Method for testing dynamic rigidity and damping characteristics of engine support

Abstract

The invention relates to the technical field of mechanical vibration testing, in particular to a method for testing dynamic rigidity and damping characteristics of an engine support, which comprises the following steps of: 1, simulating a boundary; step 2, double-source excitation loading is carried out; step 3, dynamic response acquisition: arranging vibration measurement points in the main shaft direction of the rigidity of the support to acquire acceleration signals in three directions, synchronously acquiring excitation force signals, and recording all the signals at a set sampling rate after anti-aliasing filtering; 4, constructing a frequency response matrix: performing time-frequency transformation on the exciting force signal and the acceleration signal, and calculating a cross-point frequency response function matrix; 5, parameter decoupling calculation is carried out, wherein parameter decoupling is achieved through cross iterative optimization; and 6, outputting parameters. Through the decoupling calculation of the low frequency band and the high frequency band, the cross iteration optimization method can effectively reduce the calculation error, improves the parameter decoupling precision, and guarantees the reliability of the test result under different frequency bands.

Description

Technical Field

[0001] The present invention relates to the technical field of mechanical vibration testing, and in particular to a method for testing the dynamic stiffness and damping characteristics of an engine bracket. Background Art

[0002] In the new energy vehicle sector, engine mounts are key load-bearing components of the powertrain, and their dynamic stiffness and damping characteristics directly impact the vehicle's NVH performance. With the advancement of hybrid technology, engine mounts must withstand the dual loads of engine mechanical vibration and motor electromagnetic excitation, posing new challenges to dynamic parameter testing. Existing testing methods suffer from the following insurmountable drawbacks: Traditional dynamic testing uses a single hydraulic vibrator for sinusoidal frequency sweeps, with an effective frequency band typically limited to 5-200 Hz, effectively simulating only engine inertial force excitation. However, the frequency band for high-frequency electromagnetic excitation of motors can reach 200-2000 Hz (e.g., the 48th harmonic of a permanent magnet synchronous motor). When using an electromagnetic vibrator with an extended high-frequency response, the excitation energy rapidly decays above 200 Hz (measured attenuation >60%) due to the impedance mismatch between the mechanical structure and the electromagnetic system. This makes it impossible to simultaneously reproduce the combined operating conditions of wide-band engine vibration and high-frequency electromagnetic forces of the motor using a single excitation source, resulting in significant deviations between test results and actual vehicle conditions.

[0003] In transfer function testing, the stiffness and damping matrices must be solved simultaneously using frequency response functions. However, under high-load hybrid powertrain conditions, the bracket damping exhibits strong nonlinear characteristics: viscous damping dominates in the low-frequency range (<100Hz), while structural damping dominates in the high-frequency range (>200Hz). Existing methods treat the stiffness and damping matrices as constant matrices for overall fitting, leading to two typical problems: The low-frequency damping term interferes with stiffness identification, causing the dynamic stiffness value to be 10%-15% higher. The high-frequency stiffness term affects damping identification, causing the calculated damping ratio to fluctuate by more than ±20%. This coupling error is particularly significant under transient conditions such as rapid acceleration / energy recovery of new energy vehicles.

[0004] Therefore, there is an urgent need for a test method for the dynamic stiffness and damping characteristics of an engine bracket to solve the above problems. Summary of the Invention

[0005] Based on the above objectives, the present invention provides a method for testing the dynamic stiffness and damping characteristics of an engine mount, comprising: Step 1: Boundary simulation: The engine bracket specimen is mounted on a rigid base platform using a contoured fixture. The fixture contact surface is shaped to match the mounting surface of the actual vehicle. An air-floating non-contact support unit is installed between the fixture and the platform. The airbag gas pressure is adjusted to simulate the actual vehicle bolt preload. Step 2: Dual-source excitation loading: Connect a hydraulic servo vibrator and an electromagnetic high-frequency vibrator in parallel at the power input end of the bracket, and connect them to the engine mounting point and the motor mounting point respectively. Control the hydraulic vibrator to output a broadband excitation covering the engine vibration frequency band, and the electromagnetic vibrator to output a high-frequency excitation covering the motor electromagnetic force frequency band. Step 3: Dynamic response acquisition: Vibration measurement points are arranged along the main axis of the bracket stiffness to collect three-axis acceleration signals and excitation force signals simultaneously. All signals are recorded at the set sampling rate after anti-aliasing filtering. Step 4: Frequency response matrix construction: Perform time-frequency transformation on the excitation force signal and acceleration signal, and calculate the cross-point frequency response function matrix; Step 5: Parameter decoupling calculation: In the low-frequency band, the hydraulic excitation data is used to identify the stiffness matrix. In the high-frequency band, the stiffness matrix is ​​fixed to identify the damping matrix. Parameter decoupling is achieved through cross-iterative optimization. Step 6: Parameter output: Extract the stiffness matrix elements to output the dynamic stiffness value, and calculate the equivalent viscous damping ratio based on the damping matrix.

[0006] Preferably, the process of adjusting the airbag gas pressure in step 1 includes: The preload range under different working conditions is obtained by using a real vehicle bolt preload measurement device, which includes a strain sensor array. The sensors are attached to the bolt surface and connected to a dynamic strain gauge. The target pressure per unit area is calculated based on the contact area between the profiling fixture and the airbag. The contact area is calculated by fitting the point cloud data of the fixture contact surface obtained by a 3D optical scanner. A closed-loop pressure control system is used: the proportional valve dynamically adjusts the intake flow rate according to the deviation between the actual value measured by the pressure sensor and the target value, and locks the pressure when the deviation value is less than the set threshold.

[0007] Preferably, the generation of broadband excitation in step 2 includes: Establish an engine speed-torque-inertia force mapping model: Calculate the inertia force amplitude envelope at various speeds through crankshaft dynamics simulation; The 5-200 Hz frequency band is divided into multiple sub-intervals, and the width of the sub-intervals decreases as the speed increases; A pseudo-random excitation signal is generated in each sub-interval, and its root mean square value is scaled according to the inertia force ratio of the envelope corresponding interval. The sweep rate is positively correlated with the speed change rate of the engine acceleration condition.

[0008] Preferably, the configuration of high frequency excitation in step 2 includes: Obtain the electromagnetic force spectrum characteristics of the motor: extract the fundamental wave and characteristic harmonic frequency distribution through finite element electromagnetic field simulation; Select multiple target frequency bands within the range of 200-2000 Hz, where the center frequencies of the target frequency bands are integer multiples of the characteristic harmonic frequencies; A fixed-amplitude sinusoidal sweep signal is generated in each target frequency band. The sweep step is adaptively adjusted according to the bandwidth, and the signal amplitude is distributed according to the proportion of electromagnetic force harmonic energy.

[0009] Preferably, the process of constructing the frequency response function matrix in step 4 includes: The time domain signal is processed by windowing and segmentation. The window function type is automatically selected according to the signal stability: when the short-term energy fluctuation of the signal exceeds the threshold, the suppression leakage window is used, otherwise the high-resolution window is used; Calculate the cross-power spectrum density matrix of the acceleration signal and the exciting force signal in each segment, as well as the auto-power spectrum density matrix of the exciting force signal; The frequency response function matrix is ​​obtained by dividing the cross-power spectrum density matrix by the auto-power spectrum density matrix, and the amplitude of the abnormal frequency points is smoothed.

[0010] Preferably, the cross-iterative optimization in step 5 includes: Set the initial conditions for iteration: the damping matrix in the low-frequency band is set to zero matrix, and the stiffness matrix in the high-frequency band is set to the previous identification value; Execute parameter transfer: substitute the damping matrix output in the high-frequency band into the low-frequency band stiffness identification equation, and substitute the updated stiffness matrix into the high-frequency band damping identification equation; Iteration termination judgment: Calculate the norm change rate of the stiffness matrix and damping matrix twice adjacently. When both are less than the dynamic convergence threshold, terminate the optimization. The convergence threshold is calculated based on the nonlinear error of the force sensor and the phase error of the vibrometer.

[0011] Preferably, the calculation process of the dynamic convergence threshold includes: Obtain the linearity error limit of the force measurement channel and the phase offset angle of the acceleration measurement channel through the sensor calibration certificate; Convert the force signal error limit into the relative error of the frequency response function amplitude, and convert the phase offset angle into the absolute error of the frequency response function phase; The theoretical error upper limit of the stiffness matrix elements is calculated based on the error propagation theory, and a preset ratio of the theoretical error upper limit is taken as the norm change rate threshold.

[0012] Preferably, the abnormal frequency determination rule is: Mark the frequency points where the coherence coefficient is lower than the first threshold on the frequency response function curve; Mark the frequency points where the cross power spectrum density amplitude suddenly drops and exceeds the second threshold; The frequency response function values ​​of the marked frequency points are corrected by cubic spline interpolation.

[0013] Preferably, the calculation of the equivalent viscous damping ratio in step 6 includes: Perform eigenvalue decomposition on the stiffness matrix to obtain the modal stiffness of each order; Project the damping matrix to the modal coordinate system to obtain the diagonalized modal damping matrix; Extract the diagonal elements of the modal damping matrix as the damping coefficients of each order mode; The damping ratio is obtained by dividing the modal damping coefficient by the product of twice the modal stiffness and the geometric mean of the modal mass, where the modal mass is obtained by normalizing the mode shape.

[0014] Preferably, the modal vibration shape normalization process includes: Select the reference degree of freedom in the anti-resonance point interval of the frequency response function; Scale all mode shapes by the mode component of the reference degree of freedom as 1; The modal mass matrix is ​​inversely calculated using the mass normalization coefficients.

[0015] Beneficial effects of the present invention: 1. The present invention connects a hydraulic servo vibrator and an electromagnetic high-frequency vibrator in parallel at the power input end of the bracket, covering the engine vibration frequency band and the motor electromagnetic force frequency band respectively, so that the composite excitation load of the engine and the motor can be simulated simultaneously, successfully solving the problem of being unable to reproduce double loads, and the test results are closer to the working conditions in actual applications.

[0016] 2. The present invention optimizes the excitation source configuration by refining the spectral characteristics of the motor's electromagnetic force excitation, designs multiple target frequency bands for the characteristic harmonic frequencies of the motor's electromagnetic excitation, and adaptively adjusts the sweep step size and amplitude according to the frequency band characteristics, so that high-frequency excitation can play an effective role, overcoming the problem of high-frequency excitation energy attenuation in existing test methods.

[0017] 3. The present invention adopts a cross-iteration optimization method to identify the low-frequency stiffness matrix and the high-frequency damping matrix respectively, and decouples the mutual influence of stiffness and damping through cross-iteration, so that the calculation results of the stiffness and damping matrices are more accurate, avoiding the generation of coupling errors, especially under transient conditions such as rapid acceleration and energy recovery of new energy vehicles, the test results are more accurate.

[0018] 4. This invention utilizes a polynomial window function and an adaptive window function selection strategy. Through windowing, segmentation, and signal stationarity analysis, it optimizes the frequency response function matrix construction process. This improvement effectively improves the accuracy of the frequency response function, making it particularly stable in the high-frequency band and avoiding the amplitude drops and abnormal frequency errors that occur in existing methods.

[0019] 5. This invention incorporates error propagation theory into the optimization process. Based on sensor calibration certificates, the theoretical upper bounds of the error for the force and acceleration signals are calculated and used as the norm rate of change threshold to optimize the convergence criteria for the parameter decoupling process. By precisely controlling error propagation, the reliability of the parameter decoupling process is ensured, improving the accuracy and consistency of test results. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0021] Figure 1 is a flow chart of the steps of the method of the present invention; Figure 2 This is a flowchart of the steps for calculating the equivalent viscous damping ratio in step 6 of the method of the present invention; Figure 3 This is a flow chart of the steps for modal vibration shape normalization processing of the method of the present invention. DETAILED DESCRIPTION

[0022] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It is also noted that, to provide a more detailed description, the following embodiments are best and preferred embodiments, and those skilled in the art may employ alternative methods for implementing certain known technologies. Furthermore, the accompanying drawings are intended only to provide a more detailed description of the embodiments and are not intended to limit the present invention.

[0023] See Figure 1-Figure 3 , an embodiment of the present invention provides a method for testing the dynamic stiffness and damping characteristics of an engine bracket. In step 1, first, the engine bracket specimen is installed on a rigid base platform through a contoured fixture. The contact surface shape of the contoured fixture is consistent with the installation surface of the actual vehicle to ensure that it matches the actual working conditions. An air-floating non-contact support unit is set between the fixture and the platform, and the pre-tightening force of the bolts of the actual vehicle is simulated by adjusting the airbag gas pressure. This design can avoid the rigid constraint problem in the traditional installation method, so that the bracket can more realistically reflect its dynamic response in actual use during the test process.

[0024] The air-floating support unit reduces the friction between the bracket and the platform, making the bracket free from external interference during the vibration test and maintaining the free vibration state of the specimen, thereby more accurately simulating the power transmission under actual working conditions.

[0025] In step 2, a hydraulic servo vibrator and an electromagnetic high-frequency vibrator are connected in parallel at the power input, connected to the engine and motor mounting points, respectively. The hydraulic vibrator's broadband excitation output covers the engine's vibration frequency range, while the electromagnetic vibrator provides high-frequency excitation that covers the motor's electromagnetic force frequency range. This dual-source excitation scheme simultaneously simulates the loads exerted by both the engine and motor on the bracket, effectively reproducing the complex excitation environment found in hybrid power systems.

