A method and system for monitoring the performance of an ultra-precision spring

Abstract

The application discloses a kind of ultra-precision spring performance monitoring method and system, it is related to the field of precision manufacturing, method includes: after applying compression force to ultra-precision spring, the static friction state of the contact surface of the ultra-precision spring and support clamp is determined;If the static friction state is clamp contact surface lubrication exception, after performing axial cyclic loading to ultra-precision spring, interface friction dissipation and spring body energy dissipation are calculated;After suspending the axial cyclic loading and applying compression force again, the static friction coefficient is obtained;According to the interface friction dissipation, the spring body energy dissipation and the static friction coefficient, performance detection is carried out, and performance detection result is obtained.The application realizes ultra-precision spring performance monitoring, and improves accuracy.

Description

Technical Field

[0001] The present invention relates to the field of precision manufacturing technology, and in particular to a method and system for monitoring the performance of an ultra-precision spring. Background Art

[0002] The performance of ultra-precision springs impacts product reliability, and accurately assessing their service life is a crucial component of quality control in spring production lines. Performance testing systems simulate the spring's actual operating conditions, subjecting the spring to cyclic loading and calculating energy dissipation to generate performance test results. However, in actual operation, the organic cleaning agents used can dissolve the solid lubricant film on the support fixture surface, causing localized cold welding and tearing. This increases the static friction coefficient and leads to a "stick-slip" phenomenon during cyclic loading: the contact surfaces initially "stick," accumulating stress, then suddenly "slipping" a short distance, releasing the stress, and then "sticking" again. Traditional performance testing methods mistakenly misinterpret the energy dissipated by frictional work as energy dissipated by the ultra-precision spring's inherent characteristics. This results in inflated energy dissipation values, falsely signaling performance failure alerts, and inaccurate testing.

[0003] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0004] The main purpose of the embodiments of the present invention is to provide a method and system for monitoring the performance of a super-precision spring, which can realize super-precision spring performance monitoring by identifying and removing interface friction dissipation caused by static friction, thereby improving accuracy.

[0005] In one aspect, an embodiment of the present invention provides a method for monitoring the performance of a super-precision spring, comprising the following steps:

[0006] After applying a compressive force to the super-precision spring, determining a static friction state between a contact surface of the super-precision spring and a support fixture;

[0007] If the static friction state is abnormal lubrication of the fixture contact surface, then after performing axial cyclic loading on the super-precision spring, calculate the interface friction dissipation and the energy dissipation of the spring body;

[0008] After pausing the axial cyclic loading and reapplying the compressive force, obtaining the static friction coefficient;

[0009] A performance test is performed based on the interface friction dissipation, the spring body energy dissipation and the static friction coefficient to obtain a performance test result.

[0010] In some embodiments, determining the static friction state of the contact surface between the super-precision spring and the support fixture includes:

[0011] collecting first force data and first displacement data;

[0012] generating a force-time relationship curve based on the first force data and the first displacement data;

[0013] If there is a mutation point in the force-time relationship curve and the starting force value required to overcome static friction is greater than the preset healthy lubrication state threshold, the static friction state is determined to be abnormal lubrication of the fixture contact surface; otherwise, the static friction state is determined to be normal lubrication of the fixture contact surface.

[0014] In some embodiments, the calculating of interface friction dissipation and spring body energy dissipation includes:

[0015] Get the total energy dissipation;

[0016] determining whether the support fixture has been subjected to transverse micro-vibration, wherein the direction of the transverse micro-vibration is perpendicular to the direction of the axial cyclic loading, and the transverse micro-vibration is used to suppress interface friction;

[0017] When the supporting fixture has been subjected to lateral micro-vibration, collecting second force data and second displacement data;

[0018] When the supporting fixture is not subjected to lateral micro-vibration, collecting third force data and third displacement data;

[0019] performing a local morphological analysis on the second force data and the second displacement data to obtain a first mechanical feature;

[0020] performing a local morphological analysis on the third force data and the third displacement data to obtain a second mechanical feature;

[0021] Comparing the first mechanical characteristic with the second mechanical characteristic to obtain a comparison result;

[0022] According to the comparison result, the energy dissipation corresponding to the reduced mechanical characteristics when the lateral micro-vibration is applied is used as the interface friction dissipation;

[0023] The energy dissipation of the spring body is calculated according to the total energy dissipation and the interface friction dissipation.

[0024] In some embodiments, performing local morphological analysis on the second force data and the second displacement data to obtain the first mechanical feature includes:

[0025] calculating, according to the second force data and the second displacement data, a fluctuation amplitude within a preset time window as a current noise level;

[0026] determining a mechanical characteristic threshold according to the current noise level;

[0027] calculating a second-order rate of change curve with respect to time based on the second force data and the second displacement data;

[0028] Identifying a transient peak in the second-order rate of change curve whose rate of change value is greater than the mechanical characteristic threshold;

[0029] The instantaneous peak is used as the first mechanical feature.

[0030] In some embodiments, performing performance testing based on the interface friction dissipation, the spring body energy dissipation, and the static friction coefficient to obtain a performance testing result includes:

[0031] generating time series data of the target calculation parameters according to target calculation parameters and historical data corresponding to the target calculation parameters, wherein the target calculation parameters include the interface friction dissipation, the spring body energy dissipation, or the static friction coefficient;

[0032] Performing trend analysis on the time series data of the target calculation parameter to obtain the corresponding change rate;

[0033] Determine whether a change rate of a target dissipation is greater than a preset warning threshold, and obtain a warning judgment result, wherein the target dissipation includes the interface friction dissipation or the spring body energy dissipation;

[0034] Determining whether the target dissipation is greater than a preset failure threshold, and obtaining a failure judgment result;

[0035] If the interface friction dissipation is greater than a preset interface friction threshold and the rate of change of the static friction coefficient is greater than 0, the support fixture interface is considered abnormal as the fixture judgment result; otherwise, the support fixture interface is considered normal as the fixture judgment result;

[0036] The performance test result is generated according to the early warning judgment result, the failure judgment result and the fixture judgment result.

[0037] In some embodiments, performing trend analysis on the time series data of the target calculation parameter to obtain the corresponding rate of change includes:

[0038] performing a noise level evaluation on the time series data of the target calculation parameter to obtain a noise level;

[0039] Calculating a data smoothing parameter based on the noise level;

[0040] Smoothing the time series data of the target calculation parameter according to the data smoothing parameter to obtain smoothed series data;

[0041] Identifying target key parameters of trend changes based on the smoothed sequence data, wherein the target key parameters include key points or key intervals;

[0042] Calculating a local change slope of the target calculation parameter according to the target key parameter;

[0043] The consistency of the local change slope is verified to obtain the change rate.

[0044] In some embodiments, performing noise level assessment on the time series data of the target calculation parameter to obtain the noise level includes:

[0045] Performing statistical analysis on the time series data of the target calculation parameters to obtain statistical analysis results;

[0046] Calculate the fluctuation range and dispersion degree according to the statistical analysis results;

[0047] The noise level is determined according to the fluctuation range and the dispersion degree.

[0048] In some embodiments, identifying the target key parameter of trend change based on the smoothed sequence data includes:

[0049] Calculating a first difference between every two adjacent data points in the smoothed sequence data to obtain a first difference sequence;

[0050] Calculating a second difference between every two adjacent first differences in the first difference sequence to obtain a second difference sequence;

[0051] Fitting and deriving the second difference sequence to obtain a derivative curve;

[0052] Taking a change point in the derivative curve whose value is greater than a preset change threshold as the key point;

[0053] The change interval in the derivative curve where the value is greater than a preset change threshold is used as the key interval.

[0054] In some embodiments, calculating the local change slope of the target calculation parameter according to the target key parameter includes:

[0055] Fitting the smoothed sequence data to obtain a fitting curve;

[0056] Calculating the residual between the smoothed sequence data and the fitting curve according to the target key parameter;

[0057] Determining the weight of each data point in the smoothed sequence data according to the residual;

[0058] The local change slope is calculated according to each data point in the smoothed sequence data and its corresponding weight.

[0059] On the other hand, an embodiment of the present invention provides a super-precision spring performance monitoring system, comprising:

[0060] a static friction state determination module, configured to determine the static friction state of the contact surface between the super-precision spring and the support fixture after applying a compressive force to the super-precision spring;

[0061] an energy dissipation separation module, configured to calculate interface friction dissipation and spring body energy dissipation after performing axial cyclic loading on the super-precision spring if the static friction state is abnormal lubrication of the fixture contact surface;

[0062] a static friction coefficient acquisition module, configured to acquire the static friction coefficient after pausing the axial cyclic loading and reapplying the compressive force;

[0063] The performance detection module is used to perform performance detection based on the interface friction dissipation, the spring body energy dissipation and the static friction coefficient to obtain a performance detection result.

