Predictive maintenance system for spindle of numerical control machine tool

Abstract

The invention relates to a numerical control machine tool spindle predictive maintenance system, and belongs to the technical field of mechanical equipment health management and intelligent manufacturing. The main shaft driving motor real-time power processing module is used for processing the main shaft driving motor real-time power to estimate and obtain a real-time cutting torque sequence and a real-time main shaft total power sequence, and processing the multiple paths of temperature signals to obtain multiple paths of real-time temperature sequences, and the basic degradation state evaluation module is used for calculating mechanical accumulated damage quantitative indexes based on the real-time cutting torque sequence. And calculating a cumulative thermally induced damage index based on the multi-path real-time temperature sequence, and fusing the mechanical cumulative damage quantitative index and the cumulative thermally induced damage index to generate a basic degradation index. And the robustness and operability of a prediction result in a real industrial environment are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mechanical equipment health management and intelligent manufacturing, in particular to a numerical control machine tool spindle predictive maintenance system. BACKGROUND

[0002] The numerical control machine tool spindle is a core power component that determines machining precision and efficiency, and its unexpected failure can cause serious production interruption and economic loss; therefore, it is very important to ensure production continuity by performing predictive maintenance on the spindle and accurately predicting its remaining useful life. The existing technology generally has problems such as distorted degradation rate estimation and inability to effectively distinguish between normal working condition fluctuations and real performance degradation when dealing with complex and variable machining conditions, resulting in low reliability of the prediction results and difficulty in effective application in a flexible manufacturing environment.

[0003] The above information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0004] The purpose of the present application is to provide a numerical control machine tool spindle predictive maintenance system to solve the problems raised in the above background.

[0005] The technical solution of the present application is as follows: a multi-source data acquisition and preprocessing module is used to acquire real-time power of a spindle drive motor and multi-channel temperature signals, process the real-time power of the spindle drive motor to estimate a real-time cutting torque sequence and a real-time spindle total power sequence, and process the multi-channel temperature signals to obtain a multi-channel real-time temperature sequence;

[0006] A basic degradation state evaluation module is used to calculate a mechanical cumulative damage quantitative index based on the real-time cutting torque sequence, calculate a cumulative thermal damage index based on the multi-channel real-time temperature sequence, and fuse the mechanical cumulative damage quantitative index and the cumulative thermal damage index to generate a basic degradation index;

[0007] A working condition variation index calculation module is used to identify power mutation events based on the real-time spindle total power sequence and construct a probability distribution, calculate the information entropy of the probability distribution, and normalize the information entropy to generate a working condition variation index;

[0008] An adaptive life prediction compensation module is used to determine a basic degradation rate based on the basic degradation index to estimate a basic remaining useful life, generate a prediction confidence and a prediction effective time scale based on the working condition variation index, and combine the basic remaining useful life, the prediction confidence and the prediction effective time scale to generate a composite life prediction result.

[0009] Preferably, the generating process of the basic degradation indicator is:

[0010] normalizing the mechanical cumulative damage quantification indicator and the cumulative thermal-induced damage indicator;

[0011] weighting and fusing the normalized mechanical cumulative damage quantification indicator and the cumulative thermal-induced damage indicator to obtain the basic degradation indicator.

[0012] Preferably, the calculating process of the mechanical cumulative damage quantification indicator is:

[0013] performing Fourier transform on the real-time cutting torque sequence in the time window to obtain a power spectral density, calculating an integral result of the power spectral density in a preset core frequency band, and accumulating the integral result in time to obtain the mechanical cumulative damage quantification indicator.

[0014] Preferably, the calculating process of the cumulative thermal-induced damage indicator is:

[0015] calculating an instantaneous thermal stress representation quantity based on the multi-channel real-time temperature sequence and a preset temperature measurement point spatial distance, and accumulating the instantaneous thermal stress representation quantity in time to obtain the cumulative thermal-induced damage indicator.

[0016] Preferably, the generating process of the working condition variation index is:

[0017] calculating a difference sequence of the real-time spindle total power sequence, identifying the amplitude of a power mutation event based on an adaptive threshold, counting the amplitude of the power mutation event in an analysis time window and constructing an amplitude probability distribution, calculating a power mutation information entropy based on the amplitude probability distribution, and normalizing the power mutation information entropy to obtain the working condition variation index.

[0018] Preferably, the generating process of the composite life prediction result is:

[0019] fitting the historical basic degradation indicator sequence, estimating a current degradation rate, calculating a basic remaining useful life based on the current degradation rate and a preset failure threshold, and combining the basic remaining useful life, the prediction confidence and the prediction effective time scale to generate the composite life prediction result.

[0020] Preferably, the generating process of the prediction confidence is:

[0021] calculating the prediction confidence based on the working condition variation index and a preset confidence adjustment factor.

[0022] Preferably, the generating process of the prediction effective time scale is:

[0023] The exponential decay model is adopted, and the predicted effective time scale is obtained by calculation based on the working condition variation index and a preset time scale decay coefficient.

[0024] The present application improves a numerical control machine tool spindle predictive maintenance system, which has the following improvements and advantages compared with the prior art:

[0025] 1. The present application quantifies the abstract machining condition complexity accurately by setting a working condition variation index calculation module, solves the core pain point that the prior art is difficult to distinguish between normal condition fluctuation and real performance degradation under varying conditions, enables the system to have condition perception ability, can clearly distinguish whether the sharp change of the signal is caused by normal production scheduling or abnormal performance degradation, thereby greatly improving the accuracy of prediction and effectively suppressing false positives;

[0026] 2. The basic degradation index constructed by the present application realizes more comprehensive and robust tracking of the health state of the spindle through quantification and fusion of mechanical and thermal damage sources, and the fusion of multi-physical field information ensures the comprehensiveness of the degradation evaluation, avoiding the risk of missed judgment caused by single signal source failure or insensitivity;

[0027] 3. The self-adaptive life prediction compensation module of the present application makes the final output of the composite life prediction result have dynamic and practicality, changes the situation that the traditional predictive maintenance information is single and not strong in guidance, when the working condition is stable, the system can provide a longer-term maintenance planning reference; when the working condition changes sharply, the system automatically narrows the prediction range to the short term, prompting the user to observe carefully; this design makes the output of the present application not a isolated number, but an intelligent decision support information containing prediction value, reliability, effