[0001]Technical Field The invention relates to the technical field of decoupling rate and inherent frequency robustness optimization of a vehicle power assembly suspension system, in particular to a suspension system inherent characteristic optimization method considering the inertia parameter error of a power assembly. Background Art Method for manufacturing a rubber suspension in a vehicle powertrain suspension system During installation and use, the rigidity value, the installation position and the angle can be influenced by uncertain factors, and the deviation or fluctuation of a certain degree can be avoided. Furthermore, the rigidity is suspended. In the optimization process of the position and the angle, it is generally assumed that the center of mass of the power assembly, the rotational inertia, and the measured value of the inertial product are precise, that is, no changes or fluctuations occur. At present, an optimization method for decoupling rate and natural frequency of a suspension system is mainly based on a traditional deterministic optimization model and constraint conditions. The traditional deterministic optimization method does not consider the influence of uncertain factors mentioned above, which leads to lower robustness of the optimization result and a large optimization failure risk. Based on research results of uncertain analysis and optimization methods by the inventor, a robustness optimization method considering the decoupling rate and the natural frequency of the suspension system considering the mass center position, the moment of inertia and the inertia product test error is proposed. The suspension rigidity, the installation position, the installation angle and the mass center position of the power assembly are assumed, the rotational inertia and the inertial product obey normal distribution, the decoupling rate and the natural frequency of the suspension system are optimally optimized through an advanced deterministic global optimization method, and an effective way and analysis method are provided for solving the vibration isolation performance of the vehicle suspension system containing the random design parameters and the non-design parameters. Content of the invention In order to overcome the defects of a traditional method for determining the vibration isolation performance of a vehicle suspension system, the object is achieved. The implementation steps of the invention are as follows. A six-degree-of-freedom dynamic model of the power assembly suspension system is established. X axis is parallel to the crankshaft and is directed towards the flywheel end of the engine, Z-axis is perpendicular to the crankshaft upward, Y-axis direction is determined by the right-hand rule. The suspension local coordinate system is established in 3 elastic spindle directions of the suspension, each suspension being simplified into an element having rigidity and damping in its three elastic spindle directions, and the rubber suspension damping is negligible when the suspension system is shaken slightly. Step II. A determinacy optimization model of the decoupling rate and the natural frequency of the power assembly suspension system is established by taking the test values of the mass center and the rotational inertia and the inertia product of the power assembly as a deterministic non-design parameter to suspend the installation position, the installation angle and the suspension rigidity to be deterministic design parameters. A Matlab is adopted to calculate the natural frequency and decoupling rate of the power assembly, and then a global optimization algorithm, namely an adaptive mutation genetic algorithm in Isight optimization software, is adopted to solve the deterministic optimization mathematical model. Based on the energy decoupling theory of the power assembly suspension system, the mass center, the rotation inertia and the inertia product of the power assembly are non-design parameters based on the power assembly suspension system energy decoupling theory. A Matlab is adopted to calculate the natural frequency and decoupling rate of the power assembly, and then a global optimization algorithm is adopted to solve the robustness optimization model of the suspension system. A Monte Carlo analysis method is adopted to verify the superiority of the steady optimization result with respect to the deterministic optimization result. Further, when the six-degree-of-freedom dynamic model of the power assembly suspension system is established in one step, the torsional rigidity and the suspension damping factor of the suspension are ignored and the static rigidity is converted into dynamic stiffness by the dynamic and static ratio, and the 6-degree-of-freedom dynamic model of the power assembly suspension system is established in a power assembly centroid coordinate system. Further, when a deterministic optimization model of the decoupling rate and the natural frequency of the power assembly suspension system is established in the second step, a test value of the mass center and the rotational inertia and the inertia product of the power assembly is used as a determinacy non-design parameter. Further, when the steady optimal mathematical model of the decoupling rate and the natural frequency of the suspension system is established in the three steps, uncertainty non-design parameters of the power assembly are characterized by the normal distribution random variables, and the installation angle and the