A method for identifying parameters of a hyperbolic tangent model of a magneto-rheological torsional vibration damper

Abstract

The application discloses a kind of identification methods for hyperbolic tangent model parameters of magnetorheological torsional vibration damper, first, servo fatigue motor is used to test the mechanical properties of magnetorheological torsional vibration damper, and the hysteretic curve of magnetorheological torsional vibration damper is obtained, a plurality of data is obtained by changing frequency and current, the parameters of hyperbolic tangent model are identified using experimental data, first, the stiffness coefficient k is solved according to physical law, then the viscous damping coefficient c is solved based on the mathematical characteristics of hyperbolic tangent model, finally, the parameter identification range of other parameters is limited according to the mathematical characteristics, and the parameter identification is carried out by using optimization algorithm, finally, the parameter identification is realized by fitting the parameter values under different currents.The application has eliminated the mutation and non-smooth phenomenon of zero point area, improved the identification accuracy and solution quality, greatly increased the success rate, greatly shortened the calculation time, and showed the characteristics of double advantages in efficiency and performance.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligence and intelligent materials, and more particularly, it relates to a method for identifying hyperbolic tangent model parameters of a magneto-rheological torsional vibration damper. BACKGROUND

[0002] At present, the magneto-rheological torsional vibration damper attracts the attention of many experts and scholars, which adjusts the magnetic field strength of the environment where the magneto-rheological fluid is located by controlling the current size in the coil, and then changes the viscosity coefficient of the magneto-rheological fluid. The change of the viscosity of the magneto-rheological fluid directly affects the shear force when the device rotates, so as to realize the precise control of the output damping force of the torsional vibration damper, which has the advantages of continuous adjustment, low power consumption and fast response, and is suitable for various vibration damping demand scenes. However, due to its special mechanical properties-hysteresis characteristics, the mathematical model is strongly nonlinear, making the establishment process quite complex and difficult.

[0003] Many models have been established to describe the nonlinear characteristics of magneto-rheological devices by domestic and foreign scholars. At present, the commonly used mathematical models of magneto-rheological dampers include Bingham model, Bouc-Wen model, polynomial model and hyperbolic tangent model, etc. Among them, the Bingham model was first proposed to describe the nonlinear characteristics of magneto-rheological devices, but it cannot well describe the characteristics in the low speed area. The Bouc-Wen model can accurately reflect the hysteresis characteristics of the magneto-rheological damper at low speed and high speed, but the undetermined parameters involved in the Bouc-Wen model except the viscous damping coefficient which is easy to determine are difficult to determine, and the differential term calculation is too complex, which is easy to be affected by the excitation frequency. The polynomial model has a simple expression and relatively high precision, but its parameters have no physical meaning, and there is obvious oscillation in the high speed area. Compared with the Bingham model, the hyperbolic tangent model has higher precision and can well describe the hysteresis characteristics of the magneto-rheological torsional vibration damper. Compared with the Bouc-Wen model which also reflects the hysteresis characteristics, the hyperbolic tangent model has the advantages of no differential term in the identified parameters, less affected by the excitation frequency, high precision in the speed transition area, etc. In addition, compared with the polynomial model, the parameters of the hyperbolic tangent model have certain physical meaning, and it performs more smoothly in the high speed area. Therefore, it is widely used in the establishment of forward and inverse models of magneto-rheological devices and control.

[0004] Because the hyperbolic tangent model needs to identify more parameters, generally uses intelligent algorithm to identify, Hu Haigang etc. use least square method to identify hyperbolic tangent model, fitting precision is good, but when programming, it is more complex, N.M.Kwok etc. use particle swarm optimization (particleswarm optimization, PSO) in identification process to identify, Zhu etc. pass through genetic algorithm (geneticalgorithms, GA) to the parameter identification of system.But the intelligent algorithm identification of above is easy to fall into local optimal solution, time-consuming is longer, and the size of parameter value also significantly influences the fullness degree of hysteresis loop and the continuity in zero point area, because in general only uses intelligent algorithm, such as genetic algorithm, does not propose the limiting method of parameter value range to the problem, leading to the torque-displacement curve of hyperbolic tangent model fitting, and the discontinuity of displacement is 0, influence control effect, and only using intelligent algorithm to identify the parameter of hyperbolic tangent model, will make the parameter identified without actual physical significance. SUMMARY

[0005] The purpose of the present application is to design and develop a kind of identification method for the parameters of hyperbolic tangent model of magneto-rheological torsional vibration damper, the stiffness coefficient and viscous damping coefficient are identified by constructing hyperbolic tangent model, other parameters are identified by combining genetic algorithm, improve accuracy and identification efficiency.

[0006] The technical scheme provided by the present application is as follows:

[0007] A kind of identification method for the parameters of hyperbolic tangent model of magneto-rheological torsional vibration damper, comprising the following steps:

[0008] Step one, the corresponding relative velocity, relative angular displacement input and torsional torque of loading are collected;

[0009] Step two, construct the hyperbolic tangent model of magneto-rheological torsional vibration damper:

[0010]

[0011]

[0012] In the formula, T is torsional torque, c is the viscous damping coefficient of magneto-rheological torsional vibration damper, k is the stiffness coefficient of magneto-rheological torsional vibration damper, θ (t) is relative angular displacement input, t is time, α is the expansion coefficient of hysteresis function, z (θ (t)) is hysteresis operator, Relative velocity is, β is the coefficient determined by the characteristics of magneto-rheological fluid and the structure of magneto-rheological torsional vibration damper, δ is the coefficient determined by magneto-rheological fluid and control system, sign (·) is sign function;

[0013] Step three, according to the hyperbolic tangent model of the magneto-rheological torsional vibration damper, the viscous damping coefficient of the magneto-rheological torsional vibration damper, the stiffness coefficient of the magneto-rheological torsional vibration damper, the expansion coefficient of the hysteresis function, the coefficient determined by the magneto-rheological fluid characteristics and the structure of the magneto-rheological torsional vibration damper, and the coefficient determined by the magneto-rheological fluid and the control system are obtained under the constraint condition that:

[0014]

[0015]

[0016] α>0;

[0017]

[0018] In the formula, T ′ is the corresponding torsional moment of the changed relative angular displacement input, q is a constant, is the relative speed of the i th data point in the curve, N is the number of test data points, y i is the response value of the i th data point in the curve, is the average value of the relative speed of the parallel region in the curve, y avg is the average value of the response value of the parallel region in the curve;

[0019] Step four, the viscous damping coefficient of the magneto-rheological torsional vibration damper, the stiffness coefficient of the magneto-rheological torsional vibration damper, the expansion coefficient of the hysteresis function, the coefficient determined by the magneto-rheological fluid characteristics and the structure of the magneto-rheological torsional vibration damper, and the coefficient determined by the magneto-rheological fluid and the control system satisfy:

[0020] c=0.1166+1.0806I-0.1424I 2 +0.006I 3 ;

[0021] k=0.0055;

[0022] α=0.5538+0.4176I+0.608I 2 -0.0745I 3 ;

[0023] δ=1.2894+1.4238I-0.521I 2 +0.063I 3 ;

[0024] β=6.5832+4.1262I-0.9494I 2 .

[0025] Preferably, the relative angular displacement input satisfies:

[0026] θ(t) = Asin(2πt);

[0027] where A is the amplitude of the excitation, and A > 0.

[0028] Preferably, the changed relative angular displacement input corresponds to a torsional moment that satisfies:

[0029]

[0030] where θ ′ (t) is the changed relative angular displacement input.

[0031] Preferably, the changed relative angular displacement input satisfies:

[0032] θ ′ (t) = θ(t) + q;

[0033] where q is a constant.

[0034] Preferably, the amplification coefficient of the hysteresis function, the coefficient determined by the characteristics of the MR fluid and the structure of the MR torsional vibration damper, and the coefficient determined by the MR fluid and the control system are identified by a genetic algorithm.

[0035] Preferably, the genetic algorithm comprises:

[0036] Step a, create an initial population with the corresponding relative velocity, relative angular displacement input, and torsional moment loaded;

[0037] Step b, define the fitness evaluation criteria, and set the initial population size, the probability of crossover occurrence, the probability of gene mutation, and the maximum iteration rounds;

[0038] Step c, randomly initialize the amplification coefficient of the hysteresis function, the coefficient determined by the characteristics of the MR fluid and the structure of the MR torsional vibration damper, and the coefficient determined by the MR fluid and the control system;

[0039] Step d, initialize a set of chromosomes for each parameter;

[0040] Step e, evaluate each chromosome using the fitness function value;

[0041] Step f, according to the fitness evaluation results of each chromosome, adopt a roulette wheel selection strategy to form a new population;

[0042] where the probability of each chromosome being selected is proportional to its fitness;

[0043] Step g, update the chromosomes through crossover and mutation behavior;

[0044] Step h, when the value of the fitness function converges or reaches the number of iterations, output the optimal model parameters.

[0045] Preferably, the fitness function is:

[0046]

[0047] In the formula, T i exp T is the true torsional moment of the i th data point, T i sim T is the calculated torsional moment of the i th data point.

[0048] Preferably, the initial population size is 400;

[0049] The probability of the configuration crossover is 0.8;

[0050] The probability of gene mutation is 0.2;

[0051] The maximum number of iterations is 500.

[0052] The beneficial effects of the present application are:

[0053] (1) The method for identifying the parameters of the hyperbolic tangent model of the magneto-rheological torsional vibration damper designed and developed by the present application uses the physical meaning and mathematical characteristics of the hyperbolic tangent model to identify the stiffness coefficient and the viscous damping coefficient first, then narrows the identification range, and uses intelligent algorithm to identify the remaining parameters. This method can well describe the mechanical properties of the magneto-rheological torsional vibration damper, the principle is intuitive and the calculation is efficient, and it can accurately capture the core mechanical performance characteristics of the magneto-rheological torsional vibration damper, providing a solid theoretical basis for the semi-active control strategy of the magneto-rheological torsional vibration damper, and ensuring the effective implementation and optimization of the control strategy.

[0054] (2) The method for identifying the parameters of the hyperbolic tangent model of the magneto-rheological torsional vibration damper designed and developed by the present application uses the mathematical characteristics of the hyperbolic tangent model and the physical characteristics of the magneto-rheological torsional vibration damper, aiming to give certain physical meaning to the parameters of the hyperbolic tangent model, solve the problem of possible mutation in the zero point region of the torque-angle curve, and reduce the situation that the previous algorithm identification is easy to fall into local optimal solution and takes a long time.

[0055] (3), the identification method for the hyperbolic tangent model parameters of the magneto-rheological torsional vibration damper designed and developed by the application is significantly better than the prior art in the quality and success rate of average solution, improves the accuracy, reduces the identification parameters and narrows the search range, effectively improves the iteration convergence speed, significantly shortens the running time of the algorithm, and improves the identification efficiency, and the identification method is constructed into a complete and systematic methodological system from the construction to the verification stage, the system not only has high practicability, but also can be used as an important tool for guiding theoretical learning, promotes the in-depth understanding and application of related theoretical knowledge. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 The figure is a schematic diagram of the relationship curve of the torsional moment-rotation angle of the magneto-rheological torsional vibration damper under different loads.

[0057] Figure 2 The figure is a schematic diagram of the relationship curve of the torsional moment-rotation speed of the magneto-rheological torsional vibration damper under different loads.

[0058] Figure 3 The figure is a schematic diagram of the relationship curve of the hysteresis operator-rotation speed under different load currents.

[0059] Figure 4 The figure is a schematic diagram of the relationship curve of the intermediate parameter-rotation speed under different load currents.

[0060] Figure 5 (a) is a schematic diagram of the relationship curve of the hysteresis operator-rotation angle under different beta values and delta values.

[0061] Figure 5 (b) is a schematic diagram of the relationship curve of the hysteresis operator-rotation speed under different beta values and delta values.

[0062] Figure 6 The figure is a schematic diagram of the relationship curve of the c value-current.

[0063] Figure 7 The figure is a schematic diagram of the relationship curve of the alpha value-current.

[0064] Figure 8 The figure is a schematic diagram of the relationship curve of the beta value-current.

[0065] Figure 9 The figure is a schematic diagram of the relationship curve of the delta value-current.

[0066] Figure 10 The figure is a schematic diagram of the comparison curve of the simulation value and the test value torque-rotation angle curve of the magneto-rheological torsional vibration damper model under different current loads, the frequency is 1Hz, and the amplitude is 10°.

[0067] Figure 11 The simulation value and the test value torque-angle curve comparison curve schematic diagram of the different current loaded magneto-rheological torsional vibration damper model under the frequency of 0.5Hz and the amplitude of 10°.

[0068] Figure 12 The comparison schematic diagram of the success rate and the average quality of the solution of the three parameter identification methods.

[0069] Figure 13 The comparison schematic diagram of the average fitness curve and the running time of the three parameter identification methods. DETAILED DESCRIPTION

[0070] The application will be further described in detail below, so that those skilled in the art can implement the application according to the description.

[0071] The application provides a method for identifying parameters of a hyperbolic tangent model of a magneto-rheological torsional vibration damper, which comprises the following steps:

[0072] Step 1: collecting corresponding relative speeds (rotational speeds), relative angular displacement inputs (angles of rotation) and torsional torques through different sensors;

[0073] In the embodiment, the mechanical properties of the magneto-rheological torsional vibration damper are tested by using a servo fatigue machine, and under the condition of sinusoidal cyclic loading, the loading frequency can be selected as 0.5Hz and 1Hz, the loading current can be selected as 0A, 1.0A, 2.0A, 3.0A and 4.0A, and the amplitude is 10°, as shown in Figure 1 The torque-angle relationship curve under the condition of different currents and the loading frequency of 1Hz and the amplitude of 10° is shown in Figure 2 The torque-speed relationship curve under the condition of different currents and the loading frequency of 1Hz and the amplitude of 10° is shown in, and it can be seen that the damping torque increases with the increase of the current, and the magneto-rheological torsional vibration damper has obvious mechanical hysteresis characteristics.

[0074] Step 2: constructing a hyperbolic tangent model of the magneto-rheological torsional vibration damper

[0075]

[0076] In the formula, T is the torsional torque, c is the viscous damping coefficient (N·s·m / rad) of the magneto-rheological torsional vibration damper, k is the stiffness coefficient (N·m) of the magneto-rheological torsional vibration damper, θ(t) is the relative angular displacement input (°) (i.e., the amplitude), t is the time, α is the expansion coefficient (N·m) of the hysteresis function, z(θ(t)) is the hysteresis operator, is the relative speed (rad / s)

[0077] In the formula, the hysteresis operator satisfies:

[0078]

[0079] In the formula, β is a coefficient (dimensionless) determined by the characteristics of the magnetorheological fluid and the structure of the magnetorheological torsional vibration damper, and δ is a coefficient (dimensionless) determined by the magnetorheological fluid and the control system, and sign(·) is a sign function;

[0080] The relative angular displacement input satisfies:

[0081] θ(t) = Asin(2πt);

[0082] In the formula, A is the amplitude (A>0) of the excitation, and the unit is °;

[0083] Step three, in order to accurately capture the hysteresis response of the magnetorheological torsional vibration damper through the hyperbolic tangent model table, five parameters ∩ = [c, k, α, β, δ] corresponding functions should be determined, specifically:

[0084] 1. Identification of k value:

[0085] Because the stiffness coefficient k describes the stiffness characteristics of the magnetorheological torsional vibration damper, which is independent of the characteristics of the magnetorheological fluid and is not affected by the current, the k value does not need to be identified under different currents. When the current is 0A, two sinusoidal cyclic loads are carried out, the relative angular displacement input is θ(t), the changed relative angular displacement input is θ'(t), and the changed relative angular displacement input satisfies:

[0086] θ'(t) = θ(t) + q;

[0087] In the formula, q is a constant;

[0088] When θ(t)>0, sign(θ) = 1, Therefore, it can be known that:

[0089] z(θ(t)) = z(θ'(t));

[0090] Therefore, it can be known that:

[0091]

[0092] Further, the k value satisfies:

[0093]

[0094] 2. Identification of c value:

[0095] According to the identified stiffness coefficient k value, the hyperbolic tangent model of the magnetorheological torsional vibration damper is converted:

[0096] ​​

[0097] where y is an intermediate parameter, thus the curve,

[0098] The value of c is determined by the overall trend of the data using the hyperbolic tangent model, as shown in Figure 3 , Figure 4 The curve of the hysteresis loop Z has a parallel region, i.e. the Figure 3 circle region, and the points on the curve in this region can be considered parallel to the X-axis. As can be seen from the formula for the intermediate parameter, the expansion coefficient α is only an equi-proportional expansion of the Y-axis direction of the hysteresis loop Z based on the curve, and since is the tilt of the Figure 4 circle region along the X-axis direction, c can be obtained by fitting the slope of the line segment through the data points in this region, and the value of c can be expressed as:

[0099]

[0100] where is the relative velocity of the i-th data point in the curve, N is the number of test data points, y i is the response value of the i-th data point in the curve, is the average value of the relative velocity in the parallel region (circle region) of the curve, y avg is the average value of the response value in the parallel region of the curve.