[0026] Dual-source excitation loading overcomes the limitation of traditional single excitation source that cannot simulate engine and motor vibration simultaneously, making the test more in line with actual working conditions. Especially for the high-frequency electromagnetic excitation part, it can ensure the integrity of the frequency range and avoid the problem of excitation energy attenuation.

[0027] In step 3, vibration measurement points are placed along the principal axis of the bracket's stiffness to collect acceleration signals in three directions, along with the excitation force signal. All collected signals undergo anti-aliasing filtering and are recorded at the set sampling rate. This multi-channel, multi-dimensional data acquisition method ensures the integrity and accuracy of the test signals.

[0028] This step ensures comprehensive acquisition of the dynamic response of the bracket from different directions, provides accurate acceleration and excitation force data, and ensures the accuracy of subsequent frequency response matrix and parameter decoupling calculations.

[0029] In step 4, the excitation force and acceleration signals are transformed into time-frequency signals to calculate the cross-point frequency response function matrix. This time-frequency analysis can be used to obtain dynamic response characteristics at different frequencies, providing a reliable foundation for subsequent parameter decoupling and analysis.

[0030] Time-frequency transformation technology can effectively distinguish the dynamic characteristics of different frequency bands, improve the accuracy of the frequency response matrix, and avoid confusion and errors in traditional frequency response methods.

[0031] In step 5, the stiffness matrix is ​​identified in the low-frequency range using data from the hydraulic vibrator; in the high-frequency range, the damping matrix is ​​identified by fixing the stiffness matrix. Through a cross-iterative optimization method, the interaction between the stiffness and damping matrices is decoupled, resulting in more accurate calculation results.

[0032] The cross-iterative optimization technology effectively solves the interference of low-frequency damping terms on stiffness identification and the influence of high-frequency stiffness terms on damping identification, avoids the coupling error in traditional methods, and improves the identification accuracy of stiffness and damping.

[0033] Finally, in step 6, the dynamic stiffness values ​​are extracted from the stiffness matrix elements, and the equivalent viscous damping ratio is calculated based on the damping matrix. This approach allows accurate dynamic stiffness and damping characteristic parameters to be obtained, providing data support for further vehicle performance optimization and design.

[0034] This step provides precise dynamic stiffness and damping characteristic data, providing automotive engineers with more accurate basic data to optimize the design of engine brackets and improve the NVH performance of new energy vehicles.

[0035] By combining innovative technologies such as air bearing support, dual-source excitation, time-frequency analysis, and parameter decoupling, the method effectively solves many technical difficulties encountered in existing methods in testing the dynamic stiffness and damping characteristics of new energy vehicle engine brackets, improves the accuracy and reliability of the test, and provides strong technical support for the power system design and performance optimization of new energy vehicles.

[0036] In one possible implementation, a real-vehicle bolt preload measurement device is first used to determine the bolt preload range under different operating conditions. This device comprises an array of strain sensors attached to the bolt surface, which collects strain data in real time using dynamic strain gauges. By sensing minute bolt deformations under varying loads, the strain sensor array measures and provides real-time feedback on the bolt preload. This process enables accurate bolt strain data under varying loads to be obtained, allowing the preload range to be calculated.

[0037] This step ensures the accurate measurement of the bolt preload, avoids the measurement errors that may exist in traditional methods, and provides reliable reference data for subsequent boundary simulations.

[0038] The target pressure per unit area is calculated based on the contact area between the profiling fixture and the airbag. This contact area is calculated by fitting point cloud data of the profiling fixture's contact surface acquired using a 3D optical scanner. This point cloud data comprehensively captures subtle geometric variations in the fixture's contact surface. Based on this data, the target pressure required for the airbag can be more accurately calculated, simulating the distribution of bolt preload under actual operating conditions.

[0039] Using a 3D optical scanner to acquire point cloud data can determine the contact surface shape and contact area with high precision, avoiding the simplified assumptions of complex surface contact conditions in traditional measurement methods and ensuring the accuracy of the pressure target value.

[0040] To adjust the airbag gas pressure, a closed-loop pressure control system is employed. This system consists of a proportional valve, a pressure sensor, and a target pressure value. The proportional valve dynamically adjusts the intake air flow based on the deviation between the real-time measured pressure and the target value, thereby precisely controlling the airbag pressure. When the deviation falls below a set threshold, the control system automatically locks the airbag pressure to maintain a stable pressure state.

[0041] In this embodiment of the present invention, a threshold of ±0.1% is set during the test. This means that when the deviation between the real-time pressure measured by the pressure sensor and the target value is less than 0.1%, the closed-loop system automatically stops adjusting the intake flow rate, ensuring that the airbag pressure remains stable at the desired target value. This low threshold ensures sufficient airbag pressure accuracy, thereby ensuring the accuracy of the experimental results.

[0042] The closed-loop pressure control system can achieve precise adjustment and real-time control of the airbag pressure, avoiding the errors caused by manual adjustment in traditional methods, and ensuring the stability of the airbag pressure during the test, thereby improving the repeatability and reliability of the experiment.

[0043] By combining multiple technologies, including a strain sensor array, a 3D optical scanner, and a closed-loop pressure control system, high-precision regulation of the airbag pressure was achieved, ensuring the faithful reproduction of boundary conditions during engine mount testing. This not only improved test accuracy but also provided reliable data support for subsequent dynamic analysis. Furthermore, by setting appropriate pressure deviation thresholds, refined pressure control was ensured, further enhancing test stability and reliability.

[0044] In one possible implementation, a model mapping engine speed, torque, and inertial force is first established. Specifically, crankshaft dynamics simulation is used to calculate the amplitude envelope of the inertial force generated by the engine at different speeds. Crankshaft dynamics simulation primarily integrates factors such as the engine's rotational inertia, cylinder forces, and combustion processes during operation to determine the engine's dynamic response at various speeds.

[0045] Using simulation tools such as Matlab / Simulink, modeling is performed based on the structure and dynamic characteristics of the engine, the inertia force under various speed conditions is simulated, and the envelope of the inertia force amplitude is plotted.

[0046] By establishing a high-precision simulation model, we can accurately capture the changing trend of the engine inertia force at different speeds, provide data support for generating broadband excitation signals, and ensure the rationality and repeatability of the excitation signals.

[0047] The 5-200Hz frequency band is divided into multiple subranges, with the subrange width decreasing as the engine speed increases. This design allows for accurate simulation of the engine's vibration characteristics at varying speeds. At low speeds, the inertial force changes more gradually, allowing for a wider frequency subrange. However, at high speeds, the inertial force changes rapidly, requiring a finer frequency division.

[0048] By dividing the frequency band into segments and adjusting the width of each sub-interval based on the rate of change of the inertial force in each segment, for example, at low speeds, the frequency range can be divided into 10Hz sub-intervals, while at high speeds, the frequency range can be divided into 1Hz sub-intervals, allowing for more precise control of frequency changes.

