Shock-absorbing device multi-working-condition simulation detection method and device

Abstract

The present application relates to the technical field of shock absorber performance detection, and particularly relates to a shock absorber multi-working condition simulation detection method. The method constructs a nonlinear mechanical characteristic model by obtaining dynamic response data, and extracts mechanical characteristic parameters to map to a model state space. A working condition-performance correlation mapping relationship is established to form a mapping matrix, based on which an adaptive detection criterion that can be optimized independently according to working conditions is constructed. Finally, performance evaluation is carried out according to the criterion, and the results are fed back to the model to realize closed-loop improvement of detection accuracy. The present application realizes real-time tracking and adaptive accurate evaluation of the performance evolution law of the shock absorber.

Description

Technical Field

[0001] This invention relates to the field of vibration damping device performance testing technology, and in particular to a multi-condition simulation testing method and system for vibration damping devices. Background Technology

[0002] Performance testing of vibration damping devices is crucial for ensuring their reliable operation under complex real-world conditions. Current technologies primarily rely on performance tests conducted under specific or standard operating conditions. The conventional approach involves applying a pre-set, relatively singular excitation signal to the damping device in a laboratory environment, such as a sinusoidal excitation with a specific amplitude and frequency, and then acquiring its response data. By performing independent time-domain or frequency-domain analysis on the response data, static or linearized characteristic parameters such as maximum displacement, transmissivity, or damping ratio are extracted and compared with preset standard thresholds to determine whether the vibration damping device's performance is up to standard.

[0003] These conventional testing methods have significant limitations. Because the test conditions are relatively fixed and idealized, the extracted characteristic parameters often only reflect the performance of the damping device under a few specific excitations, making it difficult to comprehensively capture its nonlinear mechanical behavior under varying and complex excitations in real service environments. The loads that damping devices bear in actual operation are often random, broadband, and vary in amplitude; their stiffness and damping characteristics exhibit complex nonlinear evolution depending on the excitation conditions. Linearized characteristic assessments based on finite operating conditions cannot accurately establish a complete dynamic correlation between operating condition changes and performance parameters.

[0004] Therefore, detection systems based on the above methods lack a deep understanding of the performance evolution patterns across the entire operating range. Their criteria are typically static and fixed, unable to adaptively adjust to changes in actual excitation conditions, leading to decreased detection sensitivity under non-standard conditions and resulting in missed detections or false positives. Furthermore, due to the lag in model and criterion updates, the detection process struggles to form an effective closed-loop optimization mechanism, limiting the continuous improvement of detection accuracy and reliability, and failing to meet the growing demand for precise condition assessment and early fault warning of vibration damping devices. Summary of the Invention

[0005] The present invention provides a multi-condition simulation testing method and system for shock absorption devices, which can solve the problems in the prior art.

[0006] A first aspect of the present invention provides a multi-condition simulation testing method for a vibration damping device, comprising: The dynamic response data of the vibration damping device under test under multiple working conditions is obtained. Based on the dynamic response data, a nonlinear mechanical characteristic model of the vibration damping device is constructed. By performing joint time-frequency domain analysis on the dynamic response data, mechanical characteristic parameters reflecting the stiffness and damping characteristics of the vibration damping device are extracted, and the mechanical characteristic parameters are mapped to the state space of the nonlinear mechanical characteristic model. Based on the nonlinear mechanical characteristic model, a working condition-performance correlation mapping relationship is established. By associating the excitation characteristics under different working conditions with the corresponding mechanical characteristic parameters in multiple dimensions, a mapping matrix is ​​formed to characterize the performance evolution law of the vibration damping device in the entire working condition range. Based on the mapping matrix, an adaptive detection criterion is constructed. By tracking the performance evolution pattern in the mapping matrix in real time, the detection threshold and anomaly identification boundary are dynamically adjusted, so that the adaptive detection criterion can be autonomously optimized according to changes in working conditions. The performance of the vibration damping device is evaluated based on the adaptive detection criteria, generating detection results that include performance degradation trends and fault feature identification. The detection results are then fed back to the nonlinear mechanical feature model to update the mechanical feature parameters, thereby achieving a closed-loop improvement in detection accuracy.

[0007] Based on the dynamic response data, a nonlinear mechanical characteristic model of the damping device is constructed. By performing joint time-frequency domain analysis on the dynamic response data, mechanical characteristic parameters reflecting the stiffness and damping characteristics of the damping device are extracted, including: The dynamic response data is synchronously transformed in the time and frequency domains. By extracting the transient force-displacement response trajectory in the time domain and the amplitude-frequency and phase-frequency characteristics of the dominant frequency components in the frequency domain, a complementary mapping relationship of time and frequency domain features is established. Based on the complementary mapping relationship, the force-displacement hysteresis feature reflecting nonlinear stiffness and the phase hysteresis feature reflecting nonlinear damping are identified in the time-frequency domain features. The nonlinear evolution law of stiffness is extracted by curve family fitting of the force-displacement hysteresis feature, and the frequency-dependent characteristics of damping are extracted by frequency dependence analysis of the phase hysteresis feature. The stiffness nonlinear evolution law is characterized as a non-monotonic function with respect to displacement amplitude, and the damping frequency variation characteristic is characterized as a continuously varying function with respect to excitation frequency. The non-monotonic function and the continuously varying function are embedded as coupled state variables into a system of differential equations to construct a nonlinear mechanical characteristic model.

[0008] The stiffness nonlinear evolution law is extracted by fitting a family of curves to the force-displacement hysteresis characteristics, and the damping frequency-varying characteristics are extracted by frequency-dependent analysis of the phase hysteresis characteristics, including: Multiple hysteresis loops under different displacement amplitude conditions are extracted from the force-displacement hysteresis characteristics to form a family of hysteresis curves. The envelope of each loop in the family of hysteresis curves is subjected to multi-order nonlinear fitting to obtain a continuous function expression of stiffness as a function of displacement amplitude. By performing piecewise gradient analysis on the continuous function expression, the nonlinear strengthening and nonlinear weakening characteristics of stiffness in different displacement ranges are identified. The nonlinear strengthening feature and the nonlinear weakening feature are used as state indicators of stiffness evolution. By constructing a transformation sequence of the state indicators as the displacement amplitude increases, a complete characterization chain describing the nonlinear evolution law of stiffness is formed. The phase difference sequence between force response and displacement response under different excitation frequencies is extracted from the phase hysteresis characteristics. The phase difference sequence is then processed to be continuous in the frequency domain to obtain a smooth functional relationship between the phase difference and the excitation frequency. By performing piecewise differentiation on the smooth functional relationship at different frequencies, the rate of change characteristics of the damping coefficient in different frequency bands are identified. Based on the frequency band differences of the rate of change characteristics, the frequency-varying characteristics of damping are extracted.

[0009] Based on the aforementioned nonlinear mechanical characteristic model, a working condition-performance correlation mapping relationship is established. By correlating the excitation characteristics under different working conditions with the corresponding mechanical characteristic parameters in multiple dimensions, a mapping matrix characterizing the performance evolution of the vibration damping device across the entire working condition range is formed, including: Based on the nonlinear mechanical characteristic model, the excitation amplitude characteristics, excitation frequency characteristics, and excitation waveform characteristics under different working conditions are extracted as working condition feature vectors. The corresponding stiffness characteristic parameters, damping characteristic parameters, and response peak parameters are extracted as performance feature vectors. A many-to-many correlation relationship is established between the working condition feature vectors and the performance feature vectors. By tensor expansion of the many-to-many relationship, an association matrix is ​​constructed with the working condition feature vector as the row index and the performance feature vector as the column index, and the association matrix is ​​filled with association weights that reflect the degree of influence of the working condition on the performance. The correlation matrix is ​​extended to cover the entire working condition range. By interpolating and extrapolating the correlation weights under known working conditions, a mapping matrix covering the entire working condition range is generated.

[0010] By tensor-quantizing the many-to-many relationship, an association matrix is ​​constructed with the operating condition feature vector as the row index and the performance feature vector as the column index, including: The many-to-many relationship is represented as a third-order tensor structure, and the third-order tensor structure is decomposed to extract the dominant association pattern between the working condition dimension and the performance dimension. Based on the dominant correlation mode, the third-order tensor structure is modally expanded, and the coupling strength is bidirectionally projected along the working condition dimension and the performance dimension to form a two-dimensional matrix structure. The two-dimensional matrix structure is filled with correlation weights by using the components of each dimension of the working condition feature vector as row indices and the components of each dimension of the performance feature vector as column indices, and filling the intersection of the row indices and the column indices with the corresponding correlation weights, thereby constructing the correlation matrix.

[0011] Based on the mapping matrix, an adaptive detection criterion is constructed. By tracking the performance evolution patterns in the mapping matrix in real time, the detection threshold and anomaly identification boundary are dynamically adjusted, enabling the adaptive detection criterion to autonomously optimize with changes in operating conditions. The evolution trajectory of performance parameters with changing operating conditions is extracted from the mapping matrix. Time series analysis is performed on the evolution trajectory to obtain the trend characteristics and fluctuation amplitude characteristics of the performance parameters. Based on the trend characteristics and fluctuation amplitude characteristics, a performance status evaluation benchmark is established. The performance status evaluation benchmark is used as the initial reference standard for constructing adaptive detection criteria. By performing a working condition sensitivity analysis on the performance status evaluation benchmark, the degree and direction of deviation of performance parameters under different working conditions are identified, and the adjustment amount of the detection threshold is dynamically calculated based on the degree and direction of deviation. The adjustment amount of the detection threshold is then superimposed on the performance status evaluation benchmark to form a dynamic detection threshold that adapts to the working conditions. Based on the dynamic detection threshold, by continuously comparing the real-time collected performance parameters with the dynamic detection threshold, abnormal states where the performance parameters exceed the dynamic detection threshold are identified. Based on the distribution position of the abnormal state in the mapping matrix, the corresponding working condition features are traced back in reverse, and the abnormal identification boundary is dynamically adjusted to match the current working condition, thereby realizing the autonomous optimization of the adaptive detection criterion as the working condition changes.

[0012] A second aspect of the present invention provides a multi-condition simulation testing system for vibration damping devices, comprising: The data acquisition unit is used to acquire the dynamic response data of the shock absorber under multiple working conditions, construct a nonlinear mechanical characteristic model of the shock absorber based on the dynamic response data, extract mechanical characteristic parameters reflecting the stiffness and damping characteristics of the shock absorber by performing joint time-frequency domain analysis on the dynamic response data, and map the mechanical characteristic parameters to the state space of the nonlinear mechanical characteristic model. The working condition performance unit is used to establish a working condition-performance correlation mapping relationship based on the nonlinear mechanical characteristic model. By associating the excitation characteristics under different working conditions with the corresponding mechanical characteristic parameters in multiple dimensions, a mapping matrix is ​​formed to characterize the performance evolution law of the vibration damping device in the entire working condition range. An adaptive judgment unit is used to construct an adaptive detection criterion based on the mapping matrix. By tracking the performance evolution law in the mapping matrix in real time, the detection threshold and anomaly identification boundary are dynamically adjusted so that the adaptive detection criterion can be autonomously optimized as the working conditions change. The performance evaluation unit is used to evaluate the performance of the shock absorption device based on the adaptive detection criteria, generate detection results including performance degradation trends and fault feature identification, and feed the detection results back to the nonlinear mechanical feature model to update the mechanical feature parameters to achieve a closed-loop improvement in detection accuracy.

[0013] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0014] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0015] This method enables in-depth analysis and modeling of the dynamic response data of vibration damping devices under multiple excitation conditions, effectively extracting key mechanical characteristic parameters reflecting their stiffness and damping properties, and accurately mapping them to the state space of a nonlinear mechanical characteristic model. This overcomes the shortcomings of traditional methods, such as incomplete feature extraction and insufficient model representation capabilities under complex excitations, laying a solid model foundation for subsequent accurate performance evaluation.

[0016] By establishing a working condition-performance correlation mapping relationship, the excitation characteristics of different working conditions are correlated with the corresponding mechanical characteristic parameters in multiple dimensions, forming a mapping matrix that characterizes the performance evolution law across the entire working condition range. This mapping matrix can systematically reveal the dynamic law of the vibration damping device's performance changing with working conditions, realizing a leap from single working condition evaluation to grasping the performance evolution law across all working conditions, and significantly improving the comprehensiveness and predictability of the detection.

[0017] The adaptive detection criterion based on the mapping matrix can track the performance evolution pattern in real time and dynamically adjust the detection threshold and anomaly identification boundary. This allows the detection criterion to autonomously optimize according to changes in actual working conditions, effectively solving the problems of poor adaptability and high false alarm / false negative rates of fixed threshold criteria under varying working conditions, and significantly improving the robustness and accuracy of the detection system. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the multi-condition simulation testing method for vibration damping devices. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0021] Figure 1 This is a flowchart illustrating the multi-condition simulation testing method for the vibration damping device according to an embodiment of the present invention. Figure 1 As shown, the multi-condition simulation testing method for vibration damping devices includes: The dynamic response data of the vibration damping device under test under multiple working conditions is obtained. Based on the dynamic response data, a nonlinear mechanical characteristic model of the vibration damping device is constructed. By performing joint time-frequency domain analysis on the dynamic response data, mechanical characteristic parameters reflecting the stiffness and damping characteristics of the vibration damping device are extracted, and the mechanical characteristic parameters are mapped to the state space of the nonlinear mechanical characteristic model. Based on the nonlinear mechanical characteristic model, a working condition-performance correlation mapping relationship is established. By associating the excitation characteristics under different working conditions with the corresponding mechanical characteristic parameters in multiple dimensions, a mapping matrix is ​​formed to characterize the performance evolution law of the vibration damping device in the entire working condition range. Based on the mapping matrix, an adaptive detection criterion is constructed. By tracking the performance evolution pattern in the mapping matrix in real time, the detection threshold and anomaly identification boundary are dynamically adjusted, so that the adaptive detection criterion can be autonomously optimized according to changes in working conditions. The performance of the vibration damping device is evaluated based on the adaptive detection criteria, generating detection results that include performance degradation trends and fault feature identification. The detection results are then fed back to the nonlinear mechanical feature model to update the mechanical feature parameters, thereby achieving a closed-loop improvement in detection accuracy.