suspension rigidity are uncertainty design parameters. The fluctuation range around the nominal value of the power assembly is generally 1% - 3%, the fluctuation range of the suspension stiffness around the nominal value is generally 10% - 20%, the fluctuation range of the suspension mounting angle near the nominal value is generally 1% - 5% 1% - 5%, although the measurement error of each parameter is not large. In order to estimate the influence of the measurement error on the optimization effect of the suspension system, the invention establishes the following robustness optimization mathematical model. Here, μ [•] and σ [•] are mean and standard deviation, w respectivelyi And vi Weights and s for mean and standard deviation of i-th order decoupling rate respectivelyμ Making μ (d)i Normalized scaling coefficients, sσ Making σ (d)i The scaling factor is normalized. fi F (i)-th order natural frequency (F)iL And fiU Respective fi Lower bound and upper bound. Σ number required for frequency-constrained robustness. di Correspond to fi Decoupling rate’ SMD for di Lower bound. Σ number required for uncoupling rate constraint robustness. xj . yj Z-zj The parameters are designed for the installation position j of the j th suspension. sum X-xj Y-yj Z-zj Lower bound and upper bound. Each is in xj Y-yj Z-zj In the vicinity of fluctuating variables, δx Δ-δy Δ-δz Each is respectively. Fluctuation amplitude. αj . βj Gamma-γj Design parameters for the installation angle j of the j-th suspension. sum αj Beta-βj Gamma-γj Lower bound and upper bound. Each is in αj Beta-βj Gamma-γj In the vicinity of fluctuating variables, δα Δ-δβ Δ-δγ Each is respectively. Fluctuation amplitude. kXj . kYj , kZj Design parameters for the stiffness of j th suspension. sum Respectively kXj , kYj , kZj Lower bound and upper bound. Respectively in kXj , kYj , kZj In the vicinity of fluctuating variables, δKx Δ-δKy Δ-δKz Each is respectively. Fluctuation amplitude. Wave non-design parameter and G for mass center coordinates of power assemblyMx . GMy , GMz A test value for the centroid coordinates is obtained. Each is respectively. Fluctuation amplitude. Fluctuation non-design parameter and I for rotational inertia and inertia product of power assembly respectivelyMxx . IMyy I-IMzz I-IMxy I-IMyz I-IMzzx The rotational inertia and the inertia product of the power assembly are respectively tested. Each is respectively. Amplitude. Further, the global optimization algorithm is an adaptive mutation genetic algorithm (AMGA) mature in Isight optimization software adaptive mutation mutation mutation mutation genetic algorithm. Further, the method of the Monte Carlo analysis method is adopted to verify the superiority of the steady optimization result with respect to the deterministic optimization result. The centroid of the suspension system is described using normal distribution random variables. An uncertainty of inertia and inertia products and suspension installation position, installation angle and suspension rigidity is determined, Monte Carlo analysis is carried out on a robust optimization method according to a common test error and design parameter of a non-design parameter, and a robust optimization method is verified to determine whether the method is more robust. The decoupling rate and the natural frequency of the suspension system are reduced, the sensitivity of the mounting angle and the suspension rigidity is reduced, and a Monte Carlo analysis is carried out on the robustness optimization result to verify the superiority of the optimization result of the suspension system with a deterministic optimization result. Compared with the traditional optimization method, the mass center position, the moment of inertia, the inertia, the mounting position, the installation angle and the like of the power assembly are assumed to be determinative numerical values. The main advantages are as follows. (1), In view of manufacturing and mounting errors of the power assembly suspension, design parameters such as the rigidity, the mounting position, the mounting angle and the like of the suspension necessarily have a certain error due to manufacturing and mounting errors. These errors are taken into consideration in the optimization of the rigidity, the installation position, the installation angle and the like of the suspension, and the robustness optimization model of the suspension system decoupling rate and the natural frequency closer to the actual engineering is established in consideration of the test error of the mass center position, the rotational inertia and the inertia product of the power assembly. (2), The invention takes into consideration the fluctuation of the rigidity, the installation position and the installation angle in the suspension assembly and the use process, and therefore, the robustness optimization method of the suspension system decoupling rate and the natural frequency is different from the traditional deterministic optimization result. Description of drawings 1 Is a flow chart of a method for optimizing an inherent characteristic of a suspension system taking a power assembly inertia parameter error into consideration. 