A method and system for testing the fatigue life of springs

Abstract

This application relates to a method and system for testing the fatigue life of a spring. The method includes: simultaneously acquiring instantaneous force and displacement signals of the spring under test, and identifying each loading cycle based on the instantaneous force and displacement signals; extracting characteristic parameters reflecting the mechanical response characteristics of the spring based on the instantaneous force and displacement signals corresponding to each loading cycle; tracking the changing trend of the characteristic parameters with the number of loading cycles, and calculating the rate of change of the characteristic parameters based on the changing trend; and determining the fatigue failure point of the spring based on the changing trend and the rate of change. This invention can significantly improve the reliability of fatigue life test results and provide accurate data support for spring design optimization.

Description

Technical Field

[0001] This application relates to the field of spring testing, and in particular to a method and system for testing the fatigue life of springs. Background Technology

[0002] Miniature springs are critical components in precision equipment, requiring reliability under high-frequency, long-term cyclic loads. Fatigue life testing often demands frequencies above 200 Hz and millions to hundreds of millions of cycles, posing a significant challenge to the testing system. During prolonged high-frequency operation, the actuator's internal energy loss is converted into heat, causing a continuous rise in temperature. Furthermore, the hydraulic oil or grease upon which its moving parts depend also experiences a significant decrease in viscosity due to high temperatures, weakening oil film stability and resulting in a gradual accumulation of lubrication performance degradation.

[0003] Lubrication failure increases friction in the moving parts inside the actuator, causing phase lag and amplitude attenuation between the actual output and control commands. Consequently, the actual waveform applied to the spring deviates from the ideal value. Although the closed-loop control system monitors and adjusts, overcompensation near the actuator's high-frequency response limit can introduce high-frequency oscillations or control stiffness, potentially causing unexpected transient shocks and stress concentrations on the spring. As testing progresses, internal temperature fluctuations in the actuator cause continuous changes in lubricant viscosity, leading to a slow drift in its response characteristics. The load and displacement data collected by sensors cannot reflect the true stress and deformation of the spring, and the waveform deviation itself changes dynamically over time.

[0004] Faced with the above situation, engineers often establish mathematical models by inputting precise test signals and calibrate and compensate the data over time. However, the actuator response characteristics dynamically evolve with lubrication conditions, and the effectiveness of compensation models established before testing or at specific times gradually decreases as testing progresses. Traditional periodic calibration cannot remain effective throughout the entire process. Ultimately, the true load borne by the spring cannot be accurately captured, rendering failure judgment criteria based on stiffness decay and force reduction ineffective. The system may misjudge premature spring failure due to uncompensated load deviations, wasting resources; it may also miss microcracks or fatigue damage, misjudging the spring as working normally. Both of these situations lead to a significant reduction in the reliability of fatigue life test results, failing to provide accurate data support for micro-spring design optimization and quality control.

[0005] Therefore, we propose a method and system for testing the fatigue life of springs. Summary of the Invention

[0006] This application provides a method and system for testing the fatigue life of springs, which at least solves the problem in the prior art that, under high-frequency and long-term cyclic loads, the lubrication performance deteriorates due to the increase in internal temperature of the actuator, which in turn causes a deviation between the actual load waveform applied to the spring and the ideal waveform. This deviation changes over time and cannot be effectively corrected by traditional calibration methods, ultimately leading to a significant decrease in the reliability of fatigue life test results.

[0007] In a first aspect, this application provides a method for testing the fatigue life of a spring, the method comprising:

[0008] The instantaneous force signal and instantaneous displacement signal of the spring under test are acquired synchronously, and each loading cycle is identified based on the instantaneous force signal and the instantaneous displacement signal;

[0009] Based on the instantaneous force signal and instantaneous displacement signal corresponding to each loading cycle, feature parameters reflecting the mechanical response characteristics of the spring are extracted;

[0010] The change trend of the feature parameter with the number of loading loops is tracked, and the change rate of the feature parameter is calculated based on the change trend of the feature parameter;

[0011] Based on the changing trend and the changing rate, the fatigue failure point of the spring is determined.

[0012] Optionally, the synchronous acquisition of the instantaneous force signal and instantaneous displacement signal of the spring under test, and the identification of each loading cycle based on the instantaneous force signal and the instantaneous displacement signal, includes:

[0013] The instantaneous force signal and instantaneous displacement signal of the spring under test, as well as the instantaneous vibration acceleration signal of the test fixture, are acquired simultaneously.

[0014] The high-frequency components that are correlated with the instantaneous vibration acceleration signal are separated from the instantaneous force signal and the instantaneous displacement signal to obtain the purified instantaneous force signal and the purified instantaneous displacement signal;

[0015] Based on the purified instantaneous force signal and the purified instantaneous displacement signal, a hysteresis loop of the loading cycle is constructed, and characteristic parameters reflecting the mechanical response characteristics of the spring are extracted from the hysteresis loop.

[0016] Optionally, tracking the trend of the feature parameters as a function of the number of loading loops includes:

[0017] The instantaneous force signal, instantaneous displacement signal, and first instantaneous local ambient temperature signal of the spring under test are simultaneously acquired, as well as the second instantaneous local ambient temperature signal of the reference object. The reference object is made of the same material as the spring under test and is not subject to loading.

[0018] A quantitative relationship function between the mechanical characteristic parameters of the reference object and temperature is established, and the theoretical temperature sensitivity value of the spring under test is estimated based on the first instantaneous local ambient temperature signal and the quantitative relationship function.

[0019] The difference between the actual characteristic parameters of the tested spring and the theoretical temperature sensitivity value is calculated to obtain the differential characteristics, and the trend of the differential characteristics with the number of loading cycles is tracked.

[0020] Optionally, determining the fatigue failure point of the spring based on the changing trend and the rate of change includes:

[0021] The rate of change of the feature parameters is normalized to obtain the normalized rate of change;

[0022] A multidimensional feature space is constructed based on the normalized rate of change, and the trajectory points of each loading loop in the multidimensional feature space are calculated in real time.

[0023] Multiple fatigue failure modes are preset with corresponding feature trajectory regions, and it is determined whether the trajectory point enters the feature trajectory region. Each feature trajectory region defines a combination of the rate of change of feature parameters under different failure modes.

[0024] Based on the characteristic trajectory region entered by the trajectory point, the fatigue failure point of the spring is determined.

[0025] Optionally, the preset feature trajectory regions corresponding to multiple fatigue failure modes include:

[0026] Short-term loading tests were conducted on a small number of springs from the same batch to obtain the rate of change data of the characteristic parameters;

[0027] Cluster analysis is performed on the rate of change data to identify potential fatigue failure modes, and feature trajectory regions corresponding to the fatigue failure modes are dynamically generated based on the clustering results.