[0022] In one optional implementation, a nonlinear mechanical characteristic model of the damping device is constructed based on the dynamic response data. By performing joint time-frequency domain analysis on the dynamic response data, mechanical characteristic parameters reflecting the stiffness and damping characteristics of the damping device are extracted, including: The dynamic response data is synchronously transformed in the time and frequency domains. By extracting the transient force-displacement response trajectory in the time domain and the amplitude-frequency and phase-frequency characteristics of the dominant frequency components in the frequency domain, a complementary mapping relationship of time and frequency domain features is established. Based on the complementary mapping relationship, the force-displacement hysteresis feature reflecting nonlinear stiffness and the phase hysteresis feature reflecting nonlinear damping are identified in the time-frequency domain features. The nonlinear evolution law of stiffness is extracted by curve family fitting of the force-displacement hysteresis feature, and the frequency-dependent characteristics of damping are extracted by frequency dependence analysis of the phase hysteresis feature. The stiffness nonlinear evolution law is characterized as a non-monotonic function with respect to displacement amplitude, and the damping frequency variation characteristic is characterized as a continuously varying function with respect to excitation frequency. The non-monotonic function and the continuously varying function are embedded as coupled state variables into a system of differential equations to construct a nonlinear mechanical characteristic model.

[0023] For the dynamic response data acquired from the vibration damping device under test, a synchronous time-frequency domain transformation is first performed. This transformation process combines short-time Fourier transform and wavelet transform to simultaneously map the original time-domain signal to both the time and frequency domains. In the time-domain extraction process, the transient force and displacement response signals of the vibration damping device under excitation are acquired to construct a force-displacement response trajectory. This trajectory reflects the instantaneous relationship between force and displacement during dynamic loading, and its morphological characteristics directly reflect the device's energy dissipation mechanism. Simultaneously, in the frequency-domain extraction process, a fast Fourier transform is performed on the response signal to obtain the dominant frequency components in the spectral distribution. For each dominant frequency, its corresponding amplitude-frequency characteristic curve and phase-frequency characteristic curve are extracted. The amplitude-frequency characteristic reflects the ratio of response amplitudes under different frequency excitations, while the phase-frequency characteristic reveals the phase lag angle of the response signal relative to the excitation signal. By establishing a correspondence between the time-domain force-displacement trajectory and the frequency-domain amplitude-phase characteristics, a complementary mapping relationship between time-frequency domain features is formed. This mapping relationship organically combines the changes in time-domain physical quantities with the energy distribution characteristics in the frequency domain.

[0024] Based on the established complementary mapping relationship, the nonlinear mechanical characteristics inherent in the process are further identified. For the force-displacement response trajectory in the time domain, the hysteresis loop morphology formed during the loading-unloading cycle is observed. When the damping device exhibits nonlinear stiffness characteristics, the slope of the hysteresis loop shows a non-constant characteristic as the displacement amplitude changes; this slope variation is the force-displacement hysteresis characteristic. Multiple hysteresis curves under different excitation amplitudes are collected to form a family of hysteresis curves. This family of curves is parametrically fitted, using polynomial functions or piecewise linear functions to describe the evolution of stiffness with displacement. During the fitting process, the change in the principal axis slope of the hysteresis loop is the focus. The first derivative of this slope reflects the stiffness softening or hardening characteristics, while the second derivative reveals the strength of the nonlinearity. Simultaneously, phase lag characteristics are identified in the phase-frequency characteristic curve in the frequency domain. When the damping device has nonlinear damping, the phase lag angle is not a constant value but shows a continuous variation trend with the excitation frequency. By extracting the phase lag angles corresponding to different frequency points, a phase-frequency relationship curve is established. Frequency dependence analysis was performed on the curve, and the trend of phase lag angle with frequency was plotted using a logarithmic coordinate system. The characteristic parameters of damping frequency-varying characteristics, including the frequency sensitivity coefficient of phase lag and the phase change rate, were extracted using curve fitting method.

[0025] After extracting the nonlinear evolution law of stiffness and the frequency-varying characteristics of damping, these characteristics need to be accurately characterized mathematically. For the nonlinear evolution law of stiffness, it is characterized as a function of displacement amplitude. Non-monotonic functions The function is chosen to be a cubic polynomial or a hyperbolic tangent function, which effectively describes the softening-hardening transition characteristic of stiffness, which initially decreases and then increases with increasing displacement. The function parameters are determined by fitting the skeleton curve of the hysteresis curve family using the least squares method, ensuring that the function curve accurately passes through the experimental data points. For the damping frequency-varying characteristics, it is characterized as a function of frequency with respect to the excitation frequency. Continuous change function This function, in either power or fractional form, reflects the monotonically increasing or decreasing trend of the damping coefficient with frequency. The function parameters are determined by fitting the phase lag data from the phase-frequency response curve, and the goodness of fit must meet the accuracy requirement of a correlation coefficient greater than 0.95.

[0026] The above stiffness function With damping function As coupled state variables, they are embedded in a set of differential equations describing the dynamic behavior of the damping device. This set of equations is based on Newton's second law and its basic form is a comprehensive equilibrium equation comprising mass, damping, and stiffness terms. The mass term uses the equivalent mass of the damping device. As a coefficient, the coefficient of the damping term is the aforementioned damping function. The coefficient of the stiffness term is the aforementioned stiffness function. Since both stiffness and damping are functions of state variables, the equations exhibit strong nonlinear characteristics. To accurately solve this system of nonlinear differential equations, a numerical integration method is employed, discretizing the time domain into equally spaced steps and progressively solving it using the Runge-Kutta method. During the solution process, the stiffness term is calculated in real-time based on the displacement value at the current moment. The damping term is calculated in real time based on the dominant frequency of the current excitation. This enables dynamic coupling of state variables.

[0027] The constructed nonlinear mechanical characteristic model includes not only a system of differential equations but also a state-space representation. Displacement... With speed The state vector is defined by two components: the external excitation force is defined as the input vector, and the response force is defined as the output vector. The stiffness and damping functions are embedded in the coefficient matrix of the state-space equations, allowing the state transition process to reflect nonlinear characteristics. This state-space representation facilitates subsequent load-performance correlation mapping analysis and allows direct extraction of state parameters reflecting the essential mechanical characteristics of the damping device. After model construction, the prediction error is calculated by comparing the model's predicted response with the measured response data. When the error exceeds a set threshold, an iterative optimization algorithm is used to adjust the parameters of the stiffness and damping functions until the root mean square error between the model output and the measured data is reduced to an acceptable range, thus ensuring the accuracy and reliability of the constructed nonlinear mechanical characteristic model. The entire modeling process realizes the transformation from raw dynamic response data to a high-fidelity nonlinear mechanical model, laying a solid theoretical foundation for the subsequent establishment of load-performance mapping relationships.

[0028] In one optional implementation, the stiffness nonlinear evolution law is extracted by fitting a family of curves to the force-displacement hysteresis characteristics, and the damping frequency-varying characteristics are extracted by frequency-dependent analysis of the phase hysteresis characteristics, including: Multiple hysteresis loops under different displacement amplitude conditions are extracted from the force-displacement hysteresis characteristics to form a family of hysteresis curves. The envelope of each loop in the family of hysteresis curves is subjected to multi-order nonlinear fitting to obtain a continuous function expression of stiffness as a function of displacement amplitude. By performing piecewise gradient analysis on the continuous function expression, the nonlinear strengthening and nonlinear weakening characteristics of stiffness in different displacement ranges are identified. The nonlinear strengthening feature and the nonlinear weakening feature are used as state indicators of stiffness evolution. By constructing a transformation sequence of the state indicators as the displacement amplitude increases, a complete characterization chain describing the nonlinear evolution law of stiffness is formed. The phase difference sequence between force response and displacement response under different excitation frequencies is extracted from the phase hysteresis characteristics. The phase difference sequence is then processed to be continuous in the frequency domain to obtain a smooth functional relationship between the phase difference and the excitation frequency. By performing piecewise differentiation on the smooth functional relationship at different frequencies, the rate of change characteristics of the damping coefficient in different frequency bands are identified. Based on the frequency band differences of the rate of change characteristics, the frequency-varying characteristics of damping are extracted.

[0029] After obtaining the force-displacement hysteresis characteristics of the vibration damping device under test, it is necessary to extract quantitative information reflecting the nonlinear evolution of stiffness. For the dynamic response of the vibration damping device under different displacement amplitudes, the correspondence between force and displacement is recorded, forming multiple closed hysteresis loops. These loops exhibit different shapes and areas in the force-displacement coordinate system, reflecting the energy dissipation and stiffness performance of the vibration damping device under different excitation intensities. These loops are arranged in ascending order of displacement amplitude, forming a family of hysteresis curves. The envelope of each loop in this family of curves represents the maximum restoring force at a specific displacement amplitude, and the slope of the envelope corresponds to the equivalent stiffness under that working condition.

[0030] When performing multi-order nonlinear fitting on the envelopes of each loop in a family of hysteresis curves, a piecewise polynomial function is used to describe the relationship between stiffness and displacement. Specifically, the displacement amplitude range is divided into several intervals, and a polynomial of appropriate order is selected for fitting within each interval. For small displacement intervals, second- or third-order polynomials are usually sufficient to describe the slow change in stiffness; for large displacement intervals, due to significant material nonlinearity or geometric nonlinearity effects, fourth- or higher-order polynomials are required. During the fitting process, the function values ​​and their first derivatives at the boundaries of adjacent intervals are ensured to be continuous, obtaining a continuous functional expression of stiffness changing with displacement amplitude. This function can quantitatively describe the continuous evolution of stiffness from its initial value as the displacement amplitude increases.

[0031] After obtaining the continuous function expression, piecewise gradient analysis is performed to calculate the derivative of the function in each displacement interval. A positive derivative indicates that the stiffness increases with increasing displacement, known as the nonlinear stiffness strengthening characteristic. This usually corresponds to the hardening effect of elastic elements in damping devices under large deformations or the gradual intervention of restraining structures. A negative derivative indicates that the stiffness decreases with increasing displacement, known as the nonlinear stiffness weakening characteristic. This is caused by material softening, the presence of gaps, or the yielding of damping elements. The absolute value of the derivative reflects the degree of stiffness change; the larger the absolute value, the stronger the nonlinear characteristic. By identifying the points where the sign of the derivative changes, the critical displacement at which stiffness evolves from strengthening to weakening or vice versa can be determined.

[0032] The identified nonlinear strengthening and weakening characteristics are used as state identifiers, and a corresponding state label is assigned to each displacement interval. As the displacement amplitude increases, the state identifiers are arranged sequentially to form a transition sequence. This sequence clearly shows the evolution path of stiffness across the entire displacement range, exhibiting a typical pattern of "initial linearity - gradual strengthening - strengthening saturation - weakening." Each state node in the transition sequence corresponds to a specific displacement amplitude and stiffness value, and the transition conditions between nodes are determined by gradient analysis results. In this way, a complete characterization chain describing the nonlinear evolution of stiffness is formed. This characterization chain not only contains numerical information but also implies the physical mechanism of stiffness change.

[0033] To extract damping characteristics, frequency dependence analysis is performed starting with phase lag features. Under different excitation frequencies, the force and displacement responses of the damping device are measured, and their phase information is extracted using Fourier transform or Hilbert transform. The phase difference is obtained by subtracting the displacement response phase from the force response phase; this phase difference reflects the degree of lag of the damping force relative to velocity. The phase differences at multiple excitation frequencies are recorded sequentially to form a phase difference sequence. This sequence typically exhibits a discrete distribution and requires frequency domain continuum conversion before it can be used for subsequent analysis.

[0034] Frequency domain continuity processing employs interpolation or curve fitting methods to transform the discrete phase difference sequence into a continuous function with respect to the excitation frequency. Commonly used methods include spline interpolation or rational fraction fitting, which can generate a smooth functional relationship while maintaining the accuracy of the original data points. The obtained smooth functional relationship describes how the phase difference changes continuously with the excitation frequency, and the shape of the function reflects the frequency sensitivity of the damping mechanism. For damping devices dominated by viscous damping, the phase difference changes relatively smoothly with frequency; for devices dominated by dry friction damping or structural damping, the phase difference exhibits abrupt changes or peaks in specific frequency ranges.

[0035] Piecewise differentiation of the smooth functional relationship is performed to calculate the derivative of the phase difference with respect to the excitation frequency. The physical meaning of the derivative lies in reflecting the rate of change of the damping coefficient with frequency, and the magnitude of the derivative characterizes the strength of the frequency-dependent damping characteristics. The excitation frequency range is divided into several frequency bands, and the average derivative value is calculated in each band as the rate of change characteristic of that band. By comparing the rate of change characteristics of different frequency bands, frequency regions with drastic and gradual changes in the damping coefficient are identified. The bandwise differences in the rate of change characteristics reveal the frequency selectivity of the damping mechanism. For example, some frequency bands are dominated by viscous damping, with smaller derivative values; while in other frequency bands, the material internal resistance or structural friction is dominant, with significantly larger derivative values.

[0036] Based on the differences in the rate of change characteristics across frequency bands, a quantitative description of the frequency-varying damping characteristics is constructed. The frequency axis is divided into three typical regions: low frequency, mid frequency, and high frequency. The average damping coefficient and its rate of change are extracted for each region. The damping characteristics in the low-frequency region are mainly affected by the fluid viscosity effect under large-amplitude motion; the mid-frequency region is affected by the material viscoelasticity and structural coupling effects; and the high-frequency region is affected by local contact friction and micro-vibration dissipation. By quantifying the damping coefficients and their frequency dependence in these three regions, a complete characterization of the frequency-varying damping characteristics is formed. This characterization, together with the aforementioned stiffness nonlinear evolution law, constitutes the core content of the mechanical characteristic parameters of the vibration damping device, providing accurate input data for subsequently establishing the working condition-performance correlation mapping relationship. The entire extraction process ensures that the conversion from the original dynamic response data to the mechanical characteristic parameters has clear physical meaning and traceability, enabling the test results to truly reflect the actual performance of the vibration damping device under different working conditions.

[0037] In one optional implementation, based on the nonlinear mechanical characteristic model, a working condition-performance correlation mapping relationship is established. By correlating the excitation characteristics under different working conditions with the corresponding mechanical characteristic parameters in multiple dimensions, a mapping matrix characterizing the performance evolution of the vibration damping device across the entire working condition range is formed, including: Based on the nonlinear mechanical characteristic model, the excitation amplitude characteristics, excitation frequency characteristics, and excitation waveform characteristics under different working conditions are extracted as working condition feature vectors. The corresponding stiffness characteristic parameters, damping characteristic parameters, and response peak parameters are extracted as performance feature vectors. A many-to-many correlation relationship is established between the working condition feature vectors and the performance feature vectors. By tensor expansion of the many-to-many relationship, an association matrix is ​​constructed with the working condition feature vector as the row index and the performance feature vector as the column index, and the association matrix is ​​filled with association weights that reflect the degree of influence of the working condition on the performance. The correlation matrix is ​​extended to cover the entire working condition range. By interpolating and extrapolating the correlation weights under known working conditions, a mapping matrix covering the entire working condition range is generated.

