A robust design method for technical parameters for performance optimization of passive suspension systems

Abstract

This invention discloses a robust design method for technical parameters of passive suspension systems aimed at performance optimization. It can robustly design two types of suspension system parameters—stiffness coefficient and damping coefficient—to cope with changes in vehicle weight. Specifically, the method constructs a four-wheel vehicle passive suspension system model and uses an improved Harris-Eagle optimization algorithm to search for the suspension stiffness coefficient and damping coefficient under varying vehicle weight conditions, thereby ensuring a balanced balance between vehicle ride smoothness and handling stability. This invention is a technical parameter design method for four-wheel vehicle passive suspension systems that does not rely on trial and error by technicians to optimize the technical parameters of the passive suspension system. More importantly, this invention considers the uncertainty of vehicle weight during actual driving; the designed parameters are optimal under different vehicle weight variations, exhibiting a high degree of robustness.

Description

Technical Field

[0001] This invention relates to a design method for a vehicle passive suspension system, and more particularly to a robust design method for technical parameters aimed at optimizing the performance of a passive suspension system. Background Technology

[0002] The automotive suspension is a flexible device that connects the chassis and axles in a car. It generally consists of elastic elements, guiding mechanisms, and damping shock absorbers. The main task of the elastic elements and damping shock absorbers is to mitigate the impact transmitted from uneven road surfaces to the chassis, thereby improving ride comfort. The design of passive suspension systems requires the design of elastic elements and damping shock absorbers with appropriate technical parameters to not only meet the requirements for vehicle ride smoothness but also to fully guarantee the suspension's handling stability. Suspension springs and dampers are two essential components of an automotive suspension system. Suspension springs and dampers with different technical parameters will result in completely different ride smoothness and handling stability. Traditional design methods rely on the experience of technicians and trial-and-error methods to continuously try and obtain suitable stiffness and damping coefficients—two key technical parameters.

[0003] Vehicle ride stability refers to the impact on the vehicle body caused by uneven road surfaces during normal driving, which can affect ride comfort or damage cargo. Prolonged exposure to such harsh environments can lead to severe fatigue and chronic illnesses. Therefore, reducing the vertical, pitch, and roll acceleration of the vehicle body is a primary goal of passive suspension system design. Furthermore, vehicle handling stability mainly addresses the changes in handling stability caused by variations in the elasticity of the suspension tires. When the elasticity of the suspension tires is too high, the friction between the tires and the ground increases, leading to higher fuel consumption; conversely, when the elasticity of the suspension tires is too low or even airborne, the vehicle's direction becomes increasingly uncontrollable, thus affecting driving performance.

[0004] Furthermore, considering that the vehicle's body mass (or sprung mass) varies with the number of passengers or cargo, different body weights will also affect the vehicle's stability and handling. Therefore, the actual load mass of a vehicle during operation is uncertain, and the body mass will constantly change depending on the number of passengers and the weight of cargo. Thus, robustly designing the most suitable technical parameters to address the impact of fluctuating body mass is of practical significance. Generally, designing appropriate stiffness and damping coefficients requires starting from the perspective of whole-vehicle modeling of the suspension system. However, considering the uncertainty of sprung mass and the resulting change in the center of gravity, robustly designing appropriate stiffness and damping coefficients for the front and rear suspensions of a vehicle remains an unsolved technical challenge. Summary of the Invention

[0005] The main technical problem this invention aims to solve is: considering uncertainties such as vehicle weight, how to robustly design the stiffness coefficient and damping coefficient of a suspension system to cope with changes in vehicle mass. Specifically, this invention constructs a four-wheel vehicle passive suspension system model and uses an improved Harris-Eagle optimization algorithm to search for the suspension stiffness coefficient and damping coefficient under the influence of changes in vehicle mass (also known as sprung mass), thereby ensuring a balanced balance between vehicle ride smoothness and handling stability.