[0028] Optionally, the normalization process for the rate of change of the feature parameters to obtain the normalized rate of change includes:

[0029] Obtain the material properties and failure mechanism of the spring under test, and determine the normalized parameter set based on the material properties and failure mechanism;

[0030] Monitor the instantaneous value and dispersion of the rate of change of each of the feature parameters, and adjust the normalization weights of the feature parameters according to the instantaneous value and the dispersion.

[0031] When the rate of change of the feature parameter is lower than a preset rate of change threshold, a nonlinear normalization function is used to normalize the rate of change of the feature parameter.

[0032] Optionally, when the rate of change of the feature parameter is lower than a preset rate of change threshold, the rate of change of the feature parameter is normalized using a nonlinear normalization function, including:

[0033] Based on the numerical range and historical distribution characteristics of the rate of change of each of the aforementioned characteristic parameters, a piecewise nonlinear normalization function is determined:

[0034] When the rate of change of the feature parameter is close to zero, the rate of change of the feature parameter is normalized using an exponential function.

[0035] When the rate of change of the feature parameter is within a small but distinguishable range, the rate of change of the feature parameter is normalized using a logarithmic function.

[0036] Optionally, when the rate of change of the feature parameter is lower than a preset rate of change threshold, the rate of change of the feature parameter is normalized using a nonlinear normalization function, including:

[0037] Based on the instantaneous value and historical distribution characteristics of the rate of change of each of the aforementioned feature parameters, a reference point and a sensitivity factor are determined;

[0038] The offset is obtained by calculating the difference between the rate of change of the feature parameter and the reference point.

[0039] The offset is multiplied by the sensitivity factor to obtain the multiplication result, and the multiplication result is normalized using a logarithmic function.

[0040] Secondly, this application provides a spring fatigue life testing system, the system comprising:

[0041] The signal acquisition module is used to synchronously acquire the instantaneous force signal and instantaneous displacement signal of the spring under test;

[0042] The loop identification module is used to identify each loading loop based on the synchronously acquired instantaneous displacement signal;

[0043] The feature parameter extraction module is used to extract feature parameters reflecting the mechanical response characteristics of the spring from the instantaneous force signal and instantaneous displacement signal corresponding to each identified loading cycle.

[0044] The trend tracking module is used to track the changing trend of the extracted feature parameters as the number of loading loops increases;

[0045] The rate of change calculation module is used to calculate the rate of change of the feature parameter based on the changing trend of the feature parameter;

[0046] The failure determination module is used to determine the fatigue failure point of the spring based on the changing trend and the rate of change of the characteristic parameters.

[0047] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0048] Compared with related technologies, the spring fatigue life testing method and system provided in this application have at least the following technical advantages:

[0049] By synchronously acquiring instantaneous force and displacement signals of the tested spring, each loading cycle is identified based on these signals. Subsequently, characteristic parameters reflecting the spring's mechanical response characteristics are extracted from each loading cycle. Then, the changing trends of these characteristic parameters with the number of loading cycles are tracked, and their rate of change is calculated. Finally, the fatigue failure point of the spring is determined based on the changing trend and rate of change. The fatigue state is assessed by monitoring the intrinsic changes in the spring's own mechanical response characteristics. Even if there are certain deviations in the external loading waveform, as long as the accumulation of fatigue damage within the spring material causes changes in its mechanical response characteristics, these changes can be captured by the characteristic parameters and their rate of change. By monitoring the changing trends and rates of change of the characteristic parameters in real time, this application can identify subtle changes in the spring's mechanical response characteristics during the accumulation of fatigue damage. Since these changes are a direct manifestation of damage to the internal structure of the spring material, this application can more accurately and reliably determine the fatigue failure point of the spring, avoiding misjudgments or omissions caused by external loading deviations. For micro-springs requiring high-frequency, long-term fatigue testing, this significantly improves the reliability of fatigue life test results, providing precise data support for spring design optimization.

[0050] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0051] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0052] Figure 1 This is a flowchart illustrating a spring fatigue life test method according to an exemplary embodiment.

[0053] Figure 2 This is a flowchart illustrating step S1 according to an exemplary embodiment.

[0054] Figure 3This is a partial flowchart illustrating step S3 according to an exemplary embodiment.

[0055] Figure 4 This is a flowchart illustrating step S4 according to an exemplary embodiment.

[0056] Figure 5 This is a partial flowchart illustrating step S44 according to an exemplary embodiment.

[0057] Figure 6 This is a partial flowchart illustrating step S41 according to an exemplary embodiment.

[0058] Figure 7 This is a flowchart illustrating step S413 according to an exemplary embodiment.

[0059] Figure 8 This is a block diagram illustrating a spring fatigue life testing system according to an exemplary embodiment. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0061] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any creative effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0062] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0063] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects are in an "or" relationship. The terms "first," "second," and "third" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0064] In related technologies, engineers often establish mathematical models by inputting precise test signals and then calibrate and compensate the data over time. However, the actuator response characteristics dynamically evolve with lubrication conditions. The effectiveness of compensation models established before testing or at specific times gradually decreases as testing progresses, and traditional periodic calibration cannot remain effective throughout the entire process. Ultimately, the true load borne by the spring cannot be accurately captured, rendering failure criteria based on stiffness decay and force reduction ineffective. The system may misjudge premature spring failure due to uncompensated load deviations, wasting resources; it may also miss microcracks or fatigue damage, mistakenly classifying the spring as functioning normally. Both of these situations significantly reduce the reliability of fatigue life test results, failing to provide accurate data support for micro-spring design optimization and quality control.

[0065] Based on the above, embodiments of the present invention provide a method and system for testing the fatigue life of springs, which will be described in detail below with reference to specific embodiments and accompanying drawings.

[0066] Example 1

[0067] This invention provides a method for testing the fatigue life of springs. Figure 1 This is a flowchart illustrating a spring fatigue life test method according to an exemplary embodiment. Figure 1 As shown, the method includes:

[0068] S1. Synchronously acquire the instantaneous force signal and instantaneous displacement signal of the spring under test, and identify each loading cycle based on the instantaneous force signal and instantaneous displacement signal;

[0069] In this embodiment, the instantaneous force signal refers to the signal obtained by the force sensor in real time during the loading process, which shows the change in magnitude and direction of the force acting on the tested spring over time; the instantaneous displacement signal refers to the signal obtained by the displacement sensor in real time, which shows the change in the amount of deformation of the tested spring under force over time. A loading cycle refers to a complete cycle in which the spring goes from its initial state through loading, unloading, and returning to its initial state, reflecting the periodic load acting on the spring during fatigue testing. The instantaneous force and displacement signals can be achieved by integrating force and displacement sensors into the fatigue testing equipment. For example, the force sensor can be a piezoelectric force sensor or a resistance strain gauge force sensor, which can measure the force acting on the spring in real time with high precision. The displacement sensor can be a laser displacement sensor or an eddy current displacement sensor, which can measure the instantaneous deformation of the spring non-contactly or in contact. These sensors need to be synchronously sampled by a data acquisition system to ensure the temporal consistency of the force and displacement signals. Based on the acquired instantaneous force and displacement signals, each loading cycle can be identified. In this embodiment, by setting a threshold for force or displacement, when the force or displacement signal reaches a preset peak or valley value, it is marked as the start or end of a loading cycle.

[0070] S2. Based on the instantaneous force signal and instantaneous displacement signal corresponding to each loading cycle, extract the characteristic parameters that reflect the mechanical response characteristics of the spring;

[0071] In this embodiment, characteristic parameters are physical quantities that quantify the mechanical response characteristics of the spring, such as spring stiffness, damping, and energy loss. Changes in these parameters directly reflect the cumulative fatigue damage of the spring material. Furthermore, the instantaneous stiffness of the spring and the hysteresis loop area for each loading cycle are obtained by calculating the slope of the force-displacement curve for each loading cycle. Instantaneous stiffness reflects the spring's resistance to deformation at different loading stages; the hysteresis loop area for each loading cycle represents the energy dissipated by the spring in one loading cycle, reflecting the material's damping characteristics and internal friction. Additionally, parameters such as maximum force, maximum displacement, and residual deformation in the loading cycle can be extracted. The extraction of these parameters can be achieved through mathematical processing and analysis of the original signal, such as integration, differentiation, and peak detection.

[0072] S3. Track the trend of feature parameters as the number of loading loops increases, and calculate the rate of change of feature parameters based on the trend of feature parameters.

[0073] In this embodiment, the rate of change refers to the rate at which the characteristic parameter changes with the number of loading cycles, and it can more sensitively capture early signs of spring fatigue damage. The characteristic parameter values ​​extracted for each loading cycle are plotted as a curve, with the horizontal axis representing the number of loading cycles and the vertical axis representing the characteristic parameter value. By observing the shape of this curve, it can be determined whether the characteristic parameter shows an upward trend, a downward trend, or remains stable. To quantify this change, various methods can be used to calculate the rate of change. In this embodiment, the rate of change is obtained by numerically differentiating the curve of the characteristic parameter changing with the number of loading cycles.

[0074] S4. Determine the fatigue failure point of the spring based on the trend and rate of change;

[0075] In this embodiment, the fatigue failure point refers to the point in time or the number of loading cycles at which the spring, due to accumulated material damage, reaches a critical state during fatigue testing, rendering it unable to continue performing its normal function. In this embodiment, a fatigue failure threshold is preset; when the rate of change of the characteristic parameters exceeds this threshold, the spring is considered to have reached the fatigue failure point. The fatigue failure threshold can be determined based on the spring's material properties, design requirements, and actual application scenarios.

[0076] The technical solutions described above extract characteristic parameters reflecting the spring's own mechanical response characteristics, such as stiffness and damping. Changes in these parameters more directly reflect the internal damage of the spring material, rather than merely deviations in external loading. Even if there is a certain deviation in external loading, as long as this deviation is within a certain range, the changing trend and rate of change of the spring's own mechanical response characteristic parameters can still relatively accurately reflect its fatigue damage state. Furthermore, by tracking the changing trend of characteristic parameters and calculating the rate of change, this application can more sensitively capture early signs of fatigue damage, thereby achieving more accurate failure prediction. This avoids the complex and difficult-to-maintain effective calibration of external loading waveforms, focusing instead on the spring's own intrinsic response, significantly improving the reliability and accuracy of fatigue life test results.

[0077] In one possible design, Figure 2 This is a flowchart illustrating step S1 according to an exemplary embodiment. (Refer to the attached document.) Figure 2 Step S1 includes:

[0078] S11. Synchronously acquire the instantaneous force signal and instantaneous displacement signal of the spring under test, as well as the instantaneous vibration acceleration signal of the test fixture;

[0079] In this embodiment, during the spring fatigue test, these physical quantities are acquired in real time and synchronously using appropriate sensors (such as force sensors, displacement sensors, and acceleration sensors). Among them, the instantaneous vibration acceleration signal is used to characterize the mechanical vibration of the test fixture or test environment, thereby identifying and eliminating the interference of these vibrations on the force and displacement signals.

[0080] S12. Separate the high-frequency components that are correlated with the instantaneous vibration acceleration signal from the instantaneous force signal and the instantaneous displacement signal to obtain the purified instantaneous force signal and the purified instantaneous displacement signal;

[0081] In this embodiment, signal processing techniques, such as digital filters, wavelet analysis, or adaptive filtering, are used to remove high-frequency noise components caused by the vibration of the test fixture from the instantaneous force and instantaneous displacement signals. These high-frequency components usually have a certain correlation with the instantaneous vibration acceleration signal in terms of frequency and phase. By analyzing this correlation, noise can be accurately separated from the original signal, thereby obtaining a purer and more realistic instantaneous force and instantaneous displacement signal.

[0082] S13. Based on the purified instantaneous force signal and the purified instantaneous displacement signal, construct the hysteresis loop of the loading cycle, and extract the characteristic parameters reflecting the mechanical response characteristics of the spring from the hysteresis loop.

[0083] In this embodiment, the purified signal can more accurately reflect the true mechanical response of the spring during the loading cycle, avoiding distortion caused by noise. The hysteresis loop, a curve describing the force-displacement relationship of the spring during loading-unloading, contains rich mechanical information in its shape and area. Characteristic parameters extracted from these purified hysteresis loops, such as the area, stiffness, and damping ratio of the hysteresis loop, will more accurately reflect the mechanical response characteristics of the spring, providing a reliable basis for subsequent fatigue life assessment.

[0084] The technical solution of the above embodiment introduces the instantaneous vibration acceleration signal of the test fixture and uses this signal as a reference to separate the high-frequency components of the original instantaneous force signal and instantaneous displacement signal, thereby effectively removing the noise introduced by external interference such as test fixture vibration and obtaining more accurate spring mechanical response characteristic parameters.

[0085] In one example, during spring fatigue life testing, a piezoelectric accelerometer can be mounted on a test fixture to synchronously acquire instantaneous vibration acceleration signals. Simultaneously, high-precision force and displacement sensors acquire the instantaneous force and displacement signals of the spring under test. During data processing, a Fast Fourier Transform (FFT) can be used to perform frequency domain analysis on the acquired instantaneous force, displacement, and vibration acceleration signals, identifying high-frequency components significantly correlated with the vibration acceleration signal. Subsequently, an adaptive filter, such as one based on the Least Mean Square (LMS) algorithm, is designed to filter the instantaneous force and displacement signals using the instantaneous vibration acceleration signal as a reference input, thereby separating high-frequency noise components and obtaining purified instantaneous force and displacement signals. Based on these purified signals, the force-displacement hysteresis loop for each loading cycle can be accurately plotted, and characteristic parameters such as the hysteresis loop area, maximum force, maximum displacement, and residual displacement can be calculated. For example, by calculating the area of ​​the hysteresis loop, the energy dissipated by the spring in each cycle can be quantified, and the trend of this energy dissipation can serve as an important indicator for judging the accumulation of fatigue damage.