[0038] After obtaining the nonlinear mechanical characteristic model, a systematic analysis is needed for the various operating conditions encountered by the vibration damping device in practical applications. The mechanical behavior of the vibration damping device varies significantly under different load levels, frequency ranges, and excitation modes, directly affecting its damping effectiveness. By establishing a quantitative correlation between operating conditions and performance, accurate prediction of the vibration damping device's performance status throughout its entire life cycle can be achieved.

[0039] The extraction of excitation features needs to cover three basic dimensions: amplitude, frequency, and waveform. Excitation amplitude features reflect the intensity level of the external load. In actual acquisition, peak detection is performed on the acquired excitation signal to identify the maximum amplitude, average amplitude, and amplitude change rate of a single impact. For periodic excitation, it is necessary to statistically analyze the amplitude distribution characteristics within the complete cycle and calculate the amplitude standard deviation and amplitude fluctuation coefficient. Excitation frequency features are obtained by performing a Fast Fourier Transform on the excitation signal, extracting the dominant frequency component, secondary frequency component, and frequency band energy distribution to form a frequency domain feature set. Excitation waveform features focus on the time-domain morphology of the excitation signal, distinguishing different waveform types such as sinusoidal excitation, triangular wave excitation, and random excitation, and calculating morphological description parameters such as waveform factor and peak factor. These feature parameters are combined to form a working condition feature vector, with each working condition corresponding to a multi-dimensional feature vector.

[0040] The construction of performance feature vectors is based on in-depth analysis of the response behavior of vibration damping devices. Stiffness feature parameters describe the ability of vibration damping devices to resist deformation, obtained by extracting the slope variation law of the force-displacement curve. Since vibration damping devices usually have nonlinear stiffness characteristics, it is necessary to extract equivalent stiffness values ​​in different displacement intervals segmentally and identify stiffness softening or hardening trends. Damping feature parameters reflect energy dissipation capacity, quantified using the logarithmic decay method or the hysteresis loop area method, and the equivalent damping ratio and energy dissipation coefficient are calculated. Peak response parameters directly measure the maximum displacement, maximum velocity, and maximum acceleration of the vibration damping device under excitation. These peak parameters are closely related to the safety margin of the vibration damping device. Under certain complex working conditions, it is also necessary to extract the spectral characteristics of the response signal to identify whether there is resonance amplification phenomenon.

[0041] When establishing the correlation between operating condition feature vectors and performance feature vectors, the nonlinear coupling effect needs to be fully considered. The same operating condition feature simultaneously influences multiple performance features, while the same performance feature is also affected by multiple operating condition features, thus forming a complex many-to-many correlation network. By conducting repeated experiments under multiple sets of operating conditions, a large number of operating condition-performance data pairs are obtained, and correlation analysis is used to identify significant correlation paths. Strongly correlated operating condition-performance pairs are assigned a high correlation strength; for weakly correlated or unrelated pairs, the correlation strength approaches zero.

[0042] Zhang's quantitative expansion process transforms many-to-many relationships into a structured numerical matrix. All operating condition feature vectors are arranged as rows of the matrix according to the experimental sequence, with each row representing a specific operating condition combination. All performance feature vectors are arranged as columns of the matrix, with each column representing a performance index. The value of each element in the matrix represents the influence weight of the corresponding operating condition feature on the performance feature. This weight can be determined using sensitivity analysis, by slightly perturbing a certain operating condition feature and observing the magnitude of change in the corresponding performance feature; the greater the magnitude of change, the higher the correlation weight. In actual calculations, the standardized change in the operating condition feature is used as input, and the relative change in the performance feature is used as output; the ratio of the two is a preliminary estimate of the correlation weight.

[0043] The construction of the correlation matrix needs to ensure the consistency and comparability of the values. Since different characteristic parameters have different dimensions and numerical ranges, normalization is required before filling the matrix to map all characteristic parameters to a unified numerical range. For excitation amplitude characteristics, the range normalization method can be used, with the ratio of the actual amplitude to the maximum amplitude as the normalization result. For frequency characteristics, normalization is performed based on the natural frequency of the damping device. Performance characteristic parameters also need to be standardized, converting the actual measured values ​​into their deviation from the design reference values. After normalization, the elements of the correlation matrix are typically distributed between 0 and 1, with values ​​closer to 1 indicating stronger correlation.

[0044] The full-condition range extension aims to map from discrete test conditions to a continuous condition space. In actual tests, only a limited number of typical conditions can be selected for testing, but the conditions encountered by the vibration damping device during actual service are continuously changing. Through interpolation techniques, the correlation weights of any intermediate condition point can be estimated based on the known correlation weights of the condition points. For single-dimensional condition changes, piecewise linear interpolation or cubic spline interpolation methods are used to construct a continuous function. For multi-dimensional coupled condition changes, multivariate interpolation algorithms, such as radial basis function interpolation or Kriging interpolation, are used to establish a smooth weight distribution surface in the multi-dimensional condition space. Reasonable boundary conditions need to be set during the interpolation process to avoid amplifying prediction errors caused by extrapolation.

[0045] The generated mapping matrix not only contains static operating condition-performance correlation information but also reflects the performance evolution pattern. By overlaying multiple mapping matrices along the time dimension, the degradation trend of performance characteristics under the same operating conditions can be observed. By performing a difference operation on the mapping matrix and calculating the rate of change of matrix elements at adjacent time points, when a certain correlation weight shows a continuous monotonic change, it indicates that the corresponding performance characteristic is undergoing a trend evolution. This performance evolution analysis based on the mapping matrix provides a quantitative data foundation for the subsequent construction of adaptive detection criteria, enabling the detection system to dynamically adjust the judgment criteria according to actual operating conditions and achieve accurate assessment of the performance status of the vibration damping device.

[0046] In one optional implementation, the association matrix is ​​constructed by tensor-quantizing the many-to-many relationship, using the operating condition feature vector as the row index and the performance feature vector as the column index, including: The many-to-many relationship is represented as a third-order tensor structure, and the third-order tensor structure is decomposed to extract the dominant association pattern between the working condition dimension and the performance dimension. Based on the dominant correlation mode, the third-order tensor structure is modally expanded, and the coupling strength is bidirectionally projected along the working condition dimension and the performance dimension to form a two-dimensional matrix structure. The two-dimensional matrix structure is filled with correlation weights by using the components of each dimension of the working condition feature vector as row indices and the components of each dimension of the performance feature vector as column indices, and filling the intersection of the row indices and the column indices with the corresponding correlation weights, thereby constructing the correlation matrix.

[0047] In practice, when structurally expressing the complex correlation between the excitation characteristics of vibration damping devices under different operating conditions and their corresponding performance responses, it is necessary to transform the originally dispersed many-to-many correlation into a computable matrix form. This process first organizes the collected operating condition-performance correlation data into a third-order tensor structure. The three dimensions of this tensor structure correspond to the operating condition feature space, the performance feature space, and the correlation strength space, respectively. Specifically, assuming that during the detection process... Typical operating condition characteristics, each operating condition characteristic is represented by: Vector representation of each component, while extracting Each performance characteristic is represented by a set of performance features. If the vector representation of each component is given, then the original association can be represented as a vector of size . The third-order tensor ,in A quantitative dimension representing the strength of the association.

