damping coefficient of the control input is c min ; The ground-hitch damping coefficient c g shall satisfy the following formula: where c max and c min are two damping coefficient values and satisfy the following equation: c max > c min > 0; Step (5): Select the vertical input displacement z of the road surface unevenness r , specifically: where, v represents the driving speed, n0 represents the reference spatial frequency, G q (n0) represents the road surface unevenness coefficient, w(t) represents Gaussian white noise with a mean of 0, and z r is the vertical input displacement of the road surface unevenness, and is the vertical input speed of the road surface unevenness; Step (6): Use a genetic algorithm to determine the response index constraint conditions of the second-order terrain positive real network: BA(z) < 1.08×BA pass , SWS(z) < SWS pass , DTL(z) < DTL pass , Among them, BA(z) is the root mean square value of the body acceleration of the second-order floor positive real network, SWS(z) is the root mean square value of the suspension dynamic stroke of the second-order floor positive real network, DTL(z) is the root mean square value of the tire dynamic load of the second-order floor positive real network, and BA pass is the root mean square value of the body acceleration of the traditional passive suspension, and SWS pass is the root mean square value of the suspension dynamic stroke of the traditional passive suspension, and DTL pass is the root mean square value of the tire dynamic load of the traditional passive suspension; Step (7): Use the genetic algorithm to obtain the dynamic inertia suspension model parameters A, B, C, D, E, F, c max , c min ; Step (8): Substitute the suspension model parameters obtained in step (7) into the velocity-type impedance transfer function T(s) of the second-order terrain positive real network and the terrain damping to obtain the dynamic inertia suspension model of the wheel hub drive vehicle.
2. The method according to claim 1, wherein Wherein, In the said step (2), g m (n) = (1 - ε)g m , g m (m) = (1 + ε)g m , |m - n| = 4; ε is: Wherein, △g is the absolute eccentricity of the motor rotor in the radial direction.
3. The method according to claim 2, wherein Wherein, In the step (2), when the number of poles of the stator in the opposite direction is m, the air gap of the motor is g m (n), the motor k m The winding inductance of the phase is as follows: Among them, N r is the number of poles of the motor rotor, a n , b n are polynomial fitting coefficients, L u is the inductance at the pole-slot alignment position, and L n is the Fourier coefficient of the inductance of each phase winding; When the number of poles of the stator in the opposite direction is n, the air gap of the motor is g m (m), motor k m The winding inductance of phase k is as follows:
4. The method according to claim 1, wherein In step (3), the positive real constraint on the velocity-type impedance transfer function T(s) of the second-order terrain positive real network is:
5. A method according to claim 1, characterized in that, Wherein, The specific content of step (7) is as follows: The genetic algorithm mainly includes generating an initial population, calculating fitness values, selection, crossover, mutation, and iteration. Specifically: Step (7.1): Generate an initial population as a randomly generated population size; Step (7.2): Taking the dynamic tire load as the optimization objective, design the calculation formula for the single-objective optimization fitness function of the genetic algorithm as: Wherein, Punishment is the penalty number; Step (7.3): Selection means directly inheriting the optimized individuals to the next generation or generating new individuals through crossover and then inheriting them to the next generation; Step (7.4): Crossover means replacing and recombining some genes of two individuals in the population to generate new individuals; Step (7.5): Mutation means changing the gene values at some gene loci of the individuals in the population; Step (7.6): Iteration means calculating whether the termination condition is satisfied. If not, continue the above operations until the maximum number of iterations is reached; if satisfied, output the optimal solution.
6. A method according to claim 5, wherein In step (7.2), the value-taking rule of the punishment number Punishment is as follows: as long as the root mean square value BA(z) of the body acceleration of the second-order floor positive-real network, the root mean square value SWS(z) of the suspension dynamic stroke of the second-order floor positive-real network, or the root mean square value DTL(z) of the tire dynamic load of the second-order floor positive-real network is greater than 1.08 times the root mean square value 1.08×BA of the body acceleration of the traditional passive suspension pass , the root mean square value SWS of the suspension dynamic stroke of the traditional passive suspension pass , the root mean square value DTL of the tire dynamic load of the traditional passive suspension pass , then the value of the punishment number Punishment is 1000; otherwise, the value is 0.
Citation Information
Patent Citations
A Method for Establishing a Second-Order Ideal Model of Vehicle ISD Suspension Based on ADD Positive Real Network Optimization
CN110001337B