[0049] This frequency band division method can effectively improve the frequency resolution of the excitation signal, especially under high-speed conditions, and can better track and reflect the rapid changes in the engine inertia force, thereby more realistically simulating the dynamic characteristics of the engine during operation.

[0050] Within each frequency subinterval, a pseudo-random excitation signal is generated, with its RMS value scaled by the inertial force ratio within the corresponding envelope interval. A pseudo-random excitation signal exhibits random characteristics within a specific frequency range. It can simulate the vibration response of an engine at different frequencies and is used to test the dynamic stiffness and damping characteristics of an engine mount.

[0051] Pseudo-random excitation signals are usually generated through digital signal processing technology, using methods such as normal distribution or Gaussian distribution to generate signals, and their root mean square values ​​are adjusted by scaling to make them consistent with the proportion of inertial force.

[0052] This excitation signal can produce uniform energy distribution within a wide frequency band, avoiding the problem of insufficient energy in the low or high frequency parts of traditional excitation signals, thereby improving the comprehensiveness and accuracy of the test.

[0053] Finally, the sweep rate is positively correlated with the rate of change of engine speed during acceleration. That is, as engine speed changes, the sweep rate adjusts accordingly. This ensures that the excitation signal changes are synchronized with the actual engine operating conditions, thereby more realistically simulating the dynamic response of the engine during acceleration.

[0054] When generating the excitation signal, the engine's acceleration condition data (such as the speed change rate) is collected and the sweep rate is adjusted so that the frequency change of the excitation signal matches the engine speed change.

[0055] This adjustment can ensure that the frequency change of the excitation signal during the test is consistent with the actual engine operating conditions, improve the reliability and accuracy of the test data, and make the test results more representative.

[0056] Through the above steps, the broadband excitation generation process accurately simulates the inertial force distribution and dynamic response of the engine under different speed conditions. Establishing a precise speed-torque-inertial force mapping model, rationally dividing the frequency bands, and generating a uniformly distributed pseudo-random excitation signal ensures the accuracy of the excitation signal across the entire frequency band. Combined with a positive correlation between the frequency sweep rate and the speed change rate, this method can better simulate the dynamic behavior of the engine under actual operating conditions. This method offers high precision, repeatability, and broad application prospects, providing a scientific and reliable excitation method for dynamic performance testing of engine mounts.

[0057] In one possible implementation, finite element electromagnetic field simulation is first performed to extract the electromagnetic force spectrum characteristics of the motor under different operating conditions, specifically the fundamental frequency distribution and characteristic harmonic frequency distribution. Finite element simulation accurately calculates the electromagnetic force distribution within the motor by taking into account factors such as the motor's geometry, material properties, and current distribution.

[0058] Use finite element simulation tools such as COMSOL Multiphysics or ANSYS Maxwell to simulate the electromagnetic force of the motor. By analyzing each frequency point of the motor, the fundamental wave of the electromagnetic force and its characteristic harmonic frequencies are extracted, and the intensity and distribution of the harmonic frequencies are determined.

[0059] This simulation method accurately captures the electromagnetic force spectrum characteristics of the motor, providing a scientific basis for configuring high-frequency excitation signals. By accurately extracting the fundamental and characteristic harmonic frequencies, it can better capture the motor's vibration patterns under actual operating conditions.

[0060] Within the 200-2000Hz frequency range, multiple target frequency bands are selected, with their center frequencies being integer multiples of the characteristic harmonic frequencies. This design ensures a close match between the excitation signal and the electromagnetic force characteristics of the motor, resulting in test results that are more consistent with actual conditions.

[0061] Based on the simulation results, select frequency bands close to the characteristic harmonic frequency and determine the center frequency of each band. For example, if the characteristic harmonic frequency is 300Hz, the center frequency of the target frequency band might be an integer multiple of 300Hz, 600Hz, 900Hz, or other frequencies. In this case, the target frequency band can be set to a 25Hz width for more detailed frequency division.

[0062] By selecting an integer multiple of the characteristic harmonic frequency as the center of the target frequency band, we can ensure that the excitation signal is synchronized with the motor's natural vibration frequency, effectively improving the accuracy and representativeness of the test. This method avoids deviation from the motor's resonant frequency and reduces test errors.

[0063] A fixed-amplitude swept sine signal is generated within each target frequency band, and the sweep step size is adaptively adjusted based on the bandwidth. The amplitude of the fixed-amplitude swept sine signal is distributed based on the proportion of electromagnetic force harmonic energy, ensuring that the amplitude of the excitation signal is consistent with the energy distribution of the electromagnetic force spectrum.

[0064] Based on the width of each target frequency band and the energy distribution of the electromagnetic force spectrum, a sinusoidal sweep signal is generated through digital signal processing. The sweep step size is adaptively adjusted based on the frequency band width to ensure uniform signal frequency distribution across the entire frequency band. The amplitude is adjusted based on the energy contribution of the characteristic harmonics to ensure that the signal amplitude reflects the energy contribution of the electromagnetic force within the frequency band.

[0065] A constant-amplitude swept sine signal provides uniform frequency variation within each frequency band, accurately stimulating the dynamic response of the engine mount. Adaptive adjustment of the sweep step size and amplitude distribution allows precise control of the signal's excitation intensity, avoiding over- or under-excitation and ensuring test signal accuracy and reliability.

[0066] This technical feature enables precise configuration of high-frequency excitation signals by extracting the electromagnetic force spectrum characteristics of the motor through finite element simulation, selecting integer multiples of the characteristic harmonic frequency band, and generating a constant-amplitude sinusoidal sweep signal. Specifically, the electromagnetic force spectrum characteristics are used to determine the frequency range and amplitude of the excitation signal, ensuring a close match between the excitation signal and the vibration characteristics of the motor. Adaptive adjustment of the sweep step size and signal amplitude further improves test accuracy, making the dynamic stiffness and damping characteristics of the engine mount more realistic and effective.

[0067] In one possible implementation, during signal processing, the time domain signal is first divided into multiple segments for analysis, a process known as windowing and segmentation processing, and an appropriate window function is automatically selected based on the stationarity of each segment.

[0068] The time domain signal is first divided into several small segments. Common window functions include rectangular windows, Hanning windows, and Blackman windows. When analyzing signal stationarity, short-term energy fluctuations are used to determine the window function type. If the signal's short-term energy fluctuations exceed a certain threshold, indicating significant signal variation within that segment, a leakage suppression window (such as a Blackman window or Hanning window) is selected to reduce edge effects and leakage. If the signal's short-term energy fluctuations are small, indicating a relatively stationary signal, a high-resolution window (such as a rectangular window) is selected to provide higher frequency resolution.