[0048] For third-order tensors Tensor decomposition is performed, employing the CA6DECOMP / PARAFAC decomposition method to extract dominant association patterns hidden in high-dimensional data. This decomposition process transforms tensors... It can be approximately decomposed into a linear combination of rank-1 tensors, i.e. ,in , , These are factor vectors representing the operating condition dimension, performance dimension, and correlation strength dimension, respectively. This represents the outer product operation. Let the rank be the tensor rank. The optimal rank is determined using an iterative optimization algorithm. This allows the decomposed tensor to reconstruct the original tensor with the fewest rank-1 components, while preserving key operational-performance correlation information. During the iteration process, the factor vectors of each dimension are updated alternately until the reconstruction error converges to a preset threshold, which is typically set to one-thousandth of the norm of the original tensor.

[0049] The factor matrix obtained after tensor decomposition and These represent the dominant correlation patterns in the operating condition and performance dimensions, respectively. To transform the third-order tensor structure into a two-dimensional matrix for easier subsequent calculations, the tensor... Modal expansion is performed along the correlation strength dimension. In practice, the indices of the correlation strength dimension are fixed, and the third-order tensor is sliced ​​and expanded according to the working condition-performance plane. Multiple slices are then merged into a single two-dimensional matrix using a weighted summation method. The weight coefficients used in the expansion process are derived from the factor vector. The components ensure that the expanded two-dimensional matrix retains the main coupling information between the operating conditions and performance in the original tensor.

[0050] In the modal expansion operation, the tensor Expand into a matrix form along the third dimension The rows of this matrix correspond to operating conditions, and the columns correspond to performance characteristics. Specifically, for each element in the tensor... ,in For operating condition index, For performance indexing, For the correlation strength index, through fixed Value and iterate through all and The combination of these elements projects the three-dimensional data onto a two-dimensional plane. Since data from different association strength levels contribute differently to the final association matrix, a hierarchical weighted fusion strategy is adopted. A weight factor is assigned to each strength level, determined by the variance proportion of the data at that level; levels with higher variance proportions receive greater weight. The fusion results in a preliminary two-dimensional association matrix, where the number of rows equals the dimension of the working condition feature vectors. Number of working conditions The product of the columns equals the dimension of the performance feature vector. With the number of performance types The product of.

[0051] To construct the final correlation matrix, specific correlation weight values ​​need to be filled into the two-dimensional matrix structure. The components of each dimension of the working condition feature vector are arranged sequentially as row indices; for example, the first working condition feature vector... Each component occupies the front of the matrix Okay, the components of the second working condition feature vector occupy the next... Rows, and so on. Similarly, the components of each dimension of the performance feature vector are arranged as column indices. After determining the row and column indices, the correlation weight value at the intersection of each index is calculated. This weight value is obtained by calculating the correlation coefficient between the corresponding operating condition feature component and the performance feature component, specifically using a correlation metric method based on mutual information, which can capture linear and nonlinear correlation relationships. For the position located at the... Line 1 The matrix elements of the column, their weight values ​​reflect the first column. The characteristic component of the first working condition is related to the first... The influence strength of each performance characteristic component.

[0052] During the process of filling in the correlation weights, the originally calculated weight values ​​need to be normalized to ensure that the numerical range in the matrix is ​​uniform and facilitates subsequent comparison and analysis. A row-wise normalization strategy is adopted, ensuring that the sum of the weights in each row equals 1, representing the relative distribution of the influence of a specific operating condition characteristic component on all performance characteristic components. The normalized correlation matrix is ​​shown below. This is the final constructed working condition-performance correlation matrix, where each element... The first quantized feature vector of the working condition The component and the th performance feature vector The matrix not only preserves the core information of the original many-to-many relationship, but also facilitates subsequent matrix operations, such as eigenvalue decomposition and singular value decomposition, through its structured matrix form, providing a mathematical basis for dynamically adjusting detection thresholds and identifying abnormal patterns.

[0053] In one optional implementation, an adaptive detection criterion is constructed based on the mapping matrix. By tracking the performance evolution patterns in the mapping matrix in real time, the detection threshold and anomaly identification boundary are dynamically adjusted, enabling the adaptive detection criterion to autonomously optimize with changes in operating conditions. This includes: The evolution trajectory of performance parameters with changing operating conditions is extracted from the mapping matrix. Time series analysis is performed on the evolution trajectory to obtain the trend characteristics and fluctuation amplitude characteristics of the performance parameters. Based on the trend characteristics and fluctuation amplitude characteristics, a performance status evaluation benchmark is established. The performance status evaluation benchmark is used as the initial reference standard for constructing adaptive detection criteria. By performing a working condition sensitivity analysis on the performance status evaluation benchmark, the degree and direction of deviation of performance parameters under different working conditions are identified, and the adjustment amount of the detection threshold is dynamically calculated based on the degree and direction of deviation. The adjustment amount of the detection threshold is then superimposed on the performance status evaluation benchmark to form a dynamic detection threshold that adapts to the working conditions. Based on the dynamic detection threshold, by continuously comparing the real-time collected performance parameters with the dynamic detection threshold, abnormal states where the performance parameters exceed the dynamic detection threshold are identified. Based on the distribution position of the abnormal state in the mapping matrix, the corresponding working condition features are traced back in reverse, and the abnormal identification boundary is dynamically adjusted to match the current working condition, thereby realizing the autonomous optimization of the adaptive detection criterion as the working condition changes.

[0054] In constructing adaptive detection criteria, the first step is to extract the evolution trajectory of performance parameters under varying operating conditions from the mapping matrix. The mapping matrix, in multi-dimensional form, records the mechanical characteristic parameters of the damping device under different operating conditions, including key performance parameters such as stiffness coefficient, damping ratio, and energy dissipation rate. For a specific performance parameter, its numerical sequence is extracted along the operating condition dimension, forming a continuous trajectory curve reflecting the parameter's variation with operating conditions. To accurately capture the evolution pattern, time-series analysis is performed on the extracted trajectory data. The sliding window method is used to calculate the rate of change of the trajectory at different time periods, identifying the rising, falling, or stable trends of the performance parameter. This rate of change sequence constitutes the trend feature. Simultaneously, by calculating the standard deviation or variance of the trajectory data at each operating condition point, the fluctuation amplitude of the performance parameter is quantified. A larger fluctuation amplitude indicates that the performance parameter is more sensitive to changes in operating conditions, while a smaller fluctuation amplitude indicates relatively stable performance.

[0055] Based on the acquired trend and fluctuation characteristics, a performance status assessment benchmark is established. This benchmark is built using statistical analysis methods to fit the distribution characteristics of performance parameters within the normal operating range, determining the expected values ​​and allowable deviation ranges of the performance parameters. In practice, historical data of performance parameters under each operating condition are grouped and statistically analyzed to calculate the mean performance parameters for different operating condition categories. This mean value serves as the performance benchmark reference value for that operating condition. Simultaneously, considering the fluctuation amplitude characteristics, a normal fluctuation range for performance parameters near the benchmark reference value is defined, typically using plus or minus three standard deviations of the benchmark reference value as the initial boundary. This performance status assessment benchmark reflects the normal performance level that the vibration damping device should possess under various operating conditions, providing an initial reference standard for subsequently constructing adaptive testing criteria.