[0064] The embodiments of the present application include at least the following beneficial effects: the embodiments of the present application first determine the static friction state of the contact surface between the super-precision spring and the support fixture after applying a compressive force to the super-precision spring. If the static friction state is abnormal lubrication of the fixture contact surface, then after performing axial cyclic loading on the super-precision spring, the interface friction dissipation and the energy dissipation of the spring body are calculated. Then, after pausing the axial cyclic loading and applying the compressive force again, the static friction coefficient is obtained. Then, based on the interface friction dissipation, the energy dissipation of the spring body and the static friction coefficient, performance testing is performed to obtain performance test results, thereby enabling super-precision spring performance monitoring to be achieved by identifying and eliminating the interface friction dissipation caused by static friction, thereby improving accuracy.

[0065] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the structures particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0067] Figure 1 This is a flow chart of a method for monitoring the performance of a super-precision spring according to an embodiment of the present invention;

[0068] Figure 2 The present invention is a schematic structural diagram of a super-precision spring performance monitoring system. DETAILED DESCRIPTION

[0069] In order to make the objectives, technical solutions, and advantages of this application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only intended to illustrate this application and are not intended to limit this application. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements.

[0070] Before explaining the embodiments of the present application in detail, some of the nouns and terms involved in the embodiments of the present application are first explained. The nouns and terms involved in the embodiments of the present application are subject to the following explanations.

[0071] Ultra-precision springs are elastic components with a wire diameter in the micron range (usually 0.06-1.6mm) and ultra-high manufacturing precision (tolerance can be controlled within ±0.01mm). Their core characteristics include extremely small dimensional tolerances, highly stable force output, and excellent fatigue resistance.

[0072] In the related art, conventional ultra-precision spring performance monitoring systems use energy dissipation calculation logic to attribute all measured energy dissipation to the intrinsic damping of the spring material during cyclic loading. When the support fixture contact surface experiences abnormal friction characteristics due to microscopic topography changes and induces microscopic slip, this slip generates additional energy dissipation, which is captured in the data by force and displacement sensors. However, existing systems lack the ability to identify energy dissipation caused by slip at the fixture interface, resulting in a systematic amplification of the calculated total energy dissipation value. This amplification effect causes the system to incorrectly assess the performance status of the ultra-precision spring, thereby affecting the accuracy and reliability of the monitoring results.

[0073] For example, during the quality control phase of a super-precision spring production, the testing department introduced a highly volatile organic cleaning agent for standardized cleaning of the contact surfaces of the support fixture before testing. This cleaning agent inadvertently dissolved the existing solid lubricant film on the fixture surface, resulting in direct metal-to-metal contact between the metal end face of the super-precision spring and the metal base of the support fixture during cyclic loading. This direct contact caused localized cold welding and tearing at the microscopic scale, resulting in irregular micromorphology and a significant increase in the static friction coefficient of the contact surface. During each loading cycle, when the loading force reached a critical point, the accumulated tangential stress between the spring end face and the fixture contact surface suddenly overcame the increased static friction, resulting in a very small, sudden slip, known as "stick-slip." This microscopic slip consumes some additional energy through frictional work and manifests as an additional closed region on the force-displacement curve. The energy dissipation calculation logic built into the existing monitoring system is designed to attribute all energy dissipation to the intrinsic damping of the spring material. Therefore, it is unable to distinguish the additional energy dissipation caused by the sliding of the fixture interface and simply adds it to the energy dissipation of the spring itself, resulting in the total energy dissipation value being systematically amplified.

[0074] If these issues are not addressed, the monitoring system will continue to mistakenly judge the performance of the super-precision springs as deteriorating based on the excessively high energy dissipation values, and may even issue an alarm indicating that the springs are approaching the failure threshold. This will force users to scrap super-precision springs that may actually be intact, resulting in financial losses and wasted resources. Furthermore, conflicts between the online monitoring system and offline manual review results will severely undermine user confidence in the reliability of the online monitoring system, challenging the decision-making basis of the entire quality control process and making it impossible to effectively guide production and maintenance.

[0075] When faced with the above problems, the first thing that comes to mind for this application is that the root of the problem lies in the additional energy dissipation introduced by the friction at the fixture interface, and the influence of this friction can be directly measured or eliminated in some way. However, it is technically challenging to accurately measure the friction of the microscopic interface directly during cyclic loading and separate it from the force of the spring body, and it is difficult to achieve without changing the main body of the existing test process. In this regard, this application further considers that when it is impossible to directly eliminate or accurately measure it, it can be distinguished from the source of energy dissipation. The key is to identify the state of abnormal interface friction, and in this state, try to decompose the total energy dissipation into interface friction dissipation and spring body energy dissipation. At the same time, in order to verify the accuracy of this separation, a parameter that can quantify the interface friction state is introduced, such as the static friction coefficient. Finally, these separated energy dissipations and static friction coefficients are combined for performance testing to obtain a more accurate performance evaluation result.

[0076] The following is a detailed explanation of the embodiments of the present application with reference to the accompanying drawings:

[0077] Figure 1 This is an optional flow chart of a method for monitoring the performance of a super-precision spring provided in an embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S104.

[0078] Step S101: After applying a compressive force to the super-precision spring, determining the static friction state of the contact surface between the super-precision spring and the support fixture;

[0079] Step S102: If the static friction state is abnormal lubrication of the fixture contact surface, then after performing axial cyclic loading on the super-precision spring, calculate the interface friction dissipation and the spring body energy dissipation;

[0080] Step S103: After pausing the axial cyclic loading and applying the compressive force again, obtaining the static friction coefficient;

[0081] Step S104: Perform performance testing based on the interface friction dissipation, the spring body energy dissipation, and the static friction coefficient to obtain a performance testing result.

[0082] Steps S101 to S104 shown in the embodiment of the present application can realize super-precision spring performance monitoring by identifying and removing interface friction dissipation caused by static friction, thereby improving accuracy.

[0083] In some embodiments, steps S101 to S104 accurately separate interface friction dissipation and spring body energy dissipation by identifying the static friction state of the contact surface, and introduce the static friction coefficient as an auxiliary judgment basis, thereby solving the problem of incorrect spring performance evaluation caused by the inability to distinguish the additional dissipation caused by interface sliding in the prior art, and achieving the effect of improving the accuracy and reliability of performance monitoring.

[0084] After applying a compressive force to the super-precision spring, the static friction state of the contact surface between the super-precision spring and the support fixture can be determined to provide a preliminary assessment of the lubrication condition of the contact surface. Abnormal contact surface lubrication can be the root cause of misjudgment of energy dissipation. If the static friction state is due to abnormal fixture contact surface lubrication, the interface friction dissipation and the energy dissipation of the spring itself are calculated after axial cyclic loading of the super-precision spring. This ensures that the true performance loss of the spring itself can be independently assessed, avoiding interference from external friction. After pausing the axial cyclic loading and reapplying the compressive force, the static friction coefficient is obtained. This coefficient serves as a supplementary indicator to quantify the friction characteristics of the contact surface and further verifies the degree of abnormality in the interface state. Performance testing is then performed based on the interface friction dissipation, the spring body energy dissipation, and the static friction coefficient to obtain the performance test results. This multi-dimensional, step-by-step testing process enables the system to comprehensively and accurately determine the performance condition of the super-precision spring, effectively avoiding misjudgments caused by friction dissipation confusion in traditional methods and ensuring the reliability of the monitoring results.

[0085] To more clearly illustrate this technical solution, a specific example is provided below. Before testing a super-precision spring, a preset compressive force is applied to the spring using a test system consisting of a force sensor and a displacement sensor. This system collects force and displacement data in real time, and a data processing unit analyzes this data, for example by monitoring the force-displacement curve for a sudden drop in force or a jump in displacement during the initial loading phase, to determine the static friction state of the contact surface between the super-precision spring and the support fixture. If the data processing unit determines that the static friction state is due to abnormal lubrication of the fixture contact surface, such as when the starting force required to overcome static friction exceeds the normal range, an axial cyclic loading procedure is initiated. During the cyclic loading process, the test system continuously collects force and displacement data and transmits this data to an energy analysis module. This module uses an energy separation algorithm, for example, by comparing the total energy dissipated with and without specific friction suppression measures to calculate and separate the interfacial friction dissipation and the energy dissipated in the spring body. After the cyclic loading process is completed and paused, the system reapplies the compressive force, and the static friction coefficient measurement unit obtains the current static friction coefficient, for example, by measuring the ratio of the minimum force required to overcome static friction to the normal force. Finally, the calculated interface friction dissipation, spring body energy dissipation, and static friction coefficient are comprehensively analyzed and compared with preset health thresholds and change trends to generate performance test results for the super-precision spring. These results are displayed on the user interface to guide subsequent decision-making.

[0086] Through the above-mentioned technical solution, this embodiment can accurately separate the actual energy dissipation of the spring body from the total energy dissipation measurement data, effectively avoiding the problem of incorrect spring performance status assessment caused by the inability to distinguish the additional dissipation caused by interface sliding in the existing technology. By introducing the judgment of static friction status and the acquisition of static friction coefficient, combined with the separate calculation of interface friction dissipation and spring body energy dissipation, this embodiment provides a multi-dimensional performance evaluation basis, thereby significantly improving the accuracy and reliability of ultra-precision spring performance monitoring results, avoiding the misjudgment and scrapping of intact springs, and reducing production costs and resource waste.