2 Is a schematic diagram of a powertrain suspension system. Mode of execution The invention will be further illustrated below with reference to the drawings and specific examples. As shown 1, the present invention proposes a suspension system inherent characteristic optimization method taking into account the error of the inertial parameters of the power assembly, and the specific implementation of the method is as follows. Six-freedom dynamic model of (1) built-up power assembly suspension system To FIG. 2, a power assembly centroid coordinate system (GCS) G0 - XYZ is established, the coordinate origin G0 is located at the center of mass of the power assembly, X-axis is parallel to the crankshaft and points to the engine flywheel end. Z-axis is perpendicular to the crankshaft upward, Y-axis direction is determined by a right-hand rule. Establishing a suspension local coordinate system (LCS) o in 3 elastic spindle directions of the suspensioni -ui vi wi U is used to three elastic main shafts of suspension respectivelyi V-vi And wi , The coordinate of the i-th suspension in the power assembly coordinate system is ri (X)i Y-yi Z-zi Each suspension is simplified to have stiffness and damping elements along its three elastic spindle directions, which is negligible when the system microamplitude vibrates. The kinetic energy and potential energy of the suspension system are calculated, a mass matrix and a stiffness matrix of the suspension system are obtained through a Lagrange equation, and a free vibration differential equation of the suspension system can be obtained. Here, [M]. [K] Are the total mass matrix and the total stiffness matrix of the system, respectively, {q} as the generalized coordinates of the rigid body. (2) Takes test values of the center of mass and moment of inertia and the inertia product of the power assembly as a deterministic non-design parameter to suspend the installation position, the installation angle and the suspension rigidity to be deterministic design parameters to establish a deterministic optimization model of the decoupling rate and the natural frequency of the power assembly. Matlab compiling programs are adopted to calculate the natural frequency and decoupling rate of the power assembly, and then a mature adaptive mutation genetic algorithm (AMGA) in Isight optimization software is adopted to solve the deterministic optimization mathematical model and the deterministic optimization model is. Wherein, fi D for i-th order natural frequenciesi Correspond to fi Decoupling rate, wi Weights. SMD for di Lower bound and fiL And fiU Respective fi Lower bound and upper bound, xj Y-yj Z-zj The parameters are designed for the installation position j of the j th suspension. sum X-xj Y-yj Z-zj Lower bound and upper bound, αj Beta-βj Gamma-γj Design parameters for the installation angle j of the j-th suspension. sum αj Beta-βj Gamma-γj Lower bound and upper bound, kXj , kYj , kZj Design parameters for the stiffness of j th suspension. sum Respectively kXj , kYj , kZj The lower bound and upper bound. (3) Uses normal distribution random variables to characterize the center of mass, moment of inertia and inertia product and suspension installation position, installation angle and suspension rigidity of the power assembly. The fluctuation range around the nominal value of the power assembly is generally 1% - 3%, the fluctuation range of the suspension stiffness around the nominal value is generally 10% - 20%, the fluctuation range of the suspension mounting angle near the nominal value is generally 1% - 5% 1% - 5%, although the measurement error of each parameter is not large. In order to estimate the influence of the measurement error on the optimization effect of the suspension system, the invention establishes the robustness optimization mathematical model of the suspension system. Here, μ [•] and σ [•] are mean and standard deviation, w respectivelyi And vi Weights, sμ And sσ The proportionality coefficient is a proportionality coefficient. fi F (i)-th order natural frequency (F)iL And fiU Respective fi Lower bound and upper bound. Σ number required for frequency-constrained robustness. di Correspond to fi Decoupling rate. SMD for di Lower bound. Σ number required for uncoupling rate constraint robustness. xj . yj Z-zj The parameters are designed for the installation position j of the j th suspension. sum X-xj Y-yj Z-zj Lower bound and upper bound. Each is in xj Y-yj Z-zj In the vicinity of fluctuating variables, δx Δ-δy Δ-δz Each is respectively. Fluctuation amplitude. αj . βj Gamma-γj Design parameters for the installation angle j of the j-th suspension. sum αj Beta-βj Gamma-γj Lower bound and upper bound. Each is in αj Beta-βj Gamma-γj In the vicinity of fluctuating variables, δα Δ-δβ Δ-δγ Each is respectively. Fluctuation amplitude. kXj . kYj , kZj Design parameters for the stiffness of j th suspension. sum Respectively kXj , kYj , kZj Lower bound and upper bound. Respectively in kXj , kYj , kZj In the vicinity of fluctuating variables, δKx Δ-δKy Δ-δKz Each is respectively. Fluctuation amplitude. Wave non-design parameter and G for mass center coordinates of power assemblyMx . GMy , GMz A test value for the centroid coordinates is obtained. Each is respectively. Fluctuation amplitude. Fluctuation non-design parameter and I for rotational inertia and inertia product of power assembly respectivelyMxx . IMyy I-IMzz I-IMxy I-IMyz I-IMzzx The rotational inertia and the inertia product of the power assembly are respectively tested. Each is respectively. Amplitude. (4): A Matlab is used to calculate the natural frequency and decoupling rate of the suspension system, wherein Matlab is used for calculating the natural frequency and the decoupling rate of the power assembly, and the robustness optimization mathematical model is solved by