[0086] In one possible design, Figure 3 This is a partial flowchart illustrating step S3 according to an exemplary embodiment. (Refer to the attached diagram.) Figure 3 Step S3 includes:

[0087] S31. Synchronously acquire the instantaneous force signal, instantaneous displacement signal, first instantaneous local ambient temperature signal of the spring under test, and the second instantaneous local ambient temperature signal of the reference object, wherein the reference object is made of the same material as the spring under test and is not subject to loading;

[0088] In this embodiment, the reference object is designed to have the same material properties as the spring under test, but it does not bear any load during the entire test, i.e., it does not suffer fatigue damage. The purpose is that the reference object is only affected by temperature, and its mechanical response characteristics change, thus serving as a benchmark for the influence of temperature on the mechanical properties of materials.

[0089] S32. Establish a quantitative relationship function between the mechanical characteristic parameters of the reference object and temperature, and estimate the theoretical temperature sensitivity value of the spring under test based on the local ambient temperature signal at the first instant and the quantitative relationship function.

[0090] In this embodiment, the quantitative relationship function is obtained by calibration experiments on the mechanical properties of the same material at different temperatures. For example, by measuring characteristic parameters such as the elastic modulus and yield strength of the material at different temperatures, and fitting the mathematical relationship between them and temperature. This quantitative relationship function is used to accurately describe the influence of temperature changes on the mechanical characteristic parameters of the material.

[0091] S33. Calculate the difference between the actual characteristic parameters of the spring under test and the theoretical temperature sensitivity value to obtain the differential characteristics, and track the trend of the differential characteristics with the number of loading cycles.

[0092] In this embodiment, the theoretical temperature influence on the mechanical characteristic parameters of the tested spring at the current temperature is the theoretical temperature sensitivity value of the spring. This theoretical temperature sensitivity value reflects the expected change in characteristic parameters caused solely by temperature changes in the absence of fatigue damage. The differential feature effectively isolates the influence of temperature on the characteristic parameters, allowing the obtained value to more purely reflect the changes in the mechanical response characteristics of the spring due to fatigue damage. Finally, the trend of this differential feature changing with the number of loading cycles is tracked to more accurately assess the fatigue damage accumulation process of the spring.

[0093] The technical solution described above effectively solves the problem of temperature fluctuations interfering with the accuracy of spring mechanical response characteristic measurements by introducing a temperature compensation mechanism. Specifically, by simultaneously acquiring local ambient temperature signals of the spring under test and an unloaded reference object, and utilizing the quantitative relationship function between the mechanical characteristic parameters established by the reference object and temperature, the theoretical influence of temperature on the mechanical characteristic parameters of the spring under test can be accurately estimated. Subtracting this theoretical temperature sensitivity value from the actually measured characteristic parameters, the resulting differential characteristics can more realistically reflect the damage accumulation of the spring under fatigue loading. It is precisely because of this temperature compensation process that the tracked characteristic parameter change trends can more accurately characterize the evolution of fatigue damage.

[0094] In one example, suppose a fatigue life test is being conducted on a car suspension spring. This spring is made of high-strength steel. During the test, in addition to installing force and displacement sensors to collect instantaneous force and displacement signals, a temperature sensor is placed near the spring under test to collect a first instantaneous local ambient temperature signal. Simultaneously, a steel sample of the exact same material as the spring is used as a reference. This sample is placed in an environment similar to the spring under test, but without any load, and another temperature sensor is installed to collect a second instantaneous local ambient temperature signal.

[0095] Before the test began, a quantitative relationship function between the elastic modulus (as a mechanical characteristic parameter) of the high-strength steel and temperature was established through laboratory calibration experiments. For example, the elastic modulus E = E0 - k * T, where E0 is the elastic modulus at the reference temperature, k is the temperature sensitivity coefficient, and T is the temperature.

[0096] When the fatigue test reaches a certain loading cycle, assume the measured actual elastic modulus of the tested spring is E_actual, and its first instantaneous local ambient temperature is T1. Simultaneously, the second instantaneous local ambient temperature of the reference object is T2. Based on a pre-established quantitative relationship function, the theoretically required elastic modulus of the material at temperature T1 can be estimated as E_theoretical = E0 - k * T1.

[0097] Subsequently, the differential characteristic is calculated: ΔE = E_actual - E_theoretical. This ΔE value represents the change in elastic modulus caused by fatigue damage after excluding the influence of temperature. By continuously calculating and tracking the trend of this ΔE value with the number of loading cycles, a purer and more accurate fatigue damage accumulation curve can be obtained, thus more reliably determining the fatigue failure point of the spring.

[0098] In one possible design, Figure 4 This is a flowchart illustrating step S4 according to an exemplary embodiment. (Refer to the attached document.) Figure 4 Step S4 includes:

[0099] S41. Normalize the rate of change of the characteristic parameters to obtain the normalized rate of change;

[0100] In this embodiment, normalization is used to eliminate differences in units and numerical ranges among different feature parameters, ensuring that the rates of change of these parameters are compared and analyzed on a uniform scale. For example, the rates of change can be scaled to the range of 0 to 1, or methods such as Z-score standardization can be used. Normalization ensures that the contribution of the rates of change of each feature parameter to the overall judgment is balanced in subsequent multidimensional analysis, preventing certain parameters with larger values ​​from dominating the judgment results.

[0101] S42. Construct a multidimensional feature space based on the normalized rate of change, and calculate the trajectory points of each loading loop in the multidimensional feature space in real time.

[0102] In this embodiment, the rate of change of multiple normalized feature parameters is used as different dimensions to form a high-dimensional coordinate system. For example, if the rate of change of spring stiffness, damping, and residual deformation are extracted, a three-dimensional feature space can be constructed. Subsequently, after each loading cycle, a unique position point in the multi-dimensional feature space is determined based on the normalized rate of change of each feature parameter at the current moment. As the number of loading cycles increases, these trajectory points will form a trajectory reflecting the fatigue evolution process of the spring.

[0103] S43. Preset multiple characteristic trajectory regions corresponding to fatigue failure modes, and determine whether the trajectory point enters the characteristic trajectory region. Each characteristic trajectory region defines the combination of the rate of change of characteristic parameters under different failure modes.