[0087] In some embodiments, in step S101, determining the static friction state of the contact surface between the super-precision spring and the support fixture may include but is not limited to the following steps:

[0088] collecting first force data and first displacement data;

[0089] generating a force-time relationship curve based on the first force data and the first displacement data;

[0090] If there is a mutation point in the force-time relationship curve and the starting force value required to overcome the static friction is greater than the preset healthy lubrication state threshold, the static friction state is determined to be abnormal lubrication of the fixture contact surface; otherwise, the static friction state is determined to be normal lubrication of the fixture contact surface.

[0091] In some cases, simple friction force measurements alone may not accurately determine the lubrication status. This is because the actual friction process is affected by multiple factors, such as surface roughness, contact pressure, and environmental conditions. These factors can cause fluctuations in friction force measurements, thus affecting the accuracy of the determination. Therefore, it is necessary to more accurately determine the static friction state of the contact surface between the super-precision spring and the support fixture to improve the reliability of subsequent performance testing.

[0092] First, the first force and displacement data are collected. Based on these data, a force-time curve is generated. This curve depicts the dynamic evolution of the friction force between the contact surface of the super-precision spring and the support fixture under compression. The force-time curve is then analyzed to identify any sudden changes. If a sudden change exists in the force-time curve and the starting force required to overcome static friction exceeds a preset healthy lubrication threshold, the static friction state is determined to indicate abnormal lubrication of the fixture contact surface. Otherwise, the static friction state is determined to be normal. The presence of a sudden change is often a sign of stick-slip, a phenomenon that is a hallmark of degraded friction characteristics of the contact surface caused by lubrication abnormalities. The starting force required to overcome static friction is also calculated and compared to a preset healthy lubrication threshold. Excessively high starting force values ​​indicate excessive frictional resistance on the contact surface, also indicating poor lubrication. By combining the dynamic characteristics of the curve (the sudden change) with the magnitude of the friction force (the starting force), a more accurate assessment of the static friction state of the contact surface between the super-precision spring and the support fixture can be achieved. This discrimination mechanism avoids the misjudgment that may be caused by a single indicator. For example, relying solely on the starting force value may ignore the existence of stick-slip phenomenon, and relying solely on the mutation point may not be able to quantify the magnitude of the friction force.

[0093] It is understood that the first force data refers to the sequence of force values ​​collected in real time by a force sensor when a compressive force is applied to the ultra-precision spring, potentially causing relative motion. This data can be acquired using a piezoelectric force sensor, a resistive strain gauge force sensor, or a capacitive force sensor. The first displacement data refers to the sequence of relative displacement values ​​of the ultra-precision spring or fixture collected in real time by a displacement sensor during the same process. This data can be acquired using a linear variable differential transformer (LVDT), an optical encoder, or a laser displacement sensor. The force-time relationship curve is a graphical representation plotted with the collected first force data as the vertical axis and the corresponding time data as the horizontal axis. Its purpose is to show the trend of force changes over time.

[0094] To more clearly illustrate this technical solution, a specific example is provided below. During the application of compression force to a super-precision spring, a high-precision force sensor can be used to collect first force data in real time, while a high-resolution displacement sensor can be used to simultaneously collect first displacement data. The analog signals output by these sensors can be converted to digital signals by a data acquisition module and transmitted to a data processing unit. This data processing unit can be an embedded industrial PC running specialized data processing software. After receiving the first force and first displacement data, the software can correlate the force data with time based on timestamps, thereby generating and displaying a force-time curve in real time. To identify abrupt changes in the curve, the data processing software can employ a signal processing algorithm. For example, by calculating the second derivative of the force-time curve and setting a dynamic threshold, a sudden change point is identified when the second derivative exceeds the threshold within a very short period of time. Furthermore, the software can automatically identify the starting force required to overcome static friction from the force-time curve, typically the maximum force reached before the curve begins to slip. The preset healthy lubrication threshold can be determined by testing a properly lubricated super-precision spring, collecting the breakout force required to overcome static friction, and performing a statistical analysis of the breakout force (for example, calculating its mean and standard deviation, and setting an upper limit based on the statistical distribution). The range of the preset healthy lubrication threshold can vary depending on factors such as the type, size, and material properties of the monitored super-precision spring, as well as the material and design of the support fixture. Finally, the data processing unit performs a logical comparison with the preset healthy lubrication threshold based on the identified breakout point information and the calculated breakout force value. If the curve contains a breakout point and the breakout force value exceeds the preset threshold, the system determines that the fixture contact surface is lubricated abnormally; otherwise, it determines that lubrication is normal. This specific implementation ensures that the static friction state is determined to take into account both the absolute magnitude of the friction force and its dynamic behavior, thereby improving the accuracy of the judgment.

[0095] Through the above technical solution, this embodiment can more accurately determine the static friction state of the contact surface between the ultra-precision spring and the support fixture. By collecting the first force data and the first displacement data and generating a force-time relationship curve, the dynamic characteristics of the friction process can be captured. Combined with the identification of the mutation point in the force-time relationship curve and the comparison of the starting force value required to overcome static friction with the preset healthy lubrication state threshold, this embodiment can distinguish between normal lubrication and abnormal lubrication states. This comprehensive judgment mechanism avoids the fluctuations and misjudgments that may be caused by a single friction force measurement, improves the accuracy of static friction state judgment, and thus provides a more reliable basis for subsequent ultra-precision spring performance testing, avoiding the problem of misjudgment of spring performance due to abnormal fixture lubrication.

[0096] In some embodiments, in step S102, calculating the interface friction dissipation and the spring body energy dissipation may include but is not limited to the following steps:

[0097] Step S201, obtaining total energy dissipation;

[0098] Step S202: Determine whether the support fixture has been subjected to transverse micro-vibration, where the direction of the transverse micro-vibration is perpendicular to the direction of the axial cyclic loading and is used to suppress interface friction;

[0099] Step S203: When the support fixture has been subjected to lateral micro-vibration, collecting second force data and second displacement data;

[0100] Step S204: when the supporting fixture is not subjected to lateral micro-vibration, collecting third force data and third displacement data;

[0101] Step S205: Perform local morphological analysis on the second force data and the second displacement data to obtain a first mechanical feature;

[0102] Step S206: performing local morphological analysis on the third force data and the third displacement data to obtain a second mechanical feature;

[0103] Step S207: Compare the first mechanical characteristic and the second mechanical characteristic to obtain a comparison result;

[0104] Step S208: Based on the comparison result, the energy dissipation corresponding to the mechanical characteristics reduced when the lateral micro-vibration is applied is used as the interface friction dissipation;

[0105] Step S209: Calculate the energy dissipation of the spring body according to the total energy dissipation and the interface friction dissipation.

[0106] In some embodiments, during actual cyclic loading, total energy dissipation includes both interface friction dissipation and energy dissipation in the spring itself. If the fixture contact surface is not lubricated properly, the interface friction dissipation can affect the calculation of total energy dissipation, potentially leading to an inaccurate assessment of spring performance. To accurately calculate interface friction dissipation and spring body energy dissipation, total energy dissipation can be first obtained. This total energy dissipation is based on the precise measurement and calculation of force and displacement data during the cyclic loading process. For example, when applying axial cyclic loading to a super-precision spring, the system simultaneously utilizes force sensors and displacement sensors to capture force and displacement data for each loading cycle. The force sensor measures the force applied to the spring in real time, while the displacement sensor measures the deformation or displacement of the spring under load. The force and displacement data collected by these sensors are either continuous or discrete data points sampled at a high frequency. This raw data can be transmitted to the core data processing unit, where the energy dissipation calculation logic uses the force and displacement data to accurately calculate the energy dissipated by the spring during each cycle, i.e., the total energy dissipation. To determine whether the support fixture is subjected to lateral micro-vibration, the test process is divided into two operating conditions: one in which lateral micro-vibration is applied perpendicular to the axial cyclic loading direction to suppress interfacial friction; the other in which lateral micro-vibration is not applied. The lateral micro-vibration is perpendicular to the axial cyclic loading direction and is used to suppress interfacial friction. Data is collected under both operating conditions. When the support fixture is subjected to lateral micro-vibration, the second force data and the second displacement data are collected; when the support fixture is not subjected to lateral micro-vibration, the third force data and the third displacement data are collected.