[0101] 3. Identification of the values of α, β and δ:

[0102] First, determine the range of values of the three parameters, α is the expansion coefficient of the hysteresis function, i.e. α > 0.

[0103] The values of β and δ have a decisive effect on the mechanical hysteresis characteristics of the MR torsional vibration damper, and the two parameters directly affect the shape of the hysteresis loop Z, as shown in Figure 5 , the orientation of the hysteresis loop Z is defined as the direction in which the protruding part in the z > 0 region points, as can be seen from Figure 5 (a), the sign of the parameter δ affects the orientation of the z-θ(t) curve, when δ > 0, the z-θ(t) curve is oriented to the right; when δ < 0, the z-θ(t) curve is oriented to the left; as can be seen from Figure 5 (b), the sign of the parameter β affects the orientation of the curve, when β > 0, the curve is oriented to the right; when β < 0, The curve points to the left, thus allowing us to determine the signs of the parameters β and δ.

[0104] The range of values ​​for parameters β and δ satisfies:

[0105]

[0106] In this embodiment, as Figures 1-3 As shown, both β and δ values ​​are positive; therefore, the range of parameter values ​​can be narrowed down to...

[0107] After determining the range of values ​​for α, β, and δ, the values ​​for α, β, and δ are then determined. In this embodiment, a genetic algorithm is used to identify the values ​​for α, β, and δ, specifically including the following steps:

[0108] Step a: Create an initial population by loading the corresponding relative velocity, relative angular displacement, and torsional torque;

[0109] Step b: Define the fitness evaluation criteria, set the initial population size to 400, configure the crossover probability to pc = 0.8, define the gene mutation probability to pm = 0.2, and limit the maximum number of iterations of the algorithm to X = 500.

[0110] Step c: Randomly initialize the values ​​of α, β, and δ;

[0111] Step d: Initialize a set of chromosomes for each parameter;

[0112] Step e: Evaluate each chromosome using the fitness function value and compare it with the previous one. The fitness function expression is:

[0113]

[0114] In the formula, T i exp Let T be the true torsional moment at the i-th data point. i sim Calculate the torsional moment for the i-th data point;

[0115] Step f: Based on the fitness assessment results of each chromosome, a roulette wheel selection strategy is adopted, in which the probability of each chromosome being selected is proportional to its fitness, thereby forming a new population.

[0116] Step g: Update chromosomes through crossover and mutation behaviors;

[0117] Step h, evaluate the new chromosome and check if the stop criterion is met, wherein the smaller the value of the fitness function, the better, if the number of iterations is higher than the limit number of iterations, report the best chromosome as the solution and output the optimal model parameters.

[0118] wherein, during the genetic algorithm running process, the constraint conditions composed of linear inequalities or equations also need to be met, and the specific expression forms of these conditions are

[0119] C x≤b;

[0120] LB≤x≤UB;

[0121] wherein, C is an inequality constraint matrix, b is an inequality constraint vector, x is a variable, LB is the lower bound of the variable, and UB is the upper bound of the variable;

[0122] substitute α>0, β>0, δ>0 and into which to obtain:

[0123] LB=[0;0;0];

[0124]

[0125] b=5.4;

[0126] Step four, according to the identified different loads, fit the change function of ∩=[c, k, α, β, δ] under multiple groups of ∩=[c, k, α, β, δ] data;

[0127] As shown in Figures 6-9 , regarding the relationship between the parameters of the magneto-rheological torsional vibration damper and the applied current, under the conditions of constant strain amplitude and frequency, increasing the current applied to the magneto-rheological torsional vibration damper will cause its key model parameters (c, α, δ) to increase significantly, and this growth trend can be approximately described by a third-order polynomial model. Meanwhile, another parameter β also changes with the current, but its change is relatively gentle and is suitable for fitting by a second-order polynomial model. Under the condition of relative angular displacement input of 10° and frequency of 1 Hz, the parameter values under five different current inputs of 0A, 1.0A, 2.0A, 3.0A, 4.0A and 5.0A are tested, and thus the change functions of each parameter are fitted as follows:

[0128] c=0.1166+1.0806I-0.1424I 2 +0.006I 3 ;

[0129] k=0.0055;

[0130] α=0.5538+0.4176I+0.608I 2 -0.0745I3 ;

[0131] δ=1.2894+1.4238I-0.521I 2 +0.063I 3 ;

[0132] β = 6.5832 + 4.1262I - 0.9494I 2 .