[0056] After establishing the performance status evaluation benchmark, a condition sensitivity analysis is performed to dynamically adjust the detection threshold. Condition sensitivity analysis focuses on identifying the degree of influence of different operating conditions on performance parameters. The real-time acquired operating conditions are matched and compared with the standard operating conditions recorded in the mapping matrix to calculate the distance metric between the current operating condition and each standard operating condition. The distance metric can use multidimensional Euclidean distance or Mahalanobis distance, comprehensively considering differences in multiple operating condition dimensions such as excitation frequency, excitation amplitude, and excitation direction. Based on the distance metric results, the relative position of the current operating condition in the operating condition space of the mapping matrix is ​​determined, and the expected value of the performance parameter corresponding to that position is extracted.

[0057] By comparing real-time performance parameters with expected values, the degree and direction of deviation of the performance parameters are identified. The degree of deviation is quantified by calculating the relative or absolute error between the actual performance parameters and the expected values, while the direction of deviation is determined by the sign of the error: positive deviation indicates that the performance parameters are higher than expected, and negative deviation indicates that the performance parameters are lower than expected. In the scenario of vibration damping device testing, a negative deviation of the stiffness coefficient indicates material fatigue or structural loosening, while a positive deviation of the damping ratio indicates an increase in the viscosity of the damping medium or blockage of the flow channels.

[0058] Based on the degree and direction of the identified deviation, the adjustment amount of the detection threshold is dynamically calculated. This adjustment takes into account both the drastic change in operating conditions and the historical trend of performance parameters. When operating conditions change significantly, a larger fluctuation range in performance parameters is allowed, and the adjustment amount of the detection threshold is increased accordingly. When operating conditions change gradually, a stricter detection standard is maintained, and the adjustment amount is relatively small. The calculation of the adjustment amount also needs to consider the trend of performance parameter changes. If performance parameters show a continuous downward trend and approach the warning level, a strict detection threshold should be maintained even if operating conditions change drastically to avoid missing potential faults.

[0059] The calculated detection threshold adjustment is superimposed on the performance status evaluation benchmark to form a dynamic detection threshold that adapts to changing operating conditions. The dynamic detection threshold includes an upper threshold and a lower threshold. The upper threshold equals the benchmark value plus a positive adjustment, and the lower threshold equals the benchmark value minus a negative adjustment. This dynamic detection threshold can adjust in real time to follow changes in operating conditions, widening the detection boundary under high-excitation conditions to accommodate normal performance fluctuations, and tightening the detection boundary under low-excitation or steady-state conditions to improve anomaly detection sensitivity.

[0060] Based on dynamic detection thresholds, real-time acquired performance parameters are continuously compared to identify abnormal states. The comparison process employs a real-time monitoring mechanism, comparing the performance parameters acquired at each sampling moment with the corresponding dynamic detection threshold under the corresponding operating condition. When a performance parameter value exceeds the upper or lower limit of the dynamic detection threshold, the damping device is determined to be in an abnormal state at that moment. Abnormal state identification does not rely on exceeding limits at a single moment; instead, a continuous judgment strategy is adopted. An abnormal state is only confirmed when the performance parameter continuously exceeds the threshold for multiple consecutive sampling moments, or when the extent of the exceedance exceeds a preset severe abnormality judgment coefficient.

[0061] After identifying an abnormal state, the corresponding operating condition characteristics are traced back based on the distribution of the abnormal state in the mapping matrix. The operating conditions and performance parameter combinations at the time of the abnormality are used as a query index to retrieve historical data samples with similar operating condition-performance combinations from the mapping matrix. By analyzing the label information of these historical samples, it is determined whether the current abnormality is a normal phenomenon under a specific operating condition or a true performance degradation or fault symptom. If a large number of similar abnormal patterns exist in the historical samples and none of them lead to a fault, it is a misjudgment due to insufficient adaptability to the operating condition, and the abnormality identification boundary under that operating condition needs to be adjusted.

[0062] The dynamic adjustment of the anomaly identification boundary is achieved by updating the boundary parameters in the mapping matrix. For operating conditions that lead to false positives, the allowable fluctuation range of the performance parameters corresponding to that condition is expanded. Specific measures include increasing the standard deviation factor or introducing a condition correction factor. For confirmed true anomalies, monitoring under that condition is strengthened to narrow the anomaly identification boundary and improve the sensitivity of subsequent detection. The adjusted boundary parameters are updated to the mapping matrix in real time and fed back into the calculation of the dynamic detection threshold, forming an adaptive optimization closed-loop mechanism.

[0063] A second aspect of the present invention provides a multi-condition simulation testing system for vibration damping devices, comprising: The data acquisition unit is used to acquire the dynamic response data of the shock absorber under multiple working conditions, construct a nonlinear mechanical characteristic model of the shock absorber based on the dynamic response data, extract mechanical characteristic parameters reflecting the stiffness and damping characteristics of the shock absorber by performing joint time-frequency domain analysis on the dynamic response data, and map the mechanical characteristic parameters to the state space of the nonlinear mechanical characteristic model. The working condition performance unit is used to establish a working condition-performance correlation mapping relationship based on the nonlinear mechanical characteristic model. By associating the excitation characteristics under different working conditions with the corresponding mechanical characteristic parameters in multiple dimensions, a mapping matrix is ​​formed to characterize the performance evolution law of the vibration damping device in the entire working condition range. An adaptive judgment unit is used to construct an adaptive detection criterion based on the mapping matrix. By tracking the performance evolution law in the mapping matrix in real time, the detection threshold and anomaly identification boundary are dynamically adjusted so that the adaptive detection criterion can be autonomously optimized as the working conditions change. The performance evaluation unit is used to evaluate the performance of the shock absorption device based on the adaptive detection criteria, generate detection results including performance degradation trends and fault feature identification, and feed the detection results back to the nonlinear mechanical feature model to update the mechanical feature parameters to achieve a closed-loop improvement in detection accuracy.

[0064] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0065] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0066] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0067] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-condition simulation testing method for vibration damping devices, characterized in that, include: The dynamic response data of the vibration damping device under test under multiple working conditions is obtained. Based on the dynamic response data, a nonlinear mechanical characteristic model of the vibration damping device is constructed. By performing joint time-frequency domain analysis on the dynamic response data, mechanical characteristic parameters reflecting the stiffness and damping characteristics of the vibration damping device are extracted, and the mechanical characteristic parameters are mapped to the state space of the nonlinear mechanical characteristic model. Based on the nonlinear mechanical characteristic model, a working condition-performance correlation mapping relationship is established. By associating the excitation characteristics under different working conditions with the corresponding mechanical characteristic parameters in multiple dimensions, a mapping matrix is ​​formed to characterize the performance evolution law of the vibration damping device in the entire working condition range. Based on the mapping matrix, an adaptive detection criterion is constructed. By tracking the performance evolution pattern in the mapping matrix in real time, the detection threshold and anomaly identification boundary are dynamically adjusted, so that the adaptive detection criterion can be autonomously optimized according to changes in working conditions. The performance of the vibration damping device is evaluated based on the adaptive detection criteria, generating detection results that include performance degradation trends and fault feature identification. The detection results are then fed back to the nonlinear mechanical feature model to update the mechanical feature parameters, thereby achieving a closed-loop improvement in detection accuracy.