[0107] Local morphological analysis is then performed on the second force and second displacement data to obtain the first mechanical characteristic, and local morphological analysis is performed on the third force and third displacement data to obtain the second mechanical characteristic. These mechanical characteristics can capture subtle differences in the system's mechanical response under different friction states. The first and second mechanical characteristics are then compared to obtain a comparison result, which quantifies the impact of lateral microvibration on mechanical behavior. Because lateral microvibration can effectively suppress interface friction, the energy dissipation corresponding to the mechanical characteristic reduced when applying lateral microvibration can be used as the interface friction dissipation based on the comparison results. Finally, the energy dissipation of the spring body is calculated based on the total energy dissipation and the interface friction dissipation. By subtracting the identified interface friction dissipation from the total energy dissipation, the energy dissipation of the spring body can be accurately calculated. This allows, during ultra-precision spring performance monitoring, even if the support fixture contact surface experiences abnormal friction characteristics and microslip due to micromorphological changes, the true energy dissipation of the spring body can be accurately separated from the total energy dissipation measurement data. This prevents the system from misassessing the spring performance status due to the inability to distinguish the additional dissipation caused by interface slip, thereby ensuring the accuracy and reliability of the monitoring results.

[0108] It is understood that lateral micro-vibration refers to a small-amplitude, high-frequency vibration whose direction is perpendicular to the axial cyclic loading direction of the spring. Its purpose is to destroy or suppress the static friction between the contact surfaces through continuous small perturbations, making it as close to a sliding friction state as possible, thereby reducing or eliminating the additional energy dissipation caused by the stick-slip effect. Specifically, this can be achieved through a piezoelectric actuator, electromagnetic drive, or mechanical vibration device. The first and second mechanical characteristics refer to quantitative indicators extracted from force and displacement data through local morphological analysis that can characterize the dynamic response or energy dissipation characteristics of the system under specific loading conditions. These characteristics may include, but are not limited to, the local slope of the force-displacement curve, the instantaneous energy dissipation rate, the amplitude response at a specific frequency, or the instantaneous spike in the second-order rate of change curve mentioned in the subsequent scheme. Its purpose is to provide a quantifiable basis for comparing the changes in the mechanical behavior of the system before and after the application of lateral micro-vibration.

[0109] To more clearly illustrate this technical solution, a specific example is provided below. After obtaining the total energy dissipation, a vibration control module can be configured to determine whether the support fixture has been subjected to transverse micro-vibration. This module controls a piezoelectric actuator or electromagnetic vibrator to apply transverse micro-vibration to the support fixture, with a frequency of 50 Hz to 200 Hz and an amplitude of 5 to 20 microns, as needed. The vibration direction is set perpendicular to the axial cyclic loading direction of the spring. When the vibration control module is activated, it is assumed that the support fixture has been subjected to transverse micro-vibration, and second force data and second displacement data are collected. When the vibration control module is inactive, third force data and third displacement data are collected. Furthermore, a local morphological analysis is performed on the second force data and the second displacement data to obtain a first mechanical characteristic. This may include calculating the fluctuation amplitude within a preset time window (e.g., a 0.1-second time window for each loading cycle) based on the collected second force data and the second displacement data, and using this as the current noise level. Based on this noise level, a mechanical characteristic threshold is determined. Subsequently, a second-order rate of change curve of the second force data and the second displacement data with respect to time is calculated. Identify instantaneous spikes in the second-order rate of change curve whose rate of change values ​​are greater than the mechanical feature threshold, and use these instantaneous spikes as the first mechanical feature. Similarly, perform the same local morphological analysis on the third force data and the third displacement data to obtain the second mechanical feature. Then, compare the first mechanical feature with the second mechanical feature. For example, the difference in the average amplitude or frequency of the instantaneous spikes before and after the application of lateral micro-vibration can be calculated to obtain a comparison result. Based on the comparison result, the energy dissipation corresponding to the reduced instantaneous spike amplitude or frequency when the lateral micro-vibration is applied is used as the interface friction dissipation. This can be achieved through a pre-established experimental calibration curve or model that associates the change in the mechanical feature with the numerical value of the energy dissipation. For example, the experimental calibration curve can be expressed as a two-dimensional rectangular coordinate system, where the horizontal axis represents the reduced instantaneous spike amplitude or frequency when the lateral micro-vibration is applied (e.g., expressed as a percentage reduction), and the vertical axis represents the corresponding interface friction dissipation (e.g., in joules). By conducting multiple experiments under different friction conditions (e.g., from well-lubricated to completely unlubricated, or varying degrees of wear) and precisely measuring the change in the transient spikes and the actual energy dissipation in each experiment, these data points can be plotted on a rectangular coordinate system. Regression analysis methods, such as linear regression, polynomial regression, or nonlinear regression, can then be used to fit these data points, yielding an experimental calibration curve in the form of a mathematical function. For example, this curve could be a simple linear function or a more complex exponential or logarithmic function to describe the relationship between the two. If the average amplitude of the transient spikes decreases by X%, the corresponding interface friction dissipation is Y joules.Finally, by subtracting the calculated interfacial friction dissipation from the total energy dissipation, we can obtain the exact value of the energy dissipation of the spring itself.

[0110] Through the above technical solution, this embodiment can accurately separate interface friction dissipation and spring body energy dissipation from total energy dissipation. This prevents the interference of interface friction dissipation caused by abnormal lubrication of the fixture contact surface on the spring body performance evaluation, thereby ensuring the accuracy of super-precision spring performance evaluation, providing a reliable data basis for spring maintenance and scrapping decisions, and effectively reducing the risk of misjudgment and unnecessary economic losses.

[0111] In some embodiments, in step S205, performing local morphological analysis on the second force data and the second displacement data to obtain the first mechanical feature may include but is not limited to the following steps:

[0112] calculating, based on the second force data and the second displacement data, a fluctuation amplitude within a preset time window as a current noise level;

[0113] Determine the mechanical feature threshold based on the current noise level;

[0114] calculating a second-order rate of change curve with respect to time based on the second force data and the second displacement data;

[0115] Identify the instantaneous peaks in the second-order rate of change curve whose rate of change is greater than the mechanical characteristic threshold;

[0116] The instantaneous spike is taken as the first mechanical feature.

[0117] In some embodiments, due to limitations of the test environment and the sensor itself, the collected second force data and second displacement data often contain noise. This noise can interfere with the accuracy of the local morphological analysis, resulting in the extracted first mechanical feature being unable to truly reflect the performance state of the spring, thereby affecting subsequent performance test results. To accurately extract the first mechanical feature, the fluctuation amplitude within a preset time window can be calculated based on the second force data and the second displacement data as the current noise level. This aims to accurately assess the interference intensity in the data and provide a basis for subsequent feature recognition. Noise usually manifests as rapid fluctuations in data over a short period of time. By calculating the fluctuation amplitude, the noise intensity of the current data can be effectively assessed.

[0118] A mechanical feature threshold is then determined based on the current noise level. This threshold is used to distinguish true mechanical features from noise interference. By associating the mechanical feature threshold with the noise level, adaptive noise filtering can be achieved. When the noise level is high, the mechanical feature threshold is increased accordingly, preventing noise from being misidentified as a mechanical feature. Conversely, when the noise level is low, the mechanical feature threshold is decreased, increasing sensitivity to subtle mechanical features. A second-order rate of change curve with respect to time is then calculated based on the second force and displacement data. This second-order rate of change curve can more sensitively reflect local changes in the mechanical data and highlight transient spikes. Compared to directly analyzing the raw data, the second-order rate of change curve can effectively amplify transient changes, making it easier to identify potential mechanical features. Finally, transient spikes in the second-order rate of change curve whose rate of change exceeds the mechanical feature threshold are identified and used as the first mechanical feature. By comparing the rate of change value with the mechanical feature threshold, noise interference can be effectively filtered out and meaningful transient spikes can be extracted. These transient spikes represent significant changes in the mechanical data and can reflect subtle behavior of the spring during cyclic loading.

[0119] It is understood that the amplitude of fluctuation refers to the degree to which the data value deviates from its mean or trend line within a specific time window. It can be measured using statistics such as standard deviation, root mean square error, peak-to-valley difference or interquartile range. Its purpose is to quantify the degree of discreteness or noise intensity of the data. The second-order rate of change curve refers to a curve that describes the speed of change of the data, that is, the result of taking the derivative of the data twice with respect to time. It can be calculated using mathematical methods such as the difference method, the least squares fitting and derivative, or the wavelet transform. Its purpose is to amplify the sudden changes and sharp features in the data. An instantaneous spike refers to a local maximum value in the second-order rate of change curve where the value is significantly higher than the surrounding level in a short period of time. It can be identified using methods such as peak detection algorithm, threshold comparison or morphological filtering. Its purpose is to locate sudden events or state transitions with important physical significance in the data.

[0120] To more clearly illustrate this technical solution, a specific example is provided below. First, after collecting the second force data and second displacement data, these time series data can be segmented, for example, with every 100 data points or every 0.1 second as a preset time window. Within each preset time window, the standard deviation or peak-to-valley difference of the data points can be calculated as the current noise level. For example, if within a window, the force data fluctuates from 10N to 10.5N and the displacement data fluctuates from 2mm to 2.05mm, the respective fluctuation amplitudes can be calculated, and the larger of the two or the weighted average can be taken as the current noise level for that window. Next, based on the calculated current noise level, a mechanical feature threshold can be dynamically determined. For example, the mechanical feature threshold can be set to three times the standard deviation of the current noise level, or a lookup table can be established based on historical data to map different noise levels to corresponding thresholds. In this way, when the noise level is high, the threshold is automatically increased to avoid misidentification of noise as a mechanical feature; when the noise level is low, the threshold is lowered, thereby capturing more subtle true mechanical features. Subsequently, the second force data and the second displacement data can be numerically differentiated to calculate a second-order rate of change curve with respect to time. For example, a central difference method or Savitzky-Golay filter can be used for smoothing and differentiation to obtain second-order derivative curves of the force and displacement data. These second-order derivative curves can highlight instantaneous acceleration or curvature changes in the data, thereby more clearly displaying potential mechanical events. Finally, a peak detection algorithm can be applied to the second-order rate of change curve to identify instantaneous spikes whose rate of change values ​​are greater than a determined mechanical characteristic threshold. For example, the second-order rate of change curve can be traversed, and when the value of a data point exceeds the mechanical characteristic threshold and the values ​​in its left and right neighborhoods are all lower than this point, the point is marked as a transient spike. These identified transient spikes, such as a sudden drop or increase in force in the force-time curve, or a sudden change in displacement velocity in the displacement-time curve, can be used as the first mechanical feature for subsequent performance evaluation.

[0121] The above technical solution effectively addresses noise interference in the data when performing local morphological analysis on the second force data and the second displacement data. By adaptively evaluating the noise level and determining the mechanical feature threshold, it is possible to avoid misjudging noise as a true mechanical feature. At the same time, the use of a second-order rate of change curve can more sensitively capture instantaneous changes in the data, thereby accurately identifying instantaneous peaks representing the spring's true behavior against a noisy background. This makes the extracted first mechanical feature more accurate and reliable, providing more precise input for subsequent calculations of interface friction dissipation and spring body energy dissipation, thereby improving the accuracy of ultra-precision spring performance monitoring.

[0122] In some embodiments, in step S104, a performance test is performed based on the interface friction dissipation, the spring body energy dissipation, and the static friction coefficient to obtain a performance test result, which may include but is not limited to the following steps:

[0123] Step S301: Generate time series data of the target calculation parameters according to the target calculation parameters and historical data corresponding to the target calculation parameters, where the target calculation parameters include interface friction dissipation, spring body energy dissipation, or static friction coefficient;

[0124] Step S302: Perform trend analysis on the time series data of the target calculation parameter to obtain the corresponding change rate;

[0125] Step S303: determining whether the rate of change of the target dissipation is greater than a preset warning threshold, and obtaining a warning determination result, wherein the target dissipation includes interface friction dissipation or spring body energy dissipation;

[0126] Step S304: determine whether the target dissipation is greater than a preset failure threshold, and obtain a failure determination result;

[0127] Step S305: If the interface friction dissipation is greater than the preset interface friction threshold and the rate of change of the static friction coefficient is greater than 0, the support fixture interface is considered abnormal as the fixture judgment result; otherwise, the support fixture interface is considered normal as the fixture judgment result;

[0128] Step S306: Generate a performance test result based on the early warning judgment result, the failure judgment result, and the fixture judgment result.

[0129] In some embodiments, since a single performance parameter may be affected by multiple factors, using these parameters alone for judgment may lead to misjudgment. For example, judging that the fixture is abnormal based solely on the fact that the interface friction dissipation exceeds a threshold value may ignore the changing trend of the static friction coefficient, resulting in inaccurate judgment. In order to improve the accuracy of performance testing, time series data of the target calculation parameters can be generated based on the target calculation parameters and the historical data corresponding to the target calculation parameters. The construction of this time series allows isolated performance parameter values ​​to be placed in a time dimension for examination. By analyzing their historical changing trends, early signs of performance degradation can be more accurately identified, avoiding misjudgments caused by relying solely on current values. Among them, the target calculation parameters include interface friction dissipation, spring body energy dissipation, or static friction coefficient.

[0130] Then, trend analysis is performed on the time series data of the target calculation parameters to quantify the rate of change of each parameter over time and obtain the corresponding rate of change. Obtaining this rate of change can reflect the speed of performance degradation and provide a more precise basis for subsequent early warning and failure judgment. For example, even if the current value of a parameter has not yet reached the failure threshold, if its rate of change has increased significantly, it may indicate a potential risk. The early warning judgment result is obtained by determining whether the rate of change of the target dissipation is greater than the preset early warning threshold. This enables the system to issue a timely early warning when the rate of change of the performance parameter exceeds a certain level, alerting the operator to potential performance issues and thus avoiding sudden failure. At the same time, the system also determines whether the current value of the target dissipation is greater than the preset failure threshold to generate a failure judgment result. This ensures that when the performance parameter reaches a critical value, a failure alarm can be issued, indicating that the spring needs to be replaced or repaired, thereby avoiding safety risks caused by performance degradation. The target dissipation includes interface friction dissipation or energy dissipation of the spring body.

[0131] The system then determines whether the target dissipation exceeds the preset failure threshold, generating a failure judgment result. If the interface friction dissipation exceeds the preset interface friction threshold and the rate of change of the static friction coefficient is greater than 0, the support fixture interface is considered abnormal as the fixture judgment result; otherwise, the support fixture interface is considered normal as the fixture judgment result. This multi-condition joint judgment mechanism can more accurately diagnose the fixture interface status and avoid misjudgments caused by single parameter anomalies. For example, even if the interface friction dissipation exceeds the threshold, if the static friction coefficient does not increase significantly, it may not be due to an abnormal fixture interface, but rather other factors, thus avoiding the incorrect scrapping of the spring body. Finally, performance test results are generated based on the early warning judgment results, failure judgment results, and fixture judgment results. This comprehensive report provides operators with comprehensive and accurate performance evaluation information, enabling them to make decisions based on more reliable data, thereby improving the accuracy and reliability of ultra-precision spring performance monitoring.

[0132] It is understood that time series data refers to a data set formed by chronologically arranging the measured values ​​of a target calculation parameter at different time points. This data can be acquired through continuous or periodic sampling. Trend analysis refers to the mathematical or statistical processing of time series data to identify the long-term direction or pattern of data change over time. This can be achieved using linear regression, moving average, exponential smoothing, or more complex machine learning algorithms. The rate of change refers to the speed at which the value of the target calculation parameter changes over a specific time period. This can be obtained by calculating the derivative, slope, or difference of the time series data.

[0133] To more clearly illustrate this technical solution, a specific example is provided below. First, during the performance monitoring of a super-precision spring, a data acquisition and processing unit can continuously receive data from force and displacement sensors and, based on this data, calculate interface friction dissipation, spring body energy dissipation, and the coefficient of static friction. These calculated parameter values ​​are stored in real time in a historical database. For example, the current parameter values ​​can be recorded along with a timestamp at regular intervals or after a certain number of loading cycles, gradually building up a time series of these target calculated parameters.

[0134] The built-in analysis module then performs trend analysis on these time series data. Specifically, this module uses a sliding window averaging algorithm to smooth the time series data to eliminate the effects of transient noise, resulting in smoothed series data. Next, by calculating the linear regression slope of the smoothed series data within a specific time window, the corresponding rate of change can be determined. For example, the average slope of change of interface friction dissipation, spring body energy dissipation, and static friction coefficient can be calculated over the past 24 hours or the past 1000 loading cycles. Based on this, the system performs multiple judgments. For example, for early warning judgments, a preset warning threshold can be set. If the rate of change of the spring body energy dissipation exceeds this threshold for three consecutive measurements, an early warning result is generated, indicating that the spring performance may be degrading at an accelerated rate. The preset warning threshold can be directly set based on expert experience or determined based on statistical analysis of historical data distribution. The threshold is determined using statistical methods (such as mean and standard deviation) based on historical data of the target dissipation rate of change under normal operating conditions or known performance degradation patterns. Furthermore, a risk assessment model can be used to determine the optimal threshold, taking into account the risk level of performance degradation, the cost of failure, and the timeliness of early warning. This risk assessment model can utilize a deep learning model architecture. For failure diagnosis, a preset failure threshold can be set. If the current value of interface friction dissipation or spring body energy dissipation exceeds this threshold, a failure determination is generated. This preset failure threshold can be obtained by collecting a large amount of performance data from super-precision springs in actual operation or accelerated aging testing over a long period of time and statistically analyzing the energy dissipation values ​​before failure, combined with their final failure scenarios. For example, the energy dissipation trend and final value of a failed spring over a period of time before failure can be analyzed to determine a critical value that effectively distinguishes normal operation from a failure state. Alternatively, the threshold can be determined based on design specifications and standards. Design standards for super-precision springs typically include life expectancy and performance degradation limits, which serve as an important basis for setting the preset failure threshold. To more accurately diagnose the fixture interface condition, the system can set a preset interface friction threshold. If the current value of the interface friction dissipation is greater than the preset interface friction threshold, and the rate of change of the static friction coefficient is detected to be greater than 0 (for example, the static friction coefficient shows a trend of continuous increase), the system will determine that the support fixture interface is abnormal and generate a corresponding fixture judgment result. On the contrary, if the interface friction dissipation does not exceed the threshold, or the rate of change of the static friction coefficient does not show an increasing trend, the fixture interface is judged to be normal. Among them, the preset interface friction threshold can be determined according to the support fixture design specifications. The support fixture design specifications usually take into account the friction characteristics of the contact surface with the spring, which can be used as the basis for setting the threshold. It can also be determined based on expert experience.