[0133] Verify the above parameter change function:

[0134] like Figure 10 , Figure 11 As shown, the simulated and experimental torque-rotation angle curves and torque-velocity curves of magnetorheological torsional vibration damper models with different current loading are presented at a frequency of 1Hz and an amplitude of 10° and at a frequency of 0.5Hz and an amplitude of 10°, respectively. The simulated values ​​and experimental values ​​of the models are in good agreement, and the torque-rotation angle curve does not have abrupt changes in the zero-point region, showing excessive smoothness. This proves that the identification method described in this invention accurately identifies the hyperbolic tangent model parameters. By proposing a method for calculating the identification range of hyperbolic tangent model parameters, the problem of abrupt changes in the torque-rotation angle curve in previous identifications is reduced.

[0135] The success rate of an optimization algorithm is defined as follows: successful identification means that the absolute error between the best solution and the optimal solution found by the fitness optimization algorithm is less than 10%. The success rate is the ratio of the number of successful identifications to the number of experiments.

[0136]

[0137] Where n1 represents the number of successful recognitions and n represents the number of experiments.

[0138] The quality of an optimization algorithm's solution is defined as how close the best solution found by the algorithm is to the optimal value.

[0139]

[0140] Where m is the fitness function value that achieves the optimal solution, is the minimum fitness function value for each current condition under 500 parameter identifications, and x is... ′ The fitness function value for each experiment;

[0141] Therefore, the average mass of the solution is:

[0142]

[0143] In the formula, ρ i Let n be the quality of the solution in the i-th experiment of the algorithm, and n be the number of experiments.

[0144] like Figure 12 ,Figure 13 As shown, the parameter identification method of the present application is compared with the simultaneous identification of all parameters using only genetic algorithm or particle swarm algorithm. With three excitation currents (I=0A; I=2A; I=4A), the average quality of the solution and the identification success rate of the present application are obviously improved relative to the genetic algorithm and the particle swarm algorithm. The average fitness curve of 30 independent running iterations 300 times shows that the genetic algorithm fails to obtain the best fitness value at the end of the simulation, and is still updating the fitness function at the 200th to 300th iteration. Both the particle swarm algorithm and the identification method of the present application can obtain the best fitness value at the end of the simulation. The identification method of the present application converges to the final value the fastest, and the calculation time of the identification method of the present application is the shortest, proving that the efficiency of the identification method of the present application in identifying the hyperbolic tangent model is better than that of simply using the genetic algorithm or the particle swarm.

[0145] Although the embodiments of the present application have been disclosed as above, it is not limited to the application listed in the specification and the embodiments, and can be fully applied to various fields suitable for the present application. Those skilled in the art can easily make further modifications, and therefore the present application is not limited to specific details and the embodiments shown and described herein, without departing from the general concept defined by the claims and the equivalent scope.