[0069] In this embodiment of the present invention, a threshold is set such that when the signal's standard deviation (or volatility) exceeds 0.1, a leakage suppression window is selected; if the standard deviation is less than 0.1, a high-resolution window is selected. This approach ensures that a window function suitable for leakage suppression is used for highly variable signal portions, while a high-resolution window is used for stationary signals to improve frequency accuracy.

[0070] This method can flexibly adjust the window function type according to the characteristics of different signals, avoiding spectrum leakage or resolution loss caused by window function mismatch, thereby improving the accuracy of spectrum analysis.

[0071] The signal in each segment needs to calculate the cross-power spectrum density matrix of the acceleration signal and the exciting force signal, as well as the auto-power spectrum density matrix of the exciting force signal.

[0072] By performing a Fast Fourier Transform (FFT) on the acceleration and excitation force signals, the power spectrum density at each frequency point is calculated. The cross-power spectrum density reflects the correlation between the acceleration and excitation force signals, while the auto-power spectrum density reflects the energy distribution of the excitation force signal.

[0073] Calculating the cross-power spectral density (CPSD) provides information about the interaction between the acceleration and excitation force signals, while the auto-power spectral density (APSD) describes the characteristics of the excitation force signal itself. Combining these two methods allows for a more accurate analysis of the system's response characteristics.

[0074] After obtaining the cross-power spectrum density matrix and the auto-power spectrum density matrix, the cross-power spectrum density matrix is ​​divided by the auto-power spectrum density matrix to obtain the frequency response function matrix.

[0075] The frequency response function matrix describes the dynamic relationship between the system's input and output, specifically the frequency response between the excitation force and acceleration. When performing division operations, ensure that the matrix dimensions match and account for any numerical errors. The frequency response function matrix provides information about the system's dynamic stiffness and damping at different frequencies.

[0076] The frequency response function matrix is ​​an important tool for analyzing the dynamic characteristics of a system. It can help accurately identify key parameters such as stiffness, damping, and resonant frequency. This is crucial for the design and optimization of engine mounts.

[0077] The abnormal frequency points in the calculated frequency response function matrix are smoothed to further eliminate noise and data deviation.

[0078] Smoothing the amplitudes of extreme values ​​or sudden frequency points in the frequency response function matrix. Common methods include moving average and Gaussian smoothing. These methods can effectively remove the impact of noise on the frequency response function matrix, making it more consistent with the dynamic response characteristics of the actual system.

[0079] Through smoothing, data fluctuations caused by the test environment or signal noise can be eliminated, thereby improving the reliability and accuracy of the frequency response function matrix and ensuring that the test results can truly reflect the dynamic stiffness and damping characteristics of the engine mount.

[0080] Through windowing and segmentation, calculation of cross- and auto-power spectral densities, construction of a frequency response function matrix, and amplitude smoothing, this method effectively extracts and analyzes the dynamic response characteristics of an engine mount. Each step is designed to improve the accuracy and reliability of signal analysis, particularly the automated selection of window functions and calculation of power spectral density and frequency response function matrices, resulting in more realistic and detailed test results.

[0081] In a possible implementation, in the initial stage of the cross-iterative optimization, it is necessary to set a suitable starting value to ensure that the optimization process can proceed smoothly.

[0082] The damping matrix for the low-frequency band is set to zero. This is because the dynamic response in this band is generally relatively stable, with minimal damping effects. Setting this to zero simplifies the initial calculation and avoids affecting subsequent optimization. The stiffness matrix for the high-frequency band is set to the value from the previous identification. This is because the stiffness in this band has been accurately estimated in the previous identification, and serving as the initial value provides a reasonable starting point for subsequent iterations.

[0083] Reasonable initial conditions can help avoid convergence problems during the iteration process, especially in complex systems. By reasonably setting the initial conditions of the low-frequency and high-frequency bands, the optimization process can be accelerated and the waste of computing resources can be reduced.

[0084] During the cross-iterative optimization process, the dynamic stiffness and damping characteristics of the low-frequency band and the high-frequency band affect each other, so parameter transfer is required.

[0085] Substitute the damping matrix output from the high-frequency band into the stiffness identification equation for the low-frequency band. Based on the current damping matrix, the low-frequency band stiffness matrix is ​​updated. Simultaneously, the updated low-frequency band stiffness matrix is ​​substituted into the damping identification equation for the high-frequency band to update the high-frequency band damping matrix. Through this parameter transfer method, the parameters of the two frequency bands are mutually constrained and optimized, resulting in more accurate dynamic characteristics.

[0086] Parameter transfer allows the dynamic characteristics of the low- and high-frequency bands to influence and constrain each other. This cross-optimization method avoids local optimal solutions and ensures comprehensive optimization of the dynamic stiffness and damping characteristics of the entire system. By continuously exchanging and updating parameters, the characteristics of each frequency band of the system can be accurately identified.

[0087] The final step of cross-iteration optimization is to determine whether to terminate the optimization. The optimization process determines whether it has converged by calculating the rate of change of the norm.

[0088] In each iteration, the norm change rate of the stiffness and damping matrices is calculated between two adjacent iterations. This norm change rate reflects the magnitude of the matrix update. When both rates of change are less than the set dynamic convergence threshold, the optimization process is considered to have converged and the optimization process is terminated. The dynamic convergence threshold is calculated by combining the nonlinear error of the force sensor and the phase error of the vibrometer. Specifically, the nonlinear error of the force sensor can cause signal deviation, while the phase error of the vibrometer can affect the accuracy of the frequency response. Therefore, these errors need to be taken into account to determine a reasonable convergence threshold.

[0089] Through precise convergence determination, the optimization process automatically terminates after reaching sufficient accuracy, avoiding over-computation and waste of computing resources. Furthermore, a convergence threshold based on sensor error ensures that the optimization results are sufficiently accurate and reliable in actual testing.

[0090] The cross-iterative optimization method can not only improve the accuracy of dynamic stiffness and damping characteristics identification through mutual optimization of low-frequency bands and high-frequency bands and accurate convergence judgment, but also effectively eliminate the influence of sensor errors and instrument noise on test results.

[0091] In one possible implementation, when calculating the dynamic convergence threshold, it is first necessary to obtain relevant error information from the sensor calibration certificate, specifically including: Linearity error limit of the force measurement channel: This refers to the linearity error of the force sensor's output signal over the entire measuring range. This error directly affects the accuracy of the force signal and, in turn, the dynamic stiffness identification results.

[0092] Phase offset angle of the acceleration measurement channel: This error represents the phase difference between the accelerometer output signal and the actual motion state. Phase offset can distort the phase information in the frequency response function, affecting the damping characteristics test.

[0093] After obtaining the sensor error information, these errors need to be converted into the error form of the frequency response function for subsequent error propagation and calculation: The linearity error limit of the force signal reflects the amplitude deviation of the force signal at different frequencies. Through conversion, the relative error of the amplitude in the frequency response function can be obtained. This error directly affects the accuracy of the stiffness matrix.

[0094] The phase offset angle of the acceleration measurement channel will introduce a phase error in the frequency response function. After converting it into an absolute error, the impact of the phase deviation on the test results can be accurately quantified.

[0095] Error propagation theory allows the amplitude error of the force signal and the phase error of the acceleration signal to be propagated into the stiffness matrix calculation. Based on the error sources (force and acceleration sensor errors) and the relationship between them, error propagation theory can calculate a theoretical upper bound on the error of the stiffness matrix elements. This upper bound reflects the worst-case impact of sensor errors on each element of the stiffness matrix.

[0096] By considering the contribution of force and acceleration signals to each element of the stiffness matrix in the frequency response function and combining the error propagation formula, the upper limit of the error of each stiffness matrix element is calculated.

[0097] This step ensures the rationality and accuracy of the error propagation process, can effectively evaluate the impact of sensor errors on the overall stiffness identification accuracy of the system, and provide a scientific basis for subsequent optimization.

[0098] Finally, based on the theoretical upper limit of the stiffness matrix error obtained from error propagation theory, a reasonable preset ratio is selected as the norm change rate threshold. This ratio is usually set based on experimental experience, application requirements, or the error range tolerated by the system.

[0099] The theoretical error limit is multiplied by the preset ratio to obtain the final norm change rate threshold. The norm change rate threshold directly determines when the optimization process ends. It ensures that the optimization process stops when the system's error change rate falls below the threshold, ensuring efficient use of computing resources.

[0100] By setting a reasonable threshold, the error transmission can be effectively controlled, over-correction can be avoided during the optimization process, and a balance between computational efficiency and optimization accuracy can be ensured.

[0101] Through the above steps, the dynamic convergence threshold calculation process fully considers the impact of sensor errors and accurately calculates the upper limit of the stiffness matrix error through error propagation theory. This process ensures a more scientific and accurate convergence judgment during the optimization process, avoids over-optimization or unnecessary calculations caused by errors, and thus improves optimization efficiency.

[0102] In one possible implementation, the coherence coefficient measures the correlation between two signals. In frequency response function testing, the coherence coefficient is typically calculated based on the relationship between the input signal and the response signal. A low coherence coefficient indicates a weak relationship between the system input and response, possibly due to noise or measurement errors.

[0103] In the frequency response function curve, a coherence coefficient value is calculated for each frequency point, which represents the correlation between the input signal and the output signal. If the coherence coefficient of a frequency point is lower than a preset first threshold (for example, 0.6), the frequency point is marked as an abnormal frequency point.

[0104] In the embodiment of the present invention, the first threshold is set to 0.6. If the coherence coefficient of a certain frequency point is 0.55, it is considered that there is a large noise or interference at the frequency point, and thus it is marked as an abnormal frequency point.

[0105] The cross-power spectral density (CPSD) describes the frequency response between the input and output signals. If the CPSD amplitude suddenly drops at certain frequencies, this could be due to a device failure, sensor issues, or environmental interference.

[0106] By calculating the cross-power spectrum density of the frequency response function and detecting its amplitude changes, if the amplitude of a certain frequency point suddenly drops by more than the second threshold (for example, the amplitude suddenly drops by more than 20dB), the frequency point is marked as an abnormal frequency point.

[0107] Specifically, if the amplitude of the cross-power spectrum density drops by more than 20 dB at a certain frequency (from 10 dB to -10 dB), the frequency is marked as an abnormal frequency because such a sudden drop in amplitude may be caused by a problem with the measurement equipment or nonlinear behavior of the system.

[0108] For frequency points marked as abnormal, interpolation correction is required to ensure the smoothness and continuity of the frequency response function data. Cubic spline interpolation is a commonly used smooth interpolation method. It corrects the data of abnormal frequency points by constructing a continuous curve with smooth derivatives between adjacent frequency points.

[0109] For frequency points marked as abnormal, the frequency response function values ​​in their neighborhood are corrected using cubic spline interpolation. Cubic spline interpolation ensures that the interpolated curve is not only continuous at the frequency point but also has a smooth derivative, avoiding data abrupt changes.

[0110] Through interpolation correction, the errors caused by abnormal frequency points can be eliminated, the smoothness and continuity of the frequency response function curve can be ensured, and the accuracy of the test results can be further improved.

[0111] The abnormal frequency determination rules can effectively ensure the reliability and accuracy of test results, guarantee the accuracy of dynamic stiffness and damping characteristic tests, and further improve the quality of performance analysis of mechanical systems such as engine mounts.

[0112] In one possible implementation, the system's stiffness matrix must first be obtained, typically through finite element analysis or experimental data. Then, by performing eigenvalue decomposition on the stiffness matrix, the modal stiffnesses of each order can be obtained. This step decomposes the complex structural stiffness characteristics into multiple simplified, independent modal stiffness values, allowing the dynamic response of each mode to be analyzed separately.

[0113] The core of eigenvalue decomposition is to mathematically solve the eigenvalues ​​and eigenvectors of the stiffness matrix. The eigenvalues ​​represent the stiffness of each mode, while the eigenvectors represent the modal shape of the system, that is, the displacement distribution of each node under that mode.

[0114] In modal analysis, the damping matrix is ​​usually not diagonal, so it is necessary to transform it through the modal coordinate system and project the damping matrix into this coordinate system. This operation can simplify the originally complex damping matrix into a diagonal modal damping matrix, where the diagonal elements are the damping coefficients for each mode.

[0115] The damping matrix of the system is projected through the modal transformation matrix to obtain the damping coefficient corresponding to each mode. These damping coefficients can reflect the damping characteristics of the system in each mode, thus providing the necessary data for subsequent calculations.

[0116] In the modal coordinate system, the modal damping matrix becomes a diagonal matrix, with the diagonal elements representing the damping coefficients for each mode. The damping coefficient is an important parameter that describes the strength of the damping effect and is closely related to the energy loss characteristics of the structure.

[0117] The damping coefficient of each mode is extracted from the diagonalized modal damping matrix to form a damping coefficient vector. These coefficients are the basis for calculating the equivalent viscous damping ratio.

[0118] Calculating the damping ratio of each mode requires the modal damping coefficient, modal stiffness, and modal mass. Modal stiffness is obtained through eigenvalue decomposition, while modal mass is obtained through modal shape normalization.

[0119] The modal mass is usually obtained by normalization, which ensures that the mass contributions of different modes are equivalent. Then, the damping ratio is calculated by the following formula: ; This formula combines the modal damping coefficient with the modal stiffness and modal mass to calculate the equivalent viscous damping ratio under this mode.

[0120] Normalization of modal shapes is done to make the masses of different modes comparable. In practical applications, modal shape normalization is usually performed based on the mass or displacement amplitude of the modal shape to ensure that the mass of each mode is calculated under the same standard.