[0104] In this embodiment, based on historical data, experimental results, or expert experience, specific regions are defined in a multidimensional feature space for different fatigue failure modes (e.g., surface cracking, internal defect propagation, material plastic deformation, etc.). Each feature trajectory region defines the range of combinations of the rates of change of various feature parameters that should be presented under a specific failure mode. Subsequently, the fatigue evolution trajectory of the spring is monitored in real time, and it is determined whether the current trajectory point falls within a preset failure mode region.

[0105] S44. Determine the fatigue failure point of the spring based on the characteristic trajectory area entered by the trajectory point;

[0106] In this embodiment, once the trajectory point enters a certain region, it indicates that the fatigue state of the spring has developed to the stage corresponding to the failure mode represented by that region.

[0107] The technical solution described above effectively solves the problem of accurately distinguishing different fatigue failure modes based solely on a single trend and rate of change by introducing normalization processing, multi-dimensional feature space construction, and preset failure mode regions. First, normalization of the rate of change of the feature parameters eliminates dimensional differences between parameters, allowing for analysis of multiple parameter changes within a unified and comparable framework. Second, by constructing a multi-dimensional feature space, the rates of change of multiple interrelated feature parameters can be integrated to form a more comprehensive and representative description of the fatigue state. The fatigue evolution of the spring is no longer merely a change in a single parameter, but a comprehensive manifestation of the coordinated changes of multiple parameters, which is represented by a unique trajectory in the multi-dimensional space. Finally, the preset feature trajectory regions corresponding to fatigue failure modes provide clear criteria for identifying and classifying different failure mechanisms. When the real-time calculated trajectory point enters a specific feature trajectory region, it not only determines whether the spring is about to or has already experienced fatigue failure, but also further identifies the specific failure mode, thereby achieving accurate judgment of fatigue failure points and effective differentiation of failure modes.

[0108] In one example, suppose that during spring fatigue testing, two key characteristic parameters need to be monitored: the rate of stiffness decay (denoted as R_k) and the rate of change of damping coefficient (denoted as R_d). First, R_k and R_d are normalized, for example, by using linear scaling to map their values ​​to the range of 0 to 1, to obtain the normalized rate of stiffness decay R_k_norm and the normalized rate of change of damping coefficient R_d_norm.

[0109] Subsequently, a two-dimensional feature space is constructed using R_k_norm and R_d_norm as two dimensions. After each loading cycle, a trajectory point is plotted in the two-dimensional feature space based on the current R_k_norm and R_d_norm values. As the number of loading cycles increases, these trajectory points will connect to form a curve reflecting the fatigue evolution path of the spring.

[0110] Furthermore, based on the accelerated fatigue test data and failure analysis results of the same batch of springs, two characteristic trajectory regions corresponding to fatigue failure modes were predefined: Region A represents "fatigue failure caused by material plastic deformation," characterized by a significant increase in R_k_norm and a relatively small change in R_d_norm; Region B represents "fatigue failure caused by microcrack propagation," characterized by a significant increase in R_d_norm and a relatively small change in R_k_norm. These regions are defined as specific boundary ranges in the two-dimensional feature space.

[0111] During real-time testing, the system continuously calculates trajectory points. When a trajectory point enters region A, the system determines that the spring has experienced fatigue failure due to plastic deformation of the material and marks this point as a fatigue failure point. When the trajectory point enters region B, the system determines that the spring has experienced fatigue failure due to microcrack propagation and marks this point as a fatigue failure point. In this way, not only can fatigue failure be determined, but the specific failure mode can also be identified, providing more detailed information for fault diagnosis and prevention.

[0112] In one possible design, Figure 5 This is a partial flowchart illustrating step S44 according to an exemplary embodiment. (Refer to the attached diagram.) Figure 5 Step S44 includes:

[0113] S441. Conduct short-term loading tests on a small number of springs from the same batch to obtain data on the rate of change of characteristic parameters;

[0114] In this embodiment, a small number of spring samples with the same material, batch, and manufacturing process as the spring under test are selected and subjected to short-duration, high-frequency loading cycle tests under controlled conditions. This allows for the rapid acquisition of data on the changes in the mechanical response characteristics of the spring in the early or middle stages of fatigue without causing complete spring failure, providing fundamental data for subsequent failure mode identification. Subsequently, during the short-term loading test, instantaneous force and displacement signals of the spring under test are simultaneously acquired, and characteristic parameters reflecting the mechanical response characteristics of the spring, such as stiffness, damping, and energy dissipation, are extracted based on these signals. The changing trends of these characteristic parameters with the number of loading cycles are tracked, and the rate of change of each characteristic parameter is calculated.

[0115] S442. Cluster analysis of the rate of change data is performed to identify potential fatigue failure modes, and based on the clustering results, the characteristic trajectory regions corresponding to the fatigue failure modes are dynamically generated.

[0116] In this embodiment, a suitable clustering algorithm (e.g., K-means, DBSCAN, hierarchical clustering, etc.) is used to analyze the rate of change data of the acquired feature parameters. The purpose of clustering analysis is to group data points with similar rate of change patterns into one category, thereby identifying different failure modes that may occur in the spring during fatigue. For example, some springs may exhibit a pattern of rapid decrease in stiffness, while others may exhibit a pattern of significant increase in damping; these different patterns correspond to different fatigue failure mechanisms. Subsequently, based on the results of clustering analysis, different data clusters are associated with specific fatigue failure modes. For example, one cluster may correspond to failure caused by plastic deformation of the material, while another cluster may correspond to failure caused by microcrack propagation. Understandably, this identification process can be combined with expert experience or existing knowledge of failure mechanisms. Finally, in the multidimensional feature space, corresponding feature trajectory regions are delineated based on the data clusters of each failure mode obtained from the clustering analysis. The boundaries of these regions can be determined by the cluster center, the degree of dispersion, and statistical methods (such as Gaussian mixture models). Since these regions are dynamically generated based on actual test data, they can more accurately reflect the true fatigue failure modes of the spring.

[0117] The technical solution described above addresses the issues of insufficient accuracy and poor adaptability inherent in traditional preset feature trajectory regions by introducing short-term loading tests and cluster analysis. Specifically, by conducting short-term loading tests on a small batch of springs, data on the changes in the mechanical response characteristics of the springs during the early or middle stages of fatigue can be obtained quickly and economically. This data includes information on the rate of change of characteristic parameters of the springs under different fatigue failure modes. Subsequently, cluster analysis is performed on these rate of change data, grouping spring samples with similar fatigue failure mechanisms into one category, thereby objectively identifying potential fatigue failure modes. Based on the results of the cluster analysis, feature trajectory regions corresponding to these failure modes are dynamically generated in a multi-dimensional feature space, ensuring that these regions are determined based on actual data rather than subjective experience. Therefore, when the trajectory point of the spring under test enters these dynamically generated regions, its fatigue failure mode can be more accurately determined, and the fatigue failure point can be identified accordingly.