[0135] Ultimately, all judgment results, including warning, failure, and fixture judgment results, are integrated into a comprehensive performance test report. This report can be output in a structured data format and visualized through the user interface. For example, it can display clear diagnostic information such as "Spring performance is normal, fixture interface is abnormal, fixture inspection recommended" or "Spring performance warning, accelerating rate of change, continued attention recommended," providing operators with clear decision-making basis.

[0136] Through the above technical solution, this embodiment can upgrade the performance detection of ultra-precision springs from static threshold judgment to dynamic time series analysis and trend judgment, so that the system can identify early signs of performance degradation earlier and avoid misjudgment caused by relying solely on instantaneous values. By analyzing the rate of change of interface friction dissipation, spring body energy dissipation and static friction coefficient, potential failure risks can be predicted more accurately and early warnings can be issued in a timely manner. In addition, combining the current value of interface friction dissipation with the changing trend of static friction coefficient, it is possible to accurately distinguish between performance problems of the spring body and abnormalities of the support fixture interface, thereby avoiding misjudgment and unnecessary scrapping of springs due to fixture problems. Ultimately, the generated performance test results are more comprehensive and reliable, providing an accurate decision-making basis for the maintenance and replacement of ultra-precision springs and improving the diagnostic capabilities of the monitoring system.

[0137] In some embodiments, in step S302, trend analysis is performed on the time series data of the target calculation parameter to obtain the corresponding change rate, which may include but is not limited to the following steps:

[0138] Step S401: performing noise level evaluation on the time series data of the target calculation parameter to obtain the noise level;

[0139] Step S402: Calculate data smoothing parameters according to the noise level;

[0140] Step S403: Smoothing the time series data of the target calculation parameter according to the data smoothing parameter to obtain smoothed series data;

[0141] Step S404: Identify target key parameters of trend changes based on the smoothed sequence data, where the target key parameters include key points or key intervals;

[0142] Step S405: Calculate the local change slope of the target calculation parameter according to the target key parameter;

[0143] Step S406: perform consistency verification on the local change slope to obtain the change rate.

[0144] In some embodiments, simply analyzing the raw time series data directly can be susceptible to noise interference, resulting in inaccurate trend analysis results and an inability to accurately reflect performance changes. To improve the accuracy of trend analysis, the time series data of the target calculation parameter can first be evaluated for noise levels to obtain a noise level. This allows the system to quantify the random fluctuations present in the data. A data smoothing parameter is then calculated based on the noise level. This ensures that the smoothing process effectively suppresses noise while maximally preserving the true trend information of the raw data, avoiding the problems of over-smoothing or under-smoothing. The time series data of the target calculation parameter is smoothed based on the data smoothing parameter to obtain smoothed sequence data. This significantly reduces the interference of noise on trend analysis and allows the underlying trends in the data to be clearly presented. Based on the smoothed sequence data, target key parameters of trend change are identified. These target key parameters include key points or key intervals, allowing trend analysis to focus on the most representative areas of change in the data. Based on the target key parameters, a local slope of change of the target calculation parameter is calculated. This provides a quantitative description of the intensity of trend change within a specific area and better reflects actual performance changes than a single global slope. Finally, the consistency of the local change slope is verified to obtain the change rate, which avoids misjudgment caused by local anomalies or accidental fluctuations and ensures that the final change rate can accurately reflect the actual performance changes of the target calculation parameters.

[0145] It is understood that the noise level refers to a numerical value or indicator obtained through noise level assessment that reflects the intensity of random fluctuations in the data. It can be a standard deviation, variance, signal-to-noise ratio, or energy value within a specific frequency range. The data smoothing parameter refers to a configuration item used to control the degree of data smoothing. It can be a smoothing window size, smoothing coefficient, or number of iterations. Its purpose is to guide the execution of the smoothing algorithm. Smoothing refers to the process of reducing noise in time series data by applying specific algorithms to reveal potential trends. It can be achieved by using methods such as moving average, exponential smoothing, Savitzky-Golay filtering, or wavelet denoising. Its purpose is to remove random fluctuations in the data. The local change slope refers to the instantaneous or average rate of change of the time series data trend near the target key parameter. It can be calculated using local linear fitting, difference calculation, or spline interpolation-based methods. Its purpose is to quantify the intensity of change in a specific area.

[0146] In order to more clearly illustrate the technical solution, a specific example is used for explanation below. When evaluating the noise level of the time series data of the target calculation parameter, the standard deviation of the time series data within the sliding window can be calculated, and the average value of the standard deviation within the window can be used as the noise level. For example, a sliding window of fixed length can be set, the standard deviation of the data points in each window can be calculated, and then the average value of these standard deviations can be used as the noise level of the current time series. When calculating the data smoothing parameter according to the noise level, a preset mapping relationship or lookup table can be established, and the parameters of the smoothing algorithm can be dynamically adjusted according to the evaluated noise level. For example, when the noise level is high, the window size of the moving average or the attenuation coefficient of the exponential smoothing can be increased to enhance the smoothing effect; when the noise level is low, these parameters can be reduced to retain more details. When smoothing the time series data according to the data smoothing parameter, the exponential weighted moving average method can be used, in which the smoothing parameter determines the weight of the influence of historical data on the current smoothing value, so that the recent data contributes more to the trend. When identifying target key parameters for trend changes based on smoothed sequence data, the first and second derivatives of the smoothed sequence data can be calculated. Key points can be identified by detecting the zero crossing points of the second-order derivative or the local extreme points of the first-order derivative, or key intervals can be identified by analyzing the trend of the first-order derivative changes of multiple consecutive data points. For example, when the second-order derivative changes from positive to negative or from negative to positive, it may indicate a turning point in a trend. When calculating the local change slope of the target calculation parameter based on the target key parameter, a linear regression fit can be performed using the least squares method within a preset data window near each identified key point or key interval, and the slope of the fitted line is used as the local change slope of the area. When verifying the consistency of the local change slope, a threshold range for the change slope can be set. If the calculated local change slope falls outside the threshold range, further data review or more complex statistical methods may be required to ensure its representativeness.

[0147] Through the above technical solution, this embodiment effectively reduces the interference of noise in the original time series data on trend analysis, allowing the calculated rate of change to more accurately reflect the actual performance changes of the target calculation parameter. This avoids misjudgments caused by data fluctuations, improves the accuracy and reliability of performance monitoring, and thus provides more refined and stable data support for the performance evaluation of super-precision springs.

[0148] In some embodiments, in step S401, performing noise level assessment on the time series data of the target calculation parameter to obtain the noise level may include but is not limited to the following steps:

[0149] Perform statistical analysis on the time series data of the target calculation parameters to obtain statistical analysis results;

[0150] Calculate the fluctuation range and dispersion degree based on the statistical analysis results;

[0151] Determine the noise level based on the fluctuation range and dispersion.

[0152] In some embodiments, relying solely on a single mean absolute deviation may not fully capture the complex noise characteristics of time series data. For example, for data with abnormal spikes or slow drift, this simple assessment method may lead to inaccurate noise level assessment, which in turn affects subsequent data smoothing and trend analysis, making the final performance test results unreliable. To improve the accuracy of noise level assessment, a statistical analysis can be performed on the time series data of the target calculation parameter to obtain statistical analysis results. This can provide in-depth information about the data distribution, such as the data's central tendency, distribution pattern, and potential outliers. Based on the statistical analysis results, the fluctuation range and dispersion are then calculated. The fluctuation range can intuitively reflect the maximum fluctuation amplitude of the data on the time axis, while the dispersion quantifies the closeness of the data points around their mean. These two indicators characterize the characteristics of random fluctuations in the data from different dimensions, avoiding the assessment bias that may be caused by a single indicator. The noise level is then determined based on the fluctuation range and dispersion to improve the accuracy of the noise level assessment.

[0153] It can be understood that statistical analysis results refer to quantitative information reflecting the overall characteristics of time series data obtained after preliminary processing. This can be achieved by calculating statistical quantities such as the mean, median, mode, quartiles, skewness, and kurtosis. Fluctuation range refers to the magnitude of the change in the values ​​of time series data within a specific time period. This can be achieved by calculating the difference between the maximum and minimum values, the range, or the difference between specific percentiles (such as the difference between the 90th and 10th percentiles). Dispersion refers to the degree to which each data point in the time series data deviates from the central trend. This can be achieved by calculating the standard deviation, variance, mean absolute deviation, or coefficient of variation.