[0006] The technical solution adopted by the method of the present invention to solve the above problems is as follows: a robust design method for technical parameters for performance optimization of passive suspension systems, comprising the following steps:

[0007] Step (1): Sequentially set the 14 state variables of the four-wheel vehicle passive suspension system as follows: vehicle vertical velocity, vehicle pitch velocity, vehicle roll velocity, vehicle vertical displacement, vehicle pitch angle, vehicle roll angle, left front suspension vertical velocity, right front suspension vertical velocity, left rear suspension vertical velocity, right rear suspension vertical velocity, left front suspension vertical displacement, right front suspension vertical displacement, left rear suspension vertical displacement, and right rear suspension vertical displacement, and set... After representing the specific values ​​of these 14 state variables at the j-th travel time, the following discrete state-space model is constructed:

[0008]

[0009] Where i equals 1, 2, 3, and 4 respectively, they represent the left front suspension, right front suspension, left rear suspension, and right rear suspension, respectively, and the time interval Δt ranges from 0.002 ≤ Δt ≤ 0.02. The values ​​of the 14 state variables at the j-th travel moment are represented sequentially. The unsprung mass m of the suspension spring is a fixed parameter. I1 represents the pitch inertia of the vehicle body, I2 represents the roll inertia of the vehicle body, L1 represents the torque between the vehicle's center of gravity and the center of gravity of the left front suspension (or right front suspension), L2 represents the torque between the vehicle's center of gravity and the center of gravity of the left rear suspension (or right rear suspension), L3 represents the torque between the center of gravity of the left front suspension and the vehicle's center of gravity, and L4 represents the torque between the center of gravity of the right front suspension and the vehicle's center of gravity. The vehicle mass M changes with the number of passengers or the weight of cargo; therefore, the vehicle mass M is within the interval [M]. min M max A changing parameter within; the specific value of tire elasticity at the j-th moment of travel. The linear elastic coefficient k1 and nonlinear elastic coefficient k2 of the tire are known and fixed parameters. The specific value of the tire deformation at the j-th driving moment is... This represents the specific value of the tire contact patch height at the j-th moment of travel.

[0010] In equation ① above, when i = 1, and They refer to and and They refer to and When i = 2 and They refer to and and They refer to and and These represent the specific values ​​of the suspension spring force and damping force at the (j-1)th driving moment, and the corresponding calculation method is shown in formula ②:

[0011]

[0012] In equation ② above, This represents the relative change in the vertical speed of the suspension at the (j-1)th moment of travel; when At that time, the sign function when At that time, the sign function K lin and k non C represents the linear stiffness coefficient and nonlinear stiffness coefficient of the suspension spring, respectively. lin C sys and C non These represent the linear damping coefficient, symmetrical damping coefficient, and nonlinear damping coefficient of the suspension damper, respectively. This represents the relative change in vertical displacement of the suspension at the (j-1)th travel moment, and its specific calculation method is shown in formula ③:

[0013]

[0014] This represents the relative change in the vertical speed of the suspension at the (j-1)th travel moment, and its specific calculation method is shown in formula ④:

[0015]

[0016] Step (2): Determine the range of variation of the technical parameters to be designed, and determine the range of variation of the vehicle body mass M, specifically including: the upper limit of the linear stiffness coefficient. and lower limit Upper limit of nonlinear stiffness coefficient and lower limit Upper limit of linear damping coefficient and lower limit Upper limit of symmetrical damping coefficient and lower limit Upper limit of nonlinear damping coefficient and lower limit The upper limit of vehicle body mass M max and lower limit M min .

[0017] It should be noted that the linear stiffness coefficient and nonlinear stiffness coefficient are key technical parameters determining the spring force of the suspension. In actual design, both linear and nonlinear stiffness coefficients have their own upper and lower design limits, which determine the range of variation for these technical parameters. In other words, the design and selection of both linear and nonlinear stiffness coefficients have a designable range. Similarly, the linear damping coefficient, symmetrical damping coefficient, and nonlinear damping coefficient of the suspension damper also have their own corresponding designable ranges.