[0121] Mass normalization is usually used, that is, normalizing the modal mass vector so that the total mass is 1, thereby simplifying the calculation process and eliminating the influence of mass differences between different modes on the calculation.

[0122] In one possible implementation, a frequency response function (FRF) describes the response characteristics of a system at different frequencies. During modal analysis, the FRF plot reveals the system's resonance and antiresonance points. The antiresonance point is the region where the system's frequency response is minimal, meaning it's less likely to resonate. By selecting the reference degrees of freedom within the antiresonance region, errors introduced by resonance effects can be avoided.

[0123] First, obtain the system's frequency response function through experimentation or calculation and identify the antiresonance point. Then, select a reasonable degree of freedom (for example, the vibration degree of freedom of a node in the system) as the reference degree of freedom, which will serve as the basis for subsequent modal normalization.

[0124] After selecting a reference degree of freedom, the modal shapes of the entire system can be scaled according to the modal components of that degree of freedom. This operation can unify all modal shapes to the base scale of the reference degree of freedom, ensuring that the modal shapes of different modes are compared at the same scale.

[0125] The modal components of the reference degree of freedom are used as a scale, and the modal components of the other modes are scaled according to the size of their modal components, so that the modal components of all modes are compared with respect to that degree of freedom. This ensures that the modal components of all modes have a consistent scale in a physical sense, avoiding errors caused by scale differences.

[0126] The mass normalization coefficient is derived from the reference degrees of freedom and scaling operations in the modal shape normalization process and reflects the mass characteristics of each mode. The modal mass matrix describes the relationship between each mode and the system mass distribution. It is typically a diagonal matrix in which each element represents the mass of the corresponding mode.

[0127] The mass normalization coefficient is usually obtained as follows: Normalize the vibration mode of each mode. Calculate the relationship between the vibration mode components of the reference degree of freedom and their corresponding modal masses. Based on this relationship, inversely calculate the mass of each mode using a formula.

[0128] The key to inversely calculating the modal mass matrix is ​​to determine the mass of each mode through the relationship between the modal vibration shape and the mass normalization coefficient. The calculation formula is: ; in, is the modal mass matrix, It is The first mode shape, is the mass normalization coefficient of the corresponding mode.

[0129] The mass normalization factor is an important parameter in modal analysis, used to normalize each mode of a system with respect to the actual mass distribution. It makes the mass characteristics of different modes comparable, thereby improving the accuracy of modal analysis.

[0130] The mass normalization factor is a numerical value that describes the relationship between the mass of the modal shape and the actual mass distribution. It is typically obtained through a normalization process, which allows the modal mass matrix of the system to conform to the actual physical constraints.

[0131] The mass normalization coefficient ensures the uniform standard of modal mass and avoids the mass error caused by the difference in modal vibration shape scale, thereby improving the accuracy and reliability of the modal mass matrix.

[0132] The following is a detailed explanation using examples: Modern engine mount designs must meet stringent dynamic performance requirements to ensure they effectively absorb and isolate vibrations during operation. In practical applications, traditional dynamic characteristic testing methods (such as direct measurement and empirical methods) often suffer from low data accuracy and complex calculations. Therefore, this paper proposes a dynamic characteristic testing method based on modal mass normalization and mode shape normalization, which can accurately evaluate the dynamic stiffness and damping characteristics of engine mounts.

[0133] This embodiment aims to verify whether the dynamic stiffness and damping characteristics of the engine bracket can be accurately evaluated through modal analysis and mass normalization technology, thereby providing a simple and efficient dynamic testing method.

[0134] In this embodiment of the present invention, a vibration test bench is used to apply a vibration signal and collect the system response. An accelerometer is used to measure the acceleration response of the bracket at different frequencies. A force sensor is used to measure the external force applied to the bracket. A data acquisition system is used to collect the frequency response function.

[0135] Engine bracket, model: ES-001, material is aluminum alloy.

[0136] Test frequency range: 0-500Hz (determined by the operating frequency of the engine mount).

[0137] Modal mass normalization coefficient: The reference degree of freedom is selected as the maximum vibration point of the bracket.

[0138] Quality normalization formula: ; in, is the mass of a single mode, is the mass of all modes, is the number of modes.

[0139] Select the main vibration mode as the reference vibration mode and normalize it using the following formula: ; in, It is The vibration shape of a mode.

[0140] A vibration test bench is used to apply excitation forces of different frequencies to the engine bracket and collect the corresponding acceleration response signals.

[0141] The frequency response function (FRF) is obtained by using the data acquisition system. By analyzing the frequency response function, the anti-resonance point range is identified and the modal characteristics within a specific frequency range are determined.

[0142] The maximum vibration point of the engine mount (usually the middle of the mount) is selected as the reference degree of freedom.

[0143] According to the vibration mode obtained by the frequency response function, the main vibration mode is selected for vibration mode normalization. Assume that the value of the vibration mode is: ; After mode normalization: ; Calculate the mass normalization factor using the formula above: ; in, is the mass sum of all modes. Assume that in the first mode, the mass , in the second mode, the mass , then the total mass is: ; The normalized quality coefficient is calculated.

[0144] Calculate the modal mass matrix: ; Fill the mass of each mode into the matrix to obtain the modal mass matrix.

[0145] The dynamic stiffness and damping characteristics of the system are calculated using the modal mass matrix and frequency response function. Assuming that the frequency of a certain mode is f=100Hz, the dynamic stiffness calculation formula is: ; According to the calculation, the dynamic stiffness of this mode is: ; The dynamic characteristics of the engine bracket were directly measured using traditional modal analysis methods, and the dynamic stiffness obtained was 3.5×106N / m.

[0146] Through mass normalization and mode normalization analysis, the dynamic stiffness obtained is 3.94×106N / m.

[0147] By comparison, it can be seen that the method of the present invention improves the accuracy by 12.6% compared with the traditional method and can more accurately reflect the real dynamic characteristics of the engine bracket.

[0148] This example demonstrates that combining modal mass normalization with mode shape normalization can more accurately assess the dynamic stiffness and damping characteristics of an engine mount. Comparative experiments demonstrate that this method offers significant accuracy advantages over traditional methods, providing more reliable data support for engine mount design and performance optimization.

[0149] The present invention encompasses any alternatives, modifications, equivalents, and solutions that fall within the spirit and scope of the present invention. To provide a thorough understanding of the present invention, specific details are described in detail below in connection with the preferred embodiments of the present invention, but those skilled in the art will be able to fully understand the present invention without these detailed descriptions. Furthermore, to avoid unnecessary confusion regarding the essence of the present invention, well-known methods, processes, procedures, components, and circuits have not been described in detail.