[0154] To more clearly illustrate this technical solution, a specific example is provided below. First, a statistical analysis is performed on the time series data of the target calculation parameter. For example, the mean, standard deviation, and quartiles (Q1, Q3) of the time series data can be calculated. These statistics constitute the statistical analysis results. Next, based on these statistical analysis results, the fluctuation range and dispersion are calculated. The fluctuation range can be specifically calculated as the data range (the difference between the maximum and minimum values); the dispersion can be specifically calculated as the data standard deviation. Finally, the noise level is determined based on the calculated range and standard deviation. For example, a weighted formula based on the range and standard deviation can be set, or a decision tree model can be constructed that uses the range and standard deviation as input features and outputs a quantified noise level value. This noise level value can be a continuous value or a discrete level (such as low, medium, or high noise). This approach comprehensively considers the overall fluctuation range of the data and the dispersion between data points, resulting in a more refined and accurate noise level assessment.

[0155] Through the above technical solution, this embodiment enables a more accurate noise level assessment of the time series data of the target calculation parameter. By performing statistical analysis on the data and calculating the fluctuation range and dispersion based on this analysis, the noise characteristics in the data can be fully captured, avoiding the bias that may be caused by single-metric evaluation. This method can more effectively identify and quantify noise in the data, thereby providing a more reliable foundation for subsequent data smoothing and trend analysis, significantly improving the accuracy and reliability of performance test results.

[0156] In some embodiments, in step S404, identifying the target key parameter with trend change based on the smoothed sequence data may include but is not limited to the following steps:

[0157] Calculate the first difference between every two adjacent data points in the smoothed sequence data to obtain a first difference sequence;

[0158] Calculating a second difference between every two adjacent first differences in the first difference sequence to obtain a second difference sequence;

[0159] Fitting and deriving the second difference sequence to obtain a derivative curve;

[0160] The change point in the derivative curve whose value is greater than the preset change threshold is taken as the key point;

[0161] The change interval in the derivative curve where the value is greater than the preset change threshold is regarded as the key interval.

[0162] In some embodiments, to accurately and efficiently identify target key parameters from smoothed sequence data so that subsequent trend analysis can accurately reflect actual data changes, the first difference between every two adjacent data points in the smoothed sequence can be calculated. This can initially capture the magnitude and direction of local data changes, resulting in a first difference sequence. Then, the second difference between every two adjacent first differences in the first difference sequence is calculated to obtain a second difference sequence. This is equivalent to taking the derivative of the data's rate of change again, thereby quantifying the "acceleration" of the data change. To eliminate the impact of noise on this "acceleration" information, the second difference sequence can be fitted and differentiated to obtain a derivative curve. This derivative curve clearly displays the trend of the data's rate of change. Its peaks and valleys correspond to the points in the original data where the rate of change is greatest, often representing key locations where data trends shift. Finally, points in the derivative curve whose values ​​exceed a preset change threshold are designated as key points, and intervals within the derivative curve where the values ​​exceed the preset change threshold are designated as key intervals. This effectively screens out truly meaningful trend changes in the data and avoids misidentifying minor fluctuations as key changes. The preset change threshold can be determined by conducting extensive testing of super-precision springs in healthy and varying degrees of degradation in a controlled laboratory environment. During these tests, time series data of the target calculation parameters are collected and the derivative curve of the second difference series is calculated. By observing the numerical characteristics of the derivative curve when normal fluctuations and when the trend changes significantly, a preset change threshold that effectively distinguishes between these two states can be determined. Alternatively, mechanical and friction models of the super-precision spring can be simulated to simulate performance data at different degradation stages, and the corresponding derivative curves can be calculated. The preset change threshold can be determined by analyzing the simulation results.

[0163] It is understood that the first difference sequence refers to the set of differences between every two adjacent data points in the smoothed sequence data. It can be obtained by point-by-point subtraction, and its purpose is to preliminarily quantify the rate of change of the data in the time dimension. The second difference sequence refers to the set of differences between every two adjacent first differences in the first difference sequence. It can be obtained by further point-by-point subtraction of the first difference sequence, and its purpose is to further quantify the change in the data change rate, that is, the "acceleration" of the data.

[0164] Through the above technical solution, this embodiment can accurately and efficiently identify the target key parameters of trend changes from the smooth sequence data. By calculating the first difference and the second difference, the rate of change and acceleration information of the data can be captured step by step, which makes the capture of the data trend more precise. Furthermore, fitting and deriving the second difference sequence can effectively eliminate noise interference and obtain a smooth and accurate derivative curve, thereby clearly revealing the turning point and change intensity of the data trend. By setting a preset change threshold to screen key points and key intervals, it can be ensured that the identified parameters are truly meaningful trend changes, avoiding misjudgment of small fluctuations. This provides a more reliable and accurate basis for subsequent trend analysis, so that the evaluation of the performance of ultra-precision springs can accurately reflect the real changes in the data, thereby improving the accuracy of the monitoring results.

[0165] In some embodiments, in step S405, calculating the local change slope of the target calculation parameter according to the target key parameter may include but is not limited to the following steps:

[0166] Fit the smoothed sequence data to obtain a fitting curve;

[0167] Calculate the residual between the smoothed sequence data and the fitted curve based on the target key parameters;

[0168] Determine the weight of each data point in the smoothed sequence data based on the residual;

[0169] Calculate the local slope of change based on each data point in the smoothed series data and its corresponding weight.

[0170] In some embodiments, because the target key parameter identified through trend analysis solely on smoothed sequence data cannot accurately reflect the true fluctuations and noise effects of the original data, the fitted curve cannot accurately capture local variations, thereby affecting the accuracy of the local variation slope. To accurately calculate the local variation slope, the smoothed sequence data can first be fitted to obtain a fitted curve that reflects the overall trend of the data. Then, based on the target key parameter, the residual between the smoothed sequence data and the fitted curve is calculated. The residual calculation focuses on the region where the target key parameter is located, making the quantification of the degree of deviation of data points more targeted and able to capture local details related to the true fluctuations that remain after the smoothing process. It can be understood that the residual refers to the numerical difference between each data point in the smoothed sequence data and the fitted curve at the corresponding position. It can be specifically calculated as the actual value of the data point minus the predicted value of the fitted curve at that data point. Its purpose is to quantify the degree to which each data point deviates from the overall trend. The residual is then used to determine the weight of each data point in the smoothed sequence data. Data points with larger residuals indicate greater deviations from the overall trend and are therefore assigned higher weights, meaning that these points will play a greater role in the subsequent slope calculation. Finally, the local change slope is calculated based on each data point in the smoothed sequence data and its corresponding weight. Through this weighted calculation, key data points that better reflect the true local fluctuations and noise effects of the data are given higher priority, allowing the calculated local change slope to more accurately reflect the actual change trend of the data and avoid the problem of ignoring important local features due to simple fitting.

[0171] The above technical solution no longer relies solely on direct fitting of smoothed sequence data when calculating the local variation slope of the target calculation parameter. Instead, by introducing a residual and weighting mechanism, the calculation process fully considers the degree to which data points deviate from the overall trend. This allows the calculation of the local variation slope to more accurately reflect the true fluctuations and local characteristics contained in the smoothed sequence data, avoiding fitting bias caused by noise or local anomalies.

[0172] The beneficial effects of implementing the embodiments of the present invention include: the embodiments of the present application first determine the static friction state of the contact surface between the super-precision spring and the support fixture after applying a compressive force to the super-precision spring; if the static friction state is abnormal lubrication of the fixture contact surface, then after performing axial cyclic loading on the super-precision spring, the interface friction dissipation and the spring body energy dissipation are calculated; then, after pausing the axial cyclic loading and applying the compressive force again, the static friction coefficient is obtained; and then, based on the interface friction dissipation, the spring body energy dissipation and the static friction coefficient, performance testing is performed to obtain performance test results, thereby enabling super-precision spring performance monitoring to be achieved by identifying and eliminating the interface friction dissipation caused by static friction, thereby improving accuracy.

[0173] like Figure 2As shown, an embodiment of the present invention further provides a super-precision spring performance monitoring system, comprising:

[0174] a static friction state determining module 501 for determining the static friction state of the contact surface between the super-precision spring and the support fixture after applying a compressive force to the super-precision spring;

[0175] an energy dissipation separation module 502 for calculating interface friction dissipation and spring body energy dissipation after performing axial cyclic loading on the super-precision spring if the static friction state is abnormal lubrication of the fixture contact surface;

[0176] a static friction coefficient acquisition module 503, configured to acquire the static friction coefficient after pausing the axial cyclic loading and applying the compressive force again;

[0177] The performance detection module 504 is configured to perform a performance detection based on the interface friction dissipation, the spring body energy dissipation, and the static friction coefficient to obtain a performance detection result.