[0118] In one example, suppose a batch of automotive suspension springs needs to be tested for fatigue life. First, five spring samples are randomly selected from the batch and subjected to short-term loading tests on a fatigue testing machine, for example, 100,000 cycles of loading, while simultaneously acquiring the instantaneous force and displacement signals for each spring. During the test, every 10,000 cycles, the spring's stiffness, damping coefficient, and energy dissipation rate are extracted as characteristic parameters, and the rate of change of these characteristic parameters relative to their initial values ​​is calculated.

[0119] Subsequently, the data on the rate of change of stiffness, damping coefficient, and energy dissipation rate of these five spring samples at different cycles were compiled into a multidimensional dataset. K-means clustering was used to analyze this data, with a cluster size of 3, to identify three potential fatigue failure modes. The clustering results showed that the first type of spring mainly exhibited a slow decrease in stiffness with little change in damping; the second type of spring exhibited a rapid decrease in stiffness with a significant increase in energy dissipation rate; and the third type of spring exhibited little decrease in stiffness but large fluctuations in the damping coefficient.

[0120] Based on these clustering results, a characteristic trajectory region is dynamically defined for each cluster in a three-dimensional feature space composed of the rate of change of stiffness, the rate of change of damping coefficient, and the rate of change of energy dissipation. For example, the data point distribution region of the first cluster is defined as the "material plastic deformation failure region," the data point distribution region of the second cluster is defined as the "macroscopic crack propagation failure region," and the data point distribution region of the third cluster is defined as the "microstructural damage failure region." The boundaries of these regions can be determined by calculating the mean and standard deviation of each cluster and combining them with statistical methods (such as the 3σ principle). When fatigue life tests are subsequently performed on other springs from the same batch, once the trajectory points calculated in real time enter these dynamically generated characteristic trajectory regions, their fatigue failure modes can be accurately determined, and the fatigue failure points can be identified accordingly.

[0121] In one possible design, Figure 6 This is a partial flowchart illustrating step S41 according to an exemplary embodiment. (Refer to the attached diagram.) Figure 6 Step S41 includes:

[0122] S411. Obtain the material properties and failure mechanism of the spring under test, and determine the normalized parameter set based on the material properties and failure mechanism;

[0123] In this embodiment, before or during fatigue testing, information about the physical and chemical properties of the spring under test, as well as the types of fatigue failure that may occur under specific operating conditions, is collected. Material properties may include, but are not limited to, the spring's chemical composition, crystal structure, heat treatment state, hardness, strength, and other mechanical performance indicators. Failure mechanisms may refer to the initiation location, propagation path, and fracture mode of fatigue cracks. Based on this information, a customized set of normalization parameters can be determined, for example, by setting specific normalization ranges, reference values, scaling factors, or offsets, to ensure that the normalization process better reflects the mechanical response characteristics of specific materials and failure modes.

[0124] S412. Monitor the instantaneous value and dispersion of the rate of change of each feature parameter, and adjust the normalization weight of the feature parameter according to the instantaneous value and dispersion.

[0125] In this embodiment, the rate of change of each characteristic parameter (e.g., stiffness, damping, hysteresis loop area, etc.) in each loading cycle is tracked in real time, and the fluctuation range or statistical distribution of these values ​​within a certain time window is evaluated. The instantaneous value reflects the trend of change at the current moment, while the degree of dispersion (e.g., standard deviation or coefficient of variation) reflects the stability or randomness of the change. Based on this monitoring data, the normalization weights of the characteristic parameters can be dynamically adjusted. For example, for characteristic parameters that change weakly in the early stage of fatigue but have important indicative significance, higher weights can be assigned so that they have a greater impact on the trajectory points in the multidimensional feature space after normalization; conversely, for characteristic parameters that fluctuate greatly but have a low correlation with fatigue failure, their weights can be appropriately reduced to suppress noise interference.

[0126] S413. When the rate of change of the feature parameter is lower than the preset rate of change threshold, the rate of change of the feature parameter is normalized using a non-linear normalization function.

[0127] In this embodiment, the preset rate of change threshold is an empirical value or a critical value determined through experiments, used to distinguish between significant and minute changes. When the rate of change of the feature parameter is small but may have important indicative significance, linear normalization may not be effective in distinguishing these subtle differences. In this case, using a nonlinear normalization function, such as a logarithmic function, exponential function, or sigmoid function, can amplify the relative differences between these minute changes, making them more clearly reflected in the normalized multidimensional feature space, thereby improving the ability to identify early fatigue damage.

[0128] The technical solutions described above enable a more refined and adaptive normalization process, overcoming the limitations of traditional fixed normalization methods in processing complex spring fatigue data. Specifically, by customizing the normalization parameter set based on material properties and failure mechanisms, the physical meaning and accuracy of the normalization results are ensured. Dynamically adjusting the normalization weights allows the system to respond more flexibly to instantaneous changes and dispersion of characteristic parameters, improving sensitivity to early fatigue damage signals. Especially when the rate of change of characteristic parameters is low, nonlinear normalization effectively amplifies weak but crucial fatigue evolution signals, significantly improving the accuracy and robustness of fatigue failure point judgment, thereby extending the service life of the spring and enhancing safety.

[0129] In one possible design, step S431 includes:

[0130] S4311. Based on the numerical range and historical distribution characteristics of the rate of change of each characteristic parameter, determine the piecewise nonlinear normalization function:

[0131] In this embodiment, before normalizing the rate of change of the characteristic parameters, it is first necessary to perform statistical analysis on the numerical distribution of these rates of change, including their minimum, maximum, average, and variance. Combined with the rate of change performance at different fatigue stages in historical test data, the entire rate of change interval is divided into several sub-intervals. Each sub-interval can be assigned a different nonlinear normalization function based on its numerical characteristics and its indicative significance for fatigue failure. This ensures that the normalization process provides optimal sensitivity and discriminability for minute rates of change of different magnitudes.

[0132] S4312A. When the rate of change of the characteristic parameter is close to zero, the rate of change of the characteristic parameter is normalized by an exponential function.

[0133] In this embodiment, when the rate of change of the detected feature parameter is very small, for example, its absolute value is much smaller than a preset minimum threshold, in order to amplify these weak changes and make them easier to identify and track in the normalized multidimensional feature space, an exponential function can be used for normalization. For example, a function of the form y = a * exp(b * x) can be used, where x is the original rate of change, y is the normalized value, and a and b are coefficients determined according to the actual data distribution, thereby enhancing the sensitivity to very early and very weak fatigue signals.

[0134] S4312B: When the rate of change of the characteristic parameter is in a small but distinguishable range, the logarithmic function is used to normalize the rate of change of the characteristic parameter.