[0018] In addition, the lower limit of vehicle body weight M min This is equal to the vehicle's mass when empty, and the upper limit of the vehicle's mass is M. max Equal to M min Including the mass under full load.

[0019] Step (3): In the interval [M] min M max Generate M on the above min and M max A equally spaced values ​​M1, M2, ..., M A Then, according to the formulas shown below, they are respectively Generate N sets of values:

[0020]

[0021] Among them, the road surface unevenness G0 = 1024 × 10 -6 random numbers It follows a normal distribution with a mean of 0 and a standard deviation of 1. When j = 1,

[0022] It should be noted that from the interval [M] min M max Generate M on the above min and M max The A equally spaced values ​​are: M1 = M mi n, M2 = M min +(M max -M min ) / A, M3=M min +2·(Mmax -M min ) / A, until M A =M max .

[0023] Step (4): Determine the quantitative index J of the passive suspension system performance, and its calculation method is shown in Formula ⑥:

[0024]

[0025] In equation ⑥ above, the weighting coefficient λ ranges from 0.2 to 0.8. This represents the change in the θ-th state variable at the j-th travel time. and They represent The maximum and minimum values ​​in the range. and They represent The maximum and minimum values ​​in the range, i = 1, 2, 3, 4; when θ equals 1, 2 and 3 respectively, In sequence and

[0026] As can be seen from the definition of the quantitative index J in Equation ⑥ above, J is equal to the weighted sum of the vehicle's vertical acceleration, pitch acceleration, roll acceleration, and the elasticity of the four tires. Acceleration measures the stability of the vehicle's movement, while positive tire elasticity leads to increased ground friction and thus increased energy consumption, while negative tire elasticity indicates insufficient grip, thus affecting the vehicle's handling performance.

[0027] Step (5): According to the uniform distribution from the interval H random numbers are generated to form a 1×H dimensional row vector p1, which is then distributed uniformly from the interval [0, 1]. H random numbers are generated from the above to form a 1×H dimensional row vector p2, which is then distributed uniformly from the interval [0, 1]. H random numbers are randomly generated to form a 1×H dimensional row vector p3, which is then distributed uniformly from the interval [0, 1]. H random numbers are generated to form a 1×H dimensional row vector p4, which is then distributed uniformly from the interval [0, 1]. First, generate H random numbers to form a 1×H dimensional row vector p4. Then, merge the row vectors p1, p2, p3, p4, and p5 into a 5×H dimensional real matrix P.

[0028] Step (6): After obtaining the optimal individual vector μ0 using the improved Harris-Eagle optimization algorithm, set K... lin K non C lin Csys and C non After the first, second, third, fourth and fifth data in μ0 are respectively equal, the suspension springs and dampers that match these five technical parameters are designed and selected; the implementation process of the improved Harris-Eagle optimization algorithm to search for the optimal individual vector β0 is specifically shown in steps (6.1) to (6.8).

[0029] Step (6.1): After initializing the number of iterations γ = 1, determine the maximum number of iterations F and the escape energy E. γ The decreasing method and the loss function value corresponding to the individual vector β The calculation method; where E γ The decreasing method is shown in Formula ⑦, where the loss function value corresponding to the individual vector β is... The specific calculation method is shown in steps (1) to (7).

[0030]

[0031] Where γ represents the number of iterations.

[0032] Step (1): Assign the 5 data points from the individual vector β to K in sequence. lin K non C lin C sys and C non That is, set K respectively lin K non C lin C sys and C non The first, second, third, fourth, and fifth data points in β are equal to α, and after initializing α = 1, the system is set to operate when the vehicle mass M = M α .

[0033] Step (2): Initialize j = 1 and set All equal 0.

[0034] Step (3): Utilize And according to the formula Calculate the elastic force of each of the four tires at the j-th moment of travel. And calculate according to formula ② as well as

[0035] Step (four): Calculate according to formula ① Then, according to the formula Calculate separately in, Let θ represent the state variable at the j-th travel time, where θ = 1, 2, ..., 14.