adopting an adaptive mutation genetic algorithm (AMGA) mature in Isight (ICI) in a Matlab. (5) A Monte Carlo analysis method is adopted to verify the superiority of the steady optimization result relative to the deterministic optimization result. The robustness optimization result is subjected to a Monte Carlo analysis, a contrast determinacy optimization and a robustness optimization model to optimize σ level of constraint conditions in the model, and whether the steady optimization method is more robust is verified by using a Descricricricricriticality optimization model. Assuming that the suspension seating angle fluctuates 10% - 20% in the vicinity of its nominal value, assuming that the suspension seating angle fluctuates 1% - 5% in the vicinity of its nominal value, 1% - 5% parameters of the power assembly centroid coordinate, rotational inertia and inertia product are assumed to fluctuate 9 1% - 3% in consideration of the superiority of the optimization scheme given by Monte Carlo analysis with respect to the deterministic optimization result. Embodiment: The suspension system is a matched product of a SUV vehicle type, and the specific application of the inherent characteristic optimization method of the suspension system considering the inertia parameter error of the power assembly is provided. The weight of the power assembly is 193 kg, and the mass center coordinates, the rotational inertia and the inertia product of the power assembly are shown 1 and 2. The original static stiffness, installation position and installation angle of the suspension are shown 3 to Table 5. On the basis of its initial value 30%, the suspension installation angle optimization variables fluctuate 10% on the basis of its initial value, 10%. parameters of the suspension installation angle optimization variables fluctuating 9 natural frequencies in the vicinity of their initial values are: 6-16 Hz, 6-16 Hz, 1% - 3% .6 8.5 - 9 . 5Hz, 6-16 Hz, @datdatdat@ and 6-16 Hz 10.5-11 . 5Hz. The decoupling rate requirement of Z-axis and Roll direction is greater 90%, and the decoupling rate in other directions is greater than 85%. Meter 1 power assembly centroid coordinates Rotational inertia and inertia product of meter 2 power assembly in centroid coordinate system Table 3 suspension initial static stiffness Table 4 suspension of original coordinates Meter 5 suspension mounting angle/(°) To the robustness optimization model provided by the invention, after the optimization result is obtained, Monte te te Carlo simulation test verification is carried out, the sampling method is a description sampling method, and the sampling frequency is improved to 10000 times. Table 6 gives the results of deterministic optimization. The Monte Carlo analysis result of the steady optimization result when the power assembly center of mass and moment of inertia and inertia product measurement error are not taken into account and considered when considering the center of mass and moment of inertia of the power assembly and the inertia product measurement error. Monte Carlo analysis versus table 6 certainty and robustness optimization results f1 To f6 The natural frequencies are first to 6 steps. DX. DY and dZ are uncoupling rates along X, Y and Z axes, dRRoll, dPPitch, dYYaw respectively indicate decoupling rates around X, Y and Z axes. From Table 6, it can be seen. The inherent frequency and decoupling rate of the optimization method can meet the requirements of constraint conditions. The natural frequency and the decoupling rate in the steady optimization result when considering the mass center of the power assembly and the moment of inertia and the inertia product measurement error and considering the mass center of the power assembly and inertia and the inertia product measurement error can meet the requirements of constraint conditions. However, when the standard deviation is not taken into consideration, the standard deviation of the frequency and the decoupling rate when the power assembly center of mass and the rotational inertia and the inertia product measurement error are not taken into account is smaller than the uncertainty optimization result. The result of optimization when the error is measured by the inertia product. In Table 6, the robustness of the inherent frequency and decoupling rate of the suspension system to the power assembly center of mass, rotational inertia and inertia product, suspension stiffness, suspension installation position and installation angle is reduced by the robustness optimization method compared with the conventional deterministic optimization method. The robustness optimization method in consideration of the mass center and moment of inertia of the power assembly and inertia product measurement error further reduces the flexibility of the power assembly center of mass, rotational inertia and inertia product, suspension rigidity, suspension installation position and installation angle. The simulation results show that the decoupling rate and the robustness of the suspension system corresponding to the suspension system robustness optimization method in consideration of the mass center of the power assembly and the rotational inertia and the inertia product measurement error obtained by the method are greatly improved compared with the traditional deterministic optimization method. The above is only a specific step of the present invention, and the scope of protection of the present invention is not limited to any limitation.