[0178] The contents of the above method embodiments are all applicable to the present system embodiments. The functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0179] The embodiments described in the embodiments of this application are intended to more clearly illustrate the technical solutions of the embodiments of this application and do not constitute a limitation on the technical solutions provided by the embodiments of this application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

Claims

1. A method for monitoring the performance of a super-precision spring, characterized in that: The following steps are involved: After applying a compressive force to the super-precision spring, determining a static friction state between a contact surface of the super-precision spring and a support fixture; If the static friction state is abnormal lubrication of the fixture contact surface, then after performing axial cyclic loading on the super-precision spring, calculate the interface friction dissipation and the energy dissipation of the spring body; After pausing the axial cyclic loading and reapplying the compressive force, obtaining the static friction coefficient; Performing a performance test based on the interface friction dissipation, the spring body energy dissipation, and the static friction coefficient to obtain a performance test result; Wherein, determining the static friction state of the contact surface between the super-precision spring and the supporting fixture includes: collecting first force data and first displacement data; generating a force-time relationship curve based on the first force data and the first displacement data; If there is a sudden change point in the force-time relationship curve and the starting force value required to overcome the static friction is greater than the preset healthy lubrication state threshold, the static friction state is determined to be abnormal lubrication of the fixture contact surface; otherwise, the static friction state is determined to be normal lubrication of the fixture contact surface; The calculation of interface friction dissipation and spring body energy dissipation includes: Get the total energy dissipation; determining whether the support fixture has been subjected to transverse micro-vibration, wherein the direction of the transverse micro-vibration is perpendicular to the direction of the axial cyclic loading, and the transverse micro-vibration is used to suppress interface friction; When the supporting fixture has been subjected to lateral micro-vibration, collecting second force data and second displacement data; When the supporting fixture is not subjected to lateral micro-vibration, collecting third force data and third displacement data; performing a local morphological analysis on the second force data and the second displacement data to obtain a first mechanical feature; performing a local morphological analysis on the third force data and the third displacement data to obtain a second mechanical feature; Comparing the first mechanical characteristic with the second mechanical characteristic to obtain a comparison result; According to the comparison result, the energy dissipation corresponding to the reduced mechanical characteristics when the lateral micro-vibration is applied is used as the interface friction dissipation; Calculating the energy dissipation of the spring body according to the total energy dissipation and the interface friction dissipation; The performing local morphological analysis on the second force data and the second displacement data to obtain the first mechanical feature includes: calculating, according to the second force data and the second displacement data, a fluctuation amplitude within a preset time window as a current noise level; determining a mechanical characteristic threshold according to the current noise level; calculating a second-order rate of change curve with respect to time based on the second force data and the second displacement data; Identifying a transient peak in the second-order rate of change curve whose rate of change value is greater than the mechanical characteristic threshold; Taking the instantaneous peak as the first mechanical feature; The performing of a performance test based on the interface friction dissipation, the energy dissipation of the spring body, and the static friction coefficient to obtain a performance test result includes: generating time series data of the target calculation parameters according to target calculation parameters and historical data corresponding to the target calculation parameters, wherein the target calculation parameters include the interface friction dissipation, the spring body energy dissipation, or the static friction coefficient; Performing trend analysis on the time series data of the target calculation parameter to obtain the corresponding change rate; Determine whether a change rate of a target dissipation is greater than a preset warning threshold, and obtain a warning judgment result, wherein the target dissipation includes the interface friction dissipation or the spring body energy dissipation; Determining whether the target dissipation is greater than a preset failure threshold, and obtaining a failure judgment result; If the interface friction dissipation is greater than a preset interface friction threshold and the rate of change of the static friction coefficient is greater than 0, the support fixture interface is considered abnormal as the fixture judgment result; otherwise, the support fixture interface is considered normal as the fixture judgment result; The performance test result is generated according to the early warning judgment result, the failure judgment result and the fixture judgment result.

2. The method according to claim 1, characterized in that The performing trend analysis on the time series data of the target calculation parameter to obtain the corresponding change rate includes: performing a noise level evaluation on the time series data of the target calculation parameter to obtain a noise level; Calculating a data smoothing parameter based on the noise level; Smoothing the time series data of the target calculation parameter according to the data smoothing parameter to obtain smoothed series data; Identifying target key parameters of trend changes based on the smoothed sequence data, wherein the target key parameters include key points or key intervals; Calculating a local change slope of the target calculation parameter according to the target key parameter; The consistency of the local change slope is verified to obtain the change rate.

3. The method according to claim 2, characterized in that The performing noise level evaluation on the time series data of the target calculation parameter to obtain the noise level includes: Performing statistical analysis on the time series data of the target calculation parameters to obtain statistical analysis results; Calculate the fluctuation range and dispersion degree according to the statistical analysis results; The noise level is determined according to the fluctuation range and the dispersion degree.

4. The method according to claim 2, characterized in that The step of identifying the target key parameters of trend changes based on the smoothed sequence data includes: Calculating a first difference between every two adjacent data points in the smoothed sequence data to obtain a first difference sequence; Calculating a second difference between every two adjacent first differences in the first difference sequence to obtain a second difference sequence; Fitting and deriving the second difference sequence to obtain a derivative curve; Taking a change point in the derivative curve whose value is greater than a preset change threshold as the key point; The change interval in the derivative curve where the value is greater than a preset change threshold is used as the key interval.

5. The method according to claim 2, characterized in that Calculating the local change slope of the target calculation parameter according to the target key parameter includes: Fitting the smoothed sequence data to obtain a fitting curve; Calculating the residual between the smoothed sequence data and the fitting curve according to the target key parameter; Determining the weight of each data point in the smoothed sequence data according to the residual; The local change slope is calculated according to each data point in the smoothed sequence data and its corresponding weight.

6. A super-precision spring performance monitoring system, characterized in that: include: a static friction state determination module, configured to determine the static friction state of the contact surface between the super-precision spring and the support fixture after applying a compressive force to the super-precision spring; an energy dissipation separation module, configured to calculate interface friction dissipation and spring body energy dissipation after performing axial cyclic loading on the super-precision spring if the static friction state is abnormal lubrication of the fixture contact surface; a static friction coefficient acquisition module, configured to acquire the static friction coefficient after pausing the axial cyclic loading and reapplying the compressive force; a performance detection module, configured to perform a performance detection based on the interface friction dissipation, the spring body energy dissipation, and the static friction coefficient to obtain a performance detection result; Wherein, determining the static friction state of the contact surface between the super-precision spring and the supporting fixture includes: collecting first force data and first displacement data; generating a force-time relationship curve based on the first force data and the first displacement data; If there is a sudden change point in the force-time relationship curve and the starting force value required to overcome the static friction is greater than the preset healthy lubrication state threshold, the static friction state is determined to be abnormal lubrication of the fixture contact surface; otherwise, the static friction state is determined to be normal lubrication of the fixture contact surface; The calculation of interface friction dissipation and spring body energy dissipation includes: Get the total energy dissipation; determining whether the support fixture has been subjected to transverse micro-vibration, wherein the direction of the transverse micro-vibration is perpendicular to the direction of the axial cyclic loading, and the transverse micro-vibration is used to suppress interface friction; When the supporting fixture has been subjected to lateral micro-vibration, collecting second force data and second displacement data; When the supporting fixture is not subjected to lateral micro-vibration, collecting third force data and third displacement data; performing a local morphological analysis on the second force data and the second displacement data to obtain a first mechanical feature; performing a local morphological analysis on the third force data and the third displacement data to obtain a second mechanical feature; Comparing the first mechanical characteristic with the second mechanical characteristic to obtain a comparison result; According to the comparison result, the energy dissipation corresponding to the reduced mechanical characteristics when the lateral micro-vibration is applied is used as the interface friction dissipation; Calculating the energy dissipation of the spring body according to the total energy dissipation and the interface friction dissipation; The performing local morphological analysis on the second force data and the second displacement data to obtain the first mechanical feature includes: calculating, according to the second force data and the second displacement data, a fluctuation amplitude within a preset time window as a current noise level; determining a mechanical characteristic threshold according to the current noise level; calculating a second-order rate of change curve with respect to time based on the second force data and the second displacement data; Identifying a transient peak in the second-order rate of change curve whose rate of change value is greater than the mechanical characteristic threshold; Taking the instantaneous peak as the first mechanical feature; The performing of a performance test based on the interface friction dissipation, the energy dissipation of the spring body, and the static friction coefficient to obtain a performance test result includes: generating time series data of the target calculation parameters according to target calculation parameters and historical data corresponding to the target calculation parameters, wherein the target calculation parameters include the interface friction dissipation, the spring body energy dissipation, or the static friction coefficient; Performing trend analysis on the time series data of the target calculation parameter to obtain the corresponding change rate; Determine whether a change rate of a target dissipation is greater than a preset warning threshold, and obtain a warning judgment result, wherein the target dissipation includes the interface friction dissipation or the spring body energy dissipation; Determining whether the target dissipation is greater than a preset failure threshold, and obtaining a failure judgment result; If the interface friction dissipation is greater than a preset interface friction threshold and the rate of change of the static friction coefficient is greater than 0, the support fixture interface is considered abnormal as the fixture judgment result; otherwise, the support fixture interface is considered normal as the fixture judgment result; The performance test result is generated according to the early warning judgment result, the failure judgment result and the fixture judgment result.

Images