[0036] Step (5): Determine if j is less than N; if yes, set j = j + 1 and return to step (3); if no, then... The maximum and minimum values ​​in the data are recorded as follows: and Will The maximum and minimum values ​​in the data are recorded as follows: and Will The maximum and minimum values ​​in the data are recorded as follows: and Then The maximum and minimum values ​​in the data are recorded as follows: and Then, proceed to step (six).

[0037] Step (six): Calculate the quantitative index J according to formula ⑥ in step (4), and record it as J. α .

[0038] Step (7): Determine if α is less than A; if yes, set α = α + 1, then set the vehicle mass M = Mα and return to step (2); if no, then follow the formula... Calculate the loss function value corresponding to the individual vector β.

[0039] Step (6.2): ​​Set individual vectors β1, β2, ..., β H After assigning the column vectors to the first, second, ..., H columns of the real matrix P in sequence, calculate the individual vectors β1, β2, ..., β1 respectively. H Corresponding loss function value Then, a random number R0 is randomly generated based on a normal distribution with a mean of 0.5 and a standard deviation of 1.

[0040] In step (6.3) above, the loss function value corresponding to the individual vector β1 is calculated. In this case, β = β1 must first be set, and then the loss function value is calculated according to the calculation process from step (i) to step (vii) above. Reset The process of calculating the loss function values ​​for the remaining individual vectors follows the same logic.

[0041] Step (6.3): ... The four smallest loss function values After recording the corresponding individual vectors as reference individual vectors μ1, μ2, μ3, and μ4, the optimal individual vector μ0 is calculated according to the formula shown below (⑧):

[0042]

[0043] Step (6.4): Determine E γ absolute value | E γ | Is it less than 1? If not, proceed to step (6.5) to update the individual vectors β1, β2, ..., β H Then, proceed to step (6.8); if so, execute step (6.6) to update the individual vectors β1, β2, ..., β H Then, proceed to step (6.8).

[0044] Step (6.5): Generate two random numbers R1 and R2 from the interval [0, 1] according to a uniform distribution, and then generate a random positive integer ε from the interval [1, H]. Update the individual vectors β1, β2, ..., β according to the formula shown below (9). H :

[0045]

[0046] In equation ⑨ above, when h equals 1, 2, ..., H respectively, β h That is, corresponding to β1, β2, ..., β H p max It is by This forms a 5×1 dimensional column vector, p min It is by This forms a 5×1 dimensional column vector.

[0047] Step (6.6): Determine if R0 is less than 0.5; if not, update the individual vectors β1, β2, ..., β according to the formula shown below (10). H If so, proceed to step (6.7).

[0048]

[0049] In equation ⑩ above, R3 represents generating a random number from the interval [0, 1] according to a uniform distribution, where h = 1, 2, ..., H.

[0050] Step (6.7): According to the formula shown below Calculate the Lever flight value L β Then, according to the formula Calculate the pseudo-individual vector δ, and then apply the formula. Update individual vectors β1, β2, ..., β H .

[0051]

[0052]

[0053]

[0054] Wherein, the gamma function Γ(2.5) = 1.3293, the gamma function Γ(1.25) = 0.9064, and R4, R5, and R6 are three random numbers generated from the interval [0, 1] according to a uniform distribution. This represents the loss function value calculated after setting β = δ. This indicates that β = δ + R5·L is set. β The loss function value is then calculated.

[0055] Step (6.8): Determine if γ is less than F; if so, set γ = γ + 1 and update R0, then calculate β1, β2, ..., β1 respectively. H Corresponding loss function value Then, return to step (6.3); otherwise, obtain the optimal individual vector μ0; where the specific method for updating R0 is as follows:

[0056]

[0057] The advantages of the method of the present invention, based on the above implementation steps, are as follows.