[0135] In this embodiment, when the rate of change of a feature parameter is small, but its value is clearly distinguishable and falls within a relatively wide range of small changes, a logarithmic function can be used for normalization to compress its numerical range while maintaining its relative diversity. For example, a function of the form y = c * log(d * x + e) ​​can be used, where x is the original rate of change, y is the normalized value, and c, d, and e are coefficients determined based on the actual data distribution. This effectively handles small changes over a wide range while maintaining discriminability, avoiding excessive data concentration or dispersion.

[0136] The technical solution described above significantly improves the sensitivity to detecting minute changes in the mechanical response of springs in the early stages of fatigue. Specifically, by determining a piecewise nonlinear normalization function based on the numerical range and historical distribution characteristics of the rate of change, and by selectively employing exponential and logarithmic functions, the rate of change of the normalized characteristic parameters can more accurately characterize the degree of fatigue damage in the spring. This not only helps to identify potential fatigue failure risks earlier but also improves the accuracy and reliability of fatigue failure point determination, thereby extending the service life of the spring and reducing the risk of accidents caused by fatigue failure.

[0137] In one possible design, Figure 7 This is a flowchart illustrating step S431 according to another exemplary embodiment. (Refer to the attached diagram.) Figure 7 Step S431 includes:

[0138] S4313. Based on the instantaneous value of the rate of change and the historical distribution characteristics of each characteristic parameter, determine the reference point and sensitivity factor;

[0139] In this embodiment, the reference benchmark is the typical rate of change value of each characteristic parameter under stable operating conditions or in the early stage of fatigue. Its determination can be based on statistical analysis of historical test data, such as calculating the average or median rate of change in the fatigue-free stage, or the stable rate of change at the beginning of the test. The sensitivity factor is used to quantify the degree of response of the characteristic parameter to fatigue damage. Its value can be set according to the material properties of the spring, the failure mechanism, and empirical knowledge, or calibrated by performing accelerated fatigue testing on springs with known failure modes. For example, for characteristic parameters that are highly sensitive to fatigue damage, a larger sensitivity factor can be set to amplify their small changes.

[0140] S4314. Calculate the difference between the rate of change of the characteristic parameter and the reference point to obtain the offset.

[0141] In this embodiment, the offset refers to the difference between the rate of change of the currently instantaneously changing characteristic parameter and the preset reference point. This difference reflects the degree of deviation of the current state from the stable state.

[0142] S4315. Multiply the offset by the sensitivity factor to obtain the multiplication result, and normalize the multiplication result using a logarithmic function.

[0143] In this embodiment, the offset is multiplied by a sensitivity factor. Based on the sensitivity of this feature parameter to fatigue, the degree of deviation is weighted to obtain a multiplication result that better reflects the degree of fatigue damage. Finally, a logarithmic function is used to normalize the multiplication result, thereby compressing the range of larger values ​​while expanding the range of smaller values. This allows even minute relative changes to be effectively distinguished and amplified in the low rate of change region, thus improving the ability to identify early fatigue damage.

[0144] The technical solutions described above can significantly improve the sensitivity and accuracy of identifying early fatigue damage in spring fatigue life testing. Specifically, by introducing personalized reference points and adjustable sensitivity factors, the normalization process can better adapt to the characteristics of different feature parameters and actual working conditions. Therefore, even in the early stages of fatigue, when the absolute value of the rate of change of feature parameters is low, their relative changes can be effectively captured and amplified, providing a more reliable data basis for the accurate determination of fatigue failure points, thereby extending the service life of the spring.

[0145] In summary, the spring fatigue life testing method provided by this invention synchronously acquires the instantaneous force and displacement signals of the spring under test, and identifies each loading cycle based on these signals. Subsequently, characteristic parameters reflecting the mechanical response characteristics of the spring are extracted from each loading cycle. Then, the changing trends of these characteristic parameters with the number of loading cycles are tracked, and their rate of change is calculated. Finally, the fatigue failure point of the spring is determined based on the changing trend and rate of change. The fatigue state is assessed by monitoring the intrinsic changes in the spring's own mechanical response characteristics. Even if there is a certain deviation in the external loading waveform, as long as the accumulation of fatigue damage within the spring material causes a change in its mechanical response characteristics, it can be captured by the characteristic parameters and their rate of change. By monitoring the changing trend and rate of change of the characteristic parameters in real time, this application can identify subtle changes in the mechanical response characteristics of the spring during the accumulation of fatigue damage. Since these changes are a direct manifestation of damage to the internal structure of the spring material, this application can more accurately and reliably determine the fatigue failure point of the spring, avoiding misjudgments or omissions caused by external loading deviations. For miniature springs that require high-frequency, long-term fatigue testing, this method can significantly improve the reliability of fatigue life test results and provide accurate data support for spring design optimization.

[0146] Example 2

[0147] Embodiment 2 of this application provides a spring fatigue life testing system. Figure 8 This is a block diagram illustrating a spring fatigue life testing system according to an exemplary embodiment. (Refer to...) Figure 8 The system includes:

[0148] The signal acquisition module 01 is used to synchronously acquire the instantaneous force signal and instantaneous displacement signal of the spring under test, and to identify each loading cycle based on the instantaneous force signal and instantaneous displacement signal;

[0149] The parameter extraction module 02 is used to extract characteristic parameters reflecting the mechanical response characteristics of the spring based on the instantaneous force signal and instantaneous displacement signal corresponding to each loading cycle.

[0150] The tracking calculation module 03 is used to track the changing trend of the feature parameters with the number of loading loops, and calculate the rate of change of the feature parameters based on the changing trend of the feature parameters;

[0151] Failure detection module 04 is used to determine the fatigue failure point of the spring based on the trend and rate of change.

[0152] In summary, the spring fatigue life testing method and system provided by this invention synchronously acquires the instantaneous force and displacement signals of the spring under test, and identifies each loading cycle based on these signals. Subsequently, characteristic parameters reflecting the mechanical response characteristics of the spring are extracted from each loading cycle. Then, the changing trends of these characteristic parameters with the number of loading cycles are tracked, and their rate of change is calculated. Finally, the fatigue failure point of the spring is determined based on the changing trend and rate of change. The fatigue state is assessed by monitoring the intrinsic changes in the spring's own mechanical response characteristics. Even if there is a certain deviation in the external loading waveform, as long as the accumulation of fatigue damage within the spring material causes changes in its mechanical response characteristics, these changes can be captured by the characteristic parameters and their rate of change. By monitoring the changing trends and rates of change of the characteristic parameters in real time, this application can identify subtle changes in the mechanical response characteristics of the spring during the accumulation of fatigue damage. Since these changes are a direct manifestation of damage to the internal structure of the spring material, this application can more accurately and reliably determine the fatigue failure point of the spring, avoiding misjudgments or omissions caused by external loading deviations. For miniature springs that require high-frequency, long-term fatigue testing, this method can significantly improve the reliability of fatigue life test results and provide accurate data support for spring design optimization.