2. The method according to claim 1, characterized in that, Based on the dynamic response data, a nonlinear mechanical characteristic model of the damping device is constructed. By performing joint time-frequency domain analysis on the dynamic response data, mechanical characteristic parameters reflecting the stiffness and damping characteristics of the damping device are extracted, including: The dynamic response data is synchronously transformed in the time and frequency domains. By extracting the transient force-displacement response trajectory in the time domain and the amplitude-frequency and phase-frequency characteristics of the dominant frequency components in the frequency domain, a complementary mapping relationship of time and frequency domain features is established. Based on the complementary mapping relationship, the force-displacement hysteresis feature reflecting nonlinear stiffness and the phase hysteresis feature reflecting nonlinear damping are identified in the time-frequency domain features. The nonlinear evolution law of stiffness is extracted by curve family fitting of the force-displacement hysteresis feature, and the frequency-dependent characteristics of damping are extracted by frequency dependence analysis of the phase hysteresis feature. The stiffness nonlinear evolution law is characterized as a non-monotonic function with respect to displacement amplitude, and the damping frequency variation characteristic is characterized as a continuously varying function with respect to excitation frequency. The non-monotonic function and the continuously varying function are embedded as coupled state variables into a system of differential equations to construct a nonlinear mechanical characteristic model.

3. The method according to claim 2, characterized in that, The stiffness nonlinear evolution law is extracted by fitting a family of curves to the force-displacement hysteresis characteristics, and the damping frequency-varying characteristics are extracted by frequency-dependent analysis of the phase hysteresis characteristics, including: Multiple hysteresis loops under different displacement amplitude conditions are extracted from the force-displacement hysteresis characteristics to form a family of hysteresis curves. The envelope of each loop in the family of hysteresis curves is subjected to multi-order nonlinear fitting to obtain a continuous function expression of stiffness as a function of displacement amplitude. By performing piecewise gradient analysis on the continuous function expression, the nonlinear strengthening and nonlinear weakening characteristics of stiffness in different displacement ranges are identified. The nonlinear strengthening feature and the nonlinear weakening feature are used as state indicators of stiffness evolution. By constructing a transformation sequence of the state indicators as the displacement amplitude increases, a complete characterization chain describing the nonlinear evolution law of stiffness is formed. The phase difference sequence between force response and displacement response under different excitation frequencies is extracted from the phase hysteresis characteristics. The phase difference sequence is then processed to be continuous in the frequency domain to obtain a smooth functional relationship between the phase difference and the excitation frequency. By performing piecewise differentiation on the smooth functional relationship at different frequencies, the rate of change characteristics of the damping coefficient in different frequency bands are identified. Based on the frequency band differences of the rate of change characteristics, the frequency-varying characteristics of damping are extracted.

4. The method according to claim 1, characterized in that, Based on the aforementioned nonlinear mechanical characteristic model, a working condition-performance correlation mapping relationship is established. By correlating the excitation characteristics under different working conditions with the corresponding mechanical characteristic parameters in multiple dimensions, a mapping matrix characterizing the performance evolution of the vibration damping device across the entire working condition range is formed, including: Based on the nonlinear mechanical characteristic model, the excitation amplitude characteristics, excitation frequency characteristics, and excitation waveform characteristics under different working conditions are extracted as working condition feature vectors. The corresponding stiffness characteristic parameters, damping characteristic parameters, and response peak parameters are extracted as performance feature vectors. A many-to-many correlation relationship is established between the working condition feature vectors and the performance feature vectors. By tensor expansion of the many-to-many relationship, an association matrix is ​​constructed with the working condition feature vector as the row index and the performance feature vector as the column index, and the association matrix is ​​filled with association weights that reflect the degree of influence of the working condition on the performance. The correlation matrix is ​​extended to cover the entire working condition range. By interpolating and extrapolating the correlation weights under known working conditions, a mapping matrix covering the entire working condition range is generated.

5. The method according to claim 4, characterized in that, By tensor-quantizing the many-to-many relationship, an association matrix is ​​constructed with the operating condition feature vector as the row index and the performance feature vector as the column index, including: The many-to-many relationship is represented as a third-order tensor structure, and the third-order tensor structure is decomposed to extract the dominant association pattern between the working condition dimension and the performance dimension. Based on the dominant correlation mode, the third-order tensor structure is modally expanded, and the coupling strength is bidirectionally projected along the working condition dimension and the performance dimension to form a two-dimensional matrix structure. The two-dimensional matrix structure is filled with correlation weights by using the components of each dimension of the working condition feature vector as row indices and the components of each dimension of the performance feature vector as column indices, and filling the intersection of the row indices and the column indices with the corresponding correlation weights, thereby constructing the correlation matrix.

6. The method according to claim 1, characterized in that, Based on the mapping matrix, an adaptive detection criterion is constructed. By tracking the performance evolution patterns in the mapping matrix in real time, the detection threshold and anomaly identification boundary are dynamically adjusted, enabling the adaptive detection criterion to autonomously optimize with changes in operating conditions. The evolution trajectory of performance parameters with changing operating conditions is extracted from the mapping matrix. Time series analysis is performed on the evolution trajectory to obtain the trend characteristics and fluctuation amplitude characteristics of the performance parameters. Based on the trend characteristics and fluctuation amplitude characteristics, a performance status evaluation benchmark is established. The performance status evaluation benchmark is used as the initial reference standard for constructing adaptive detection criteria. By performing a working condition sensitivity analysis on the performance status evaluation benchmark, the degree and direction of deviation of performance parameters under different working conditions are identified, and the adjustment amount of the detection threshold is dynamically calculated based on the degree and direction of deviation. The adjustment amount of the detection threshold is then superimposed on the performance status evaluation benchmark to form a dynamic detection threshold that adapts to the working conditions. Based on the dynamic detection threshold, by continuously comparing the real-time collected performance parameters with the dynamic detection threshold, abnormal states where the performance parameters exceed the dynamic detection threshold are identified. Based on the distribution position of the abnormal state in the mapping matrix, the corresponding working condition features are traced back in reverse, and the abnormal identification boundary is dynamically adjusted to match the current working condition, thereby realizing the autonomous optimization of the adaptive detection criterion as the working condition changes.

7. A multi-condition simulation testing system for vibration damping devices, used to implement the method as described in any one of claims 1-6, characterized in that, include: The data acquisition unit is used to acquire the dynamic response data of the shock absorber under multiple working conditions, construct a nonlinear mechanical characteristic model of the shock absorber based on the dynamic response data, extract mechanical characteristic parameters reflecting the stiffness and damping characteristics of the shock absorber by performing joint time-frequency domain analysis on the dynamic response data, and map the mechanical characteristic parameters to the state space of the nonlinear mechanical characteristic model. The working condition performance unit is used to establish a working condition-performance correlation mapping relationship based on the nonlinear mechanical characteristic model. By associating the excitation characteristics under different working conditions with the corresponding mechanical characteristic parameters in multiple dimensions, a mapping matrix is ​​formed to characterize the performance evolution law of the vibration damping device in the entire working condition range. An adaptive judgment unit is used to construct an adaptive detection criterion based on the mapping matrix. By tracking the performance evolution law in the mapping matrix in real time, the detection threshold and anomaly identification boundary are dynamically adjusted so that the adaptive detection criterion can be autonomously optimized as the working conditions change. The performance evaluation unit is used to evaluate the performance of the shock absorption device based on the adaptive detection criteria, generate detection results including performance degradation trends and fault feature identification, and feed the detection results back to the nonlinear mechanical feature model to update the mechanical feature parameters to achieve a closed-loop improvement in detection accuracy.

8. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 6.

Images