[0058] This invention provides a technical parameter design method for optimizing the performance of a four-wheeled vehicle passive suspension system. It does not rely on the experience of technicians, nor does it require continuous trial and error to optimize the technical parameters of the suspension springs and dampers. More importantly, this invention considers the uncertainty of vehicle weight during actual driving, ensuring that the designed parameters are optimal under different vehicle weight conditions, exhibiting a high degree of robustness. Therefore, the technical parameters designed by this invention are also a robust design method, reliably guaranteeing the vehicle's smoothness and handling stability even with changes in vehicle weight. Attached Figure Description

[0059] Figure 1 This is a schematic diagram illustrating the implementation process of the method of the present invention. Detailed Implementation

[0060] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0061] This invention discloses a robust design method for technical parameters for performance optimization of passive suspension systems. The following is a combination of... Figure 1 The implementation flowchart shown below illustrates the specific implementation method of the present invention.

[0062] Step (1): Determine the 14 state variables of the four-wheel vehicle passive suspension system as follows: vehicle vertical velocity, vehicle pitch velocity, vehicle roll velocity, vehicle vertical displacement, vehicle pitch angle, vehicle roll angle, left front suspension vertical velocity, right front suspension vertical velocity, left rear suspension vertical velocity, right rear suspension vertical velocity, left front suspension vertical displacement, right front suspension vertical displacement, left rear suspension vertical displacement, and right rear suspension vertical displacement, and set... After representing the specific values ​​of these 14 state variables at the j-th travel time, a discrete state-space model as shown in formula ① is constructed.

[0063] Step (2): Determine the range of values ​​for the technical parameters to be designed, and then determine the range of variation of the vehicle body mass M.

[0064] The four-wheeled vehicle has four suspensions, each equipped with one suspension spring and one damper. All suspension springs and dampers use the same stiffness coefficient.

[0065] Step (3): Generate a file containing M min and M max A equally spaced values ​​M1, M2, ..., M A Then, according to formula ⑤ above, they are respectively Generate specific values ​​for N travel times.

[0066] Step (4): Determine the quantitative index J of the passive suspension system performance, and its calculation method is shown in Formula ⑥.

[0067] Step (5): According to the uniform distribution from the interval H random numbers are generated to form a 1×H dimensional row vector p1, which is then distributed uniformly from the interval [0, 1]. H random numbers are generated from the above to form a 1×H dimensional row vector p2, which is then distributed uniformly from the interval [0, 1]. H random numbers are randomly generated to form a 1×H dimensional row vector p3, which is then distributed uniformly from the interval [0, 1]. H random numbers are generated to form a 1×H dimensional row vector p4, which is then distributed uniformly from the interval [0, 1]. First, generate H random numbers to form a 1×H dimensional row vector p4. Then, merge the row vectors p1, p2, p3, p4, and p5 into a 5×H dimensional real matrix P.

[0068] Step (6): After obtaining the optimal individual vector μ0 using the improved Harris-Eagle optimization algorithm, set K... lin K non C lin C sys and Cnon After the values ​​are successively equal to the 1st, 2nd, 3rd, 4th and 5th data in μ0, the suspension springs and dampers with the same technical parameters as these 5 are selected in the design.