[0153] Example 3

[0154] Embodiment 3 of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, it implements the steps of the method provided in Embodiment 1 above.

[0155] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0156] In a possible implementation, the present invention can also be implemented as a program product comprising program code that, when the program product is run on a terminal device, causes the terminal device to perform steps implementing the method provided in Embodiment 1.

[0157] The program code for executing the present invention can be written in any combination of one or more programming languages. The program code can be executed entirely on the user device, partially on the user device, as a standalone software package, partially on the user device and partially on a remote device, or entirely on a remote device.

[0158] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0159] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for testing the fatigue life of a spring, characterized in that, The method includes: The instantaneous force and displacement signals of the spring under test are simultaneously acquired, and each loading cycle is identified based on the instantaneous force and displacement signals. This process includes the following steps: simultaneously acquiring the instantaneous force and displacement signals of the spring under test, as well as the instantaneous vibration acceleration signal of the test fixture; separating the high-frequency components correlated with the instantaneous vibration acceleration signal from the instantaneous force and displacement signals to obtain purified instantaneous force and displacement signals; constructing the hysteresis loop of the loading cycle based on the purified instantaneous force and displacement signals, and extracting characteristic parameters reflecting the mechanical response characteristics of the spring from the hysteresis loop. Based on the instantaneous force signal and instantaneous displacement signal corresponding to each loading cycle, feature parameters reflecting the mechanical response characteristics of the spring are extracted; The change trend of the feature parameter with the number of loading loops is tracked, and the change rate of the feature parameter is calculated based on the change trend of the feature parameter; Based on the changing trend and the changing rate, the fatigue failure point of the spring is determined.

2. The spring fatigue life testing method according to claim 1, characterized in that, The tracking of the trend of the feature parameters with the number of loading loops includes: The instantaneous force signal, instantaneous displacement signal, and first instantaneous local ambient temperature signal of the spring under test are simultaneously acquired, as well as the second instantaneous local ambient temperature signal of the reference object. The reference object is made of the same material as the spring under test and is not subject to loading. A quantitative relationship function between the mechanical characteristic parameters of the reference object and temperature is established, and the theoretical temperature sensitivity value of the spring under test is estimated based on the first instantaneous local ambient temperature signal and the quantitative relationship function. The difference between the actual characteristic parameters of the tested spring and the theoretical temperature sensitivity value is calculated to obtain the differential characteristics, and the trend of the differential characteristics with the number of loading cycles is tracked.

3. The spring fatigue life testing method according to claim 1, characterized in that, The step of determining the fatigue failure point of the spring based on the changing trend and the rate of change includes: The rate of change of the feature parameters is normalized to obtain the normalized rate of change; A multidimensional feature space is constructed based on the normalized rate of change, and the trajectory points of each loading loop in the multidimensional feature space are calculated in real time. Multiple fatigue failure modes are preset with corresponding feature trajectory regions, and it is determined whether the trajectory point enters the feature trajectory region. Each feature trajectory region defines a combination of the rate of change of feature parameters under different failure modes. Based on the characteristic trajectory region entered by the trajectory point, the fatigue failure point of the spring is determined.

4. The spring fatigue life testing method according to claim 3, characterized in that, The feature trajectory regions corresponding to the preset multiple fatigue failure modes include: Short-term loading tests were conducted on a small number of springs from the same batch to obtain the rate of change data of the characteristic parameters; Cluster analysis is performed on the rate of change data to identify potential fatigue failure modes, and feature trajectory regions corresponding to the fatigue failure modes are dynamically generated based on the clustering results.

5. The spring fatigue life testing method according to claim 3, characterized in that, The normalization process for the rate of change of the feature parameters, to obtain the normalized rate of change, includes: Obtain the material properties and failure mechanism of the spring under test, and determine the normalized parameter set based on the material properties and failure mechanism; Monitor the instantaneous value and dispersion of the rate of change of each of the feature parameters, and adjust the normalization weights of the feature parameters according to the instantaneous value and the dispersion. When the rate of change of the feature parameter is lower than a preset rate of change threshold, a nonlinear normalization function is used to normalize the rate of change of the feature parameter.

6. The spring fatigue life testing method according to claim 5, characterized in that, When the rate of change of the feature parameter is lower than a preset rate of change threshold, the rate of change of the feature parameter is normalized using a nonlinear normalization function, including: Based on the numerical range and historical distribution characteristics of the rate of change of each of the aforementioned characteristic parameters, a piecewise nonlinear normalization function is determined: When the rate of change of the feature parameter is close to zero, the rate of change of the feature parameter is normalized using an exponential function. When the rate of change of the feature parameter is within a small but distinguishable range, the rate of change of the feature parameter is normalized using a logarithmic function.

7. The spring fatigue life testing method according to claim 5, characterized in that, When the rate of change of the feature parameter is lower than a preset rate of change threshold, the rate of change of the feature parameter is normalized using a nonlinear normalization function, including: Based on the instantaneous value and historical distribution characteristics of the rate of change of each of the aforementioned feature parameters, a reference point and a sensitivity factor are determined; The offset is obtained by calculating the difference between the rate of change of the feature parameter and the reference point. The offset is multiplied by the sensitivity factor to obtain the multiplication result, and the multiplication result is normalized using a logarithmic function.

8. A spring fatigue life testing system, characterized in that, The system includes: The signal acquisition module is used to synchronously acquire the instantaneous force signal and instantaneous displacement signal of the spring under test; The cycle identification module is used to identify each loading cycle based on the synchronously acquired instantaneous displacement signal, including the following steps: synchronously acquiring the instantaneous force signal and instantaneous displacement signal of the spring under test, as well as the instantaneous vibration acceleration signal of the test fixture; separating the high-frequency components that are correlated with the instantaneous vibration acceleration signal from the instantaneous force signal and the instantaneous displacement signal to obtain purified instantaneous force signal and purified instantaneous displacement signal; constructing the hysteresis loop of the loading cycle based on the purified instantaneous force signal and the purified instantaneous displacement signal, and extracting characteristic parameters reflecting the mechanical response characteristics of the spring from the hysteresis loop; The feature parameter extraction module is used to extract feature parameters reflecting the mechanical response characteristics of the spring from the instantaneous force signal and instantaneous displacement signal corresponding to each identified loading cycle. The trend tracking module is used to track the changing trend of the extracted feature parameters as the number of loading loops increases; The rate of change calculation module is used to calculate the rate of change of the feature parameter based on the changing trend of the feature parameter; The failure determination module is used to determine the fatigue failure point of the spring based on the changing trend and the rate of change of the characteristic parameters.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method described in any one of claims 1-7.

Images