[0150] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for testing the dynamic stiffness and damping characteristics of an engine mount, characterized in that: include: Step 1: Boundary simulation: The engine bracket specimen is mounted on a rigid base platform using a contoured fixture. The fixture contact surface is shaped to match the mounting surface of the actual vehicle. An air-floating non-contact support unit is installed between the fixture and the platform. The airbag gas pressure is adjusted to simulate the actual vehicle bolt preload. Step 2: Dual-source excitation loading: Connect a hydraulic servo vibrator and an electromagnetic high-frequency vibrator in parallel at the power input end of the bracket, and connect them to the engine mounting point and the motor mounting point respectively. Control the hydraulic vibrator to output a broadband excitation covering the engine vibration frequency band, and the electromagnetic vibrator to output a high-frequency excitation covering the motor electromagnetic force frequency band. Step 3: Dynamic response acquisition: Vibration measurement points are arranged along the main axis of the bracket stiffness to collect three-axis acceleration signals and excitation force signals simultaneously. All signals are recorded at the set sampling rate after anti-aliasing filtering. Step 4: Frequency response matrix construction: Perform time-frequency transformation on the excitation force signal and acceleration signal, and calculate the cross-point frequency response function matrix; Step 5: Parameter decoupling calculation: In the low-frequency band, the hydraulic excitation data is used to identify the stiffness matrix. In the high-frequency band, the stiffness matrix is ​​fixed to identify the damping matrix. Parameter decoupling is achieved through cross-iterative optimization. Step 6: Parameter output: Extract the stiffness matrix elements to output the dynamic stiffness value, and calculate the equivalent viscous damping ratio based on the damping matrix.

2. The method for testing the dynamic stiffness and damping characteristics of an engine mount according to claim 1, characterized in that: The process of adjusting the airbag gas pressure in step 1 includes: The preload range under different working conditions is obtained by using a real vehicle bolt preload measurement device, which includes a strain sensor array. The sensors are attached to the bolt surface and connected to a dynamic strain gauge. The target pressure per unit area is calculated based on the contact area between the profiling fixture and the airbag. The contact area is calculated by fitting the point cloud data of the fixture contact surface obtained by a 3D optical scanner. A closed-loop pressure control system is used: the proportional valve dynamically adjusts the intake flow rate according to the deviation between the actual value measured by the pressure sensor and the target value, and locks the pressure when the deviation value is less than the set threshold.

3. The method for testing the dynamic stiffness and damping characteristics of an engine mount according to claim 1, characterized in that: The generation of broadband excitation in step 2 includes: Establish an engine speed-torque-inertia force mapping model: Calculate the inertia force amplitude envelope at various speeds through crankshaft dynamics simulation; The 5-200 Hz frequency band is divided into multiple sub-intervals, and the width of the sub-intervals decreases as the speed increases; A pseudo-random excitation signal is generated in each sub-interval, and its root mean square value is scaled according to the inertia force ratio of the envelope corresponding interval. The sweep rate is positively correlated with the speed change rate of the engine acceleration condition.

4. The method for testing the dynamic stiffness and damping characteristics of an engine mount according to claim 1, characterized in that: The configuration of high frequency excitation in step 2 includes: Obtain the electromagnetic force spectrum characteristics of the motor: extract the fundamental wave and characteristic harmonic frequency distribution through finite element electromagnetic field simulation; Select multiple target frequency bands within the range of 200-2000 Hz, where the center frequencies of the target frequency bands are integer multiples of the characteristic harmonic frequencies; A fixed-amplitude sinusoidal sweep signal is generated in each target frequency band. The sweep step is adaptively adjusted according to the bandwidth, and the signal amplitude is distributed according to the proportion of electromagnetic force harmonic energy.

5. The method for testing the dynamic stiffness and damping characteristics of an engine mount according to claim 1, characterized in that: The process of constructing the frequency response function matrix in step 4 includes: The time domain signal is processed by windowing and segmentation. The window function type is automatically selected according to the signal stability: when the short-term energy fluctuation of the signal exceeds the threshold, the suppression leakage window is used, otherwise the high-resolution window is used; Calculate the cross-power spectrum density matrix of the acceleration signal and the exciting force signal in each segment, as well as the auto-power spectrum density matrix of the exciting force signal; The frequency response function matrix is ​​obtained by dividing the cross-power spectrum density matrix by the auto-power spectrum density matrix, and the amplitude of the abnormal frequency points is smoothed.

6. The method for testing the dynamic stiffness and damping characteristics of an engine mount according to claim 1, characterized in that: The cross-iterative optimization in step 5 includes: Set the initial conditions for iteration: the damping matrix in the low-frequency band is set to zero matrix, and the stiffness matrix in the high-frequency band is set to the previous identification value; Execute parameter transfer: substitute the damping matrix output in the high-frequency band into the low-frequency band stiffness identification equation, and substitute the updated stiffness matrix into the high-frequency band damping identification equation; Iteration termination judgment: Calculate the norm change rate of the stiffness matrix and damping matrix twice adjacently. When both are less than the dynamic convergence threshold, terminate the optimization. The convergence threshold is calculated based on the nonlinear error of the force sensor and the phase error of the vibrometer.

7. The method for testing the dynamic stiffness and damping characteristics of an engine mount according to claim 6, characterized in that: The calculation process of the dynamic convergence threshold includes: Obtain the linearity error limit of the force measurement channel and the phase offset angle of the acceleration measurement channel through the sensor calibration certificate; Convert the force signal error limit into the relative error of the frequency response function amplitude, and convert the phase offset angle into the absolute error of the frequency response function phase; The theoretical error upper limit of the stiffness matrix elements is calculated based on the error propagation theory, and a preset ratio of the theoretical error upper limit is taken as the norm change rate threshold.

8. The method for testing the dynamic stiffness and damping characteristics of an engine mount according to claim 5, characterized in that: The determination rule of the abnormal frequency point is: Mark the frequency points where the coherence coefficient is lower than the first threshold on the frequency response function curve; Mark the frequency points where the cross power spectrum density amplitude suddenly drops and exceeds the second threshold; The frequency response function values ​​of the marked frequency points are corrected by cubic spline interpolation.

9. The method for testing the dynamic stiffness and damping characteristics of an engine mount according to claim 1, characterized in that: The calculation of the equivalent viscous damping ratio in step 6 includes: Perform eigenvalue decomposition on the stiffness matrix to obtain the modal stiffness of each order; Project the damping matrix to the modal coordinate system to obtain the diagonalized modal damping matrix; Extract the diagonal elements of the modal damping matrix as the damping coefficients of each order mode; The damping ratio is obtained by dividing the modal damping coefficient by the product of twice the modal stiffness and the geometric mean of the modal mass, where the modal mass is obtained by normalizing the mode shape.

10. The method for testing the dynamic stiffness and damping characteristics of an engine mount according to claim 9, characterized in that: The modal vibration shape normalization process includes: Select the reference degree of freedom in the anti-resonance point interval of the frequency response function; Scale all mode shapes by the mode component of the reference degree of freedom as 1; The modal mass matrix is ​​inversely calculated using the mass normalization coefficients.

Images