Claims

1. A method for identifying parameters of a hyperbolic tangent model of a magneto-rheological torsional vibration damper, characterized by The method comprises the following steps: Step 1: collecting corresponding relative speed, relative angular displacement input and torsional moment; Step 2: constructing a hyperbolic tangent model of the magneto-rheological torsional vibration damper; In the formula, T is a torsional moment, c is a viscous damping coefficient of the magneto-rheological torsional vibration damper, k is a stiffness coefficient of the magneto-rheological torsional vibration damper, θ(t) is a relative angular displacement input, t is a time, α is an amplification coefficient of a hysteresis function, z(θ(t)) is a hysteresis operator, is a relative velocity, β is a coefficient determined by magneto-rheological fluid characteristics and a structure of the magneto-rheological torsional vibration damper, δ is a coefficient determined by the magneto-rheological fluid and a control system, sign(·) is a sign function, and tanh(·) is a hyperbolic tangent function. Step 3: obtaining constraint conditions of viscous damping coefficient c of the magneto-rheological torsional vibration damper, stiffness coefficient k of the magneto-rheological torsional vibration damper, expansion coefficient α of a hysteresis function, coefficient β determined by magneto-rheological fluid characteristics and magneto-rheological torsional vibration damper structure, and coefficient δ determined by the magneto-rheological fluid and a control system according to the hyperbolic tangent model of the magneto-rheological torsional vibration damper; α>0； where T' is the changed relative angular displacement input corresponding to the torsional moment, q is a constant, is y the relative velocity of the i-th data point in the graph, N is the number of test data points, y i is the response value of the i-th data point in the graph, is the average value of the relative velocity of the parallel region in the graph, y avg is the average value of the response value of the parallel region in the graph, A is the amplitude of the excitation; Step 4: fitting a change function of ∩=[c, k, α, β, δ] according to the identified multiple groups of ∩=[c, k, α, β, δ] data under different loads; The viscous damping coefficient of the magneto-rheological torsional vibration damper, the stiffness coefficient of the magneto-rheological torsional vibration damper, the expansion coefficient of the hysteresis function, the coefficient determined by the magneto-rheological fluid characteristics and the magneto-rheological torsional vibration damper structure, and the coefficient determined by the magneto-rheological fluid and the control system satisfy: c = 0.1166 + 1.0806I - 0.1424I 2 + 0.006I 3 ; k＝0.0055； a = 0.5538 + 0.4176I + 0.608I 2 -0.0745I 3 ; δ = 1.2894 + 1.4238I - 0.5211 2 + 0.0631 3 ; β = 6.5832 + 4.1262I - 0.9494I 2 ; In the formula, I is a current.

2. The method for identifying hyperbolic tangent model parameters of a magneto-rheological torsion vibration damper according to claim 1, characterized in that, The relative angular displacement input satisfies: θ(t)=Asin(2πt); In the formula, A is an amplitude of excitation, and A>0.

3. The method for identifying the parameters of the hyperbolic tangent model of the magneto-rheological torsion vibration damper according to claim 2, characterized in that, The torsional moment corresponding to the changed relative angular displacement input satisfies: In the formula, θ′(t) is the changed relative angular displacement input.

4. The method for identifying hyperbolic tangent model parameters of a magneto-rheological torsion vibration damper according to claim 3, characterized in that, The changed relative angular displacement input satisfies: θ′(t)=θ(t)+q; In the formula, q is a constant.

5. The method for identifying hyperbolic tangent model parameters of a magneto-rheological torsion vibration damper according to claim 4, characterized in that The expansion coefficient of the hysteresis function, the coefficient determined by the magneto-rheological fluid characteristics and the magneto-rheological torsional vibration damper structure, and the coefficient determined by the magneto-rheological fluid and the control system are all identified through a genetic algorithm.

6. The method for identifying hyperbolic tangent model parameters of a magneto-rheological torsion vibration damper according to claim 5, characterized in that, The genetic algorithm comprises: Step a: creating an initial population with corresponding relative speed, relative angular displacement input and torsional moment; Step b: defining an adaptability evaluation criterion, and setting an initial population size, a probability of cross occurrence, a probability of gene mutation and a maximum iteration round; Step c: randomly initializing the expansion coefficient of the hysteresis function, the coefficient determined by the magneto-rheological fluid characteristics and the magneto-rheological torsional vibration damper structure, and the coefficient determined by the magneto-rheological fluid and the control system; Step d: initializing a group of chromosomes for each parameter; Step e: evaluating each chromosome by using a fitness function value; Step f: according to the adaptability evaluation results of each chromosome, adopting a roulette wheel selection strategy to form a new population, wherein the probability of each chromosome being selected is proportional to its adaptability; Step g: updating the chromosomes through cross and mutation behaviors; Step h: when the value of the fitness function converges or reaches the iteration number, outputting optimal model parameters.

7. The method for identifying hyperbolic tangent model parameters of a magneto-rheological torsion vibration damper according to claim 6, characterized in that The fitness function is: where T i exp T is the true torsional moment for the i-th data point i sim T is the calculated torsional moment for the i-th data point.

8. The method for identifying hyperbolic tangent model parameters of a magneto-rheological torsion vibration damper according to claim 7, characterized in that, The initial population size is 400; The probability of cross occurrence is 0.8; The probability of gene mutation is 0.2; The maximum iteration round is 500.

Images