Claims

1. A technical parameter robust design method for passive suspension system performance optimization, characterized in that, Specifically comprising the steps shown below: Step (1): Set the 14 state variables of the four-wheel passive suspension system in turn as the body vertical velocity, the body pitch velocity, the body roll velocity, the body vertical displacement, the body pitch angle, the body roll angle, the left front suspension vertical velocity, the right front suspension vertical velocity, the left rear suspension vertical velocity, the right rear suspension vertical velocity, the left front suspension vertical displacement, the right front suspension vertical displacement, the left rear suspension vertical displacement and the right rear suspension vertical displacement, and set After sequentially representing the specific values of the 14 state variables at the jth driving moment, build the discrete state space model as shown in formula ①: wherein, when i is equal to 1, 2, 3 and 4 respectively, it refers to the left front suspension, the right front suspension, the left rear suspension and the right rear suspension in turn, Δt represents the time interval, denote the change value of the 14 state variables at the jth driving moment; m represents the unsprung mass of the suspension spring, I1 represents the moment of inertia of the body pitch, I2 represents the moment of inertia of the body roll, L1 represents the moment between the vehicle mass center and the left front suspension mass center, L2 represents the moment between the vehicle mass center and the left rear suspension mass center, L3 represents the moment between the left front suspension mass center and the vehicle mass center, L4 represents the moment between the right front suspension mass center and the vehicle mass center, M represents the body mass; the specific value of the tire elastic force at the jth driving moment k1 and k2 represent the linear and nonlinear elastic coefficients of the tire respectively, and the specific value of the tire deformation at the jth driving moment denote the specific value of the protrusion height of the tire contacting the ground at the jth driving moment; and denote the specific value of the suspension elastic force and damping force at the (j-1)th driving moment, and the corresponding calculation method is shown in formula ②: In the above formula 2, denotes the relative change amount of the suspension vertical velocity at the j-1th driving time; when , the sign function , the sign function , the sign function K lin and k non respectively denote the linear stiffness coefficient and the nonlinear stiffness coefficient, C lin , C sys and C non respectively denote the linear damping coefficient, the symmetric damping coefficient and the nonlinear damping coefficient, denotes the relative change amount of the suspension vertical displacement at the j-1th driving time, and the specific calculation method is shown in formula 3: represents the relative change amount of the suspension vertical velocity at the j-1th driving time, and the specific calculation method is shown in formula (IV): Step (2): determining the change range of the technical parameters to be designed, and determining the change range of the vehicle body mass M, specifically including: the upper limit and the lower limit of the linear stiffness coefficient and the lower limit of the nonlinear stiffness coefficient and the lower limit of the linear damping coefficient and the lower limit of the symmetric damping coefficient and the lower limit of the nonlinear damping coefficient max and the upper limit M min and the lower limit M of the vehicle body mass M Step (3): In the interval [M] min M max Generate M on the above min and M max A equally spaced values ​​M1, M2, ..., M A Then, according to the formulas shown below, they are respectively Generate N sets of specific values: Wherein, the road roughness G0=1024x10 -6 , random number subject to a normal distribution with mean equal to 0 and standard deviation equal to 1, when j=1, Step (4): determining the quantitative index J of passive suspension system performance, which is calculated in the manner shown in formula (6): In the above formula (6), λ represents a weight coefficient, represents a change value of the θth state variable at the jth travel time, and respectively represent a maximum value and a minimum value in , and respectively represent a maximum value and a minimum value in , i = 1, 2, 3, 4; and when θ is equal to 1, 2, and 3, respectively, in turn represent and Step (5): Randomly generate H random numbers from the interval above to form a 1 x H dimensional row vector p1, randomly generate H random numbers from the interval above to form a 1 x H dimensional row vector p2, randomly generate H random numbers from the interval above to form a 1 x H dimensional row vector p3, randomly generate H random numbers from the interval above to form a 1 x H dimensional row vector p4, randomly generate H random numbers from the interval above to form a 1 x H dimensional row vector p4, and then combine the row vectors p1, p2, p3, p4, p5 into a 5 x H dimensional real number matrix P. Step (6): After the optimal individual vector μ0 is searched by the improved Harris-Hawk optimization algorithm, set K lin , K non , C lin , C sys and C non equal to the 1st, 2nd, 3rd, 4th and 5th data in μ0 respectively, and then design and select the suspension spring and damper matched with the five technical parameters.

2. The technical parameter robust design method for passive suspension system performance optimization according to claim 1, wherein The implementation process of searching for the optimal individual vector β0 in the step (6) by using the improved Harris-Hawk optimization algorithm is shown below: Step (6.1): Set the iteration number γ equal to 1, set the maximum iteration number equal to F, and then determine the escape energy E γ in a decreasing manner, and the loss function value corresponding to the individual vector β is calculated Step (6.2): Set individual vectors β1, β2, …, β H After sequentially equaling the column vectors of the 1st column, the 2nd column to the Hth column in the real number matrix P, respectively calculate individual vectors β1, β2, …, β H The corresponding loss function value Then, according to the normal distribution with the mean of 0.5 and the standard deviation of 1, a random number R0 is randomly generated; Step (6.3): The following steps are performed: The four smallest loss function values among the 1000 loss function values After the four smallest loss function values and their corresponding individual vectors are recorded as reference individual vectors μ1, μ2, μ3, μ4, the optimal individual vector μ0 is calculated according to the following formula (VI): Step (6.4): judging E γ the absolute value of E γ is less than 1; if not, then performing step (6.5) to update the individual vectors β1, β2,..., β H , and then jumping to step (6.8); if yes, then performing step (6.6) to update the individual vectors β1, β2,..., β H , and then jumping to step (6.8); Step (6.5): Update the individual vector β1, β2,..., βnaccording to the following formula (VII) after generating two random numbers R1and R2from the interval [0, 1] and a random positive integer ε from the interval [1, H] according to uniform distribution. H : In the above equation (7), β h corresponding to β1, β2,..., β H , p max is a 5 x 1 column vector composed of p min is a 5 x 1 column vector composed of p Step (6.6): Determine whether Ro is less than 0.5; if not, update the individual vectors β1, β2,..., β H ; if yes, perform step (6.7); In the above formula (9), R3 represents generating a random number from the interval [0, 1] according to a uniform distribution, h = 1, 2, …, H; Step (6.7): Calculate the Levy flight value L according to the following formula (X) β After that, calculate the pseudo individual vector δ according to the formula Calculate the pseudo individual vector δ, and then update the individual vectors β1, β2, …, β Update the individual vectors β1, β2, …, β H ; where the gamma function Γ(2.5) = 1.3293, the gamma function Γ(1.25) = 0.9064, R4, R5and R6are three random numbers generated from the interval [0, 1] according to a uniform distribution, denotes the loss function value calculated after setting β = δ, denotes the loss function value calculated after setting β = δ + R5·L β denotes the loss function value calculated after setting β = δ + R5·L Step (6.8): Determine whether γ is less than F; if yes, set γ = γ + 1 and update R0, and then calculate β1, β2, …, β H The corresponding loss function value After that, return to step (6.3); if no, the optimal individual vector μ0 is obtained.

3. The method of claim 2, wherein, The step (6.1) of decreasing E γ is specified as follows: Wherein, γ represents the iteration number.

4. The technical parameter robust design method for passive suspension system performance optimization according to claim 2, characterized in that, The loss function value corresponding to the individual vector β in step (6.1) The calculation method is specifically shown in steps (i) to (vi): Step (i): assign the 5 data in individual vector β to K lin , K non , C lin , C sys and C non in turn, and initialize α = 1, then set the body mass M = M α ; Step ii: Initialize j = 1 and set all equal to 0; Step iii) calculating the value of the four tire spring forces at the jth driving moment using the formula and the formula respectively and according to the formula ② respectively and Step iv: Calculate according to formula Then, calculate according to formula respectively wherein, represents the specific value of the θth state variable at the jth driving time, θ = 1, 2, …, 14; Step 5: Determine if j is less than N; if yes, set j = j + 1 and return to step 3; if no, then... The maximum and minimum values ​​in the data are recorded as follows: and Will The maximum and minimum values ​​in the data are recorded as follows: and Will The maximum and minimum values ​​in the data are recorded as follows: and Then The maximum and minimum values ​​in the data are recorded as follows: and Then, proceed to step (vi); Step (vii): After calculating the quantification index J according to the formula (vi) in step (4), set J α = J; Step viii: judging whether a is smaller than A; if yes, setting a = a + 1 and then setting the mass of the vehicle body M = M α and returning to step ii; if no, calculating the loss function value corresponding to the individual vector β according to the formula ​ 5. The technical parameter robust design method for passive suspension system performance optimization according to claim 2, wherein, The specific manner of updating R0 in the step (6.8) is: if R0 is less than 0.7, then the calculation result of R0 / 0.7 is assigned to R0; otherwise, the calculation result of 10·(1-R0) / 3 is assigned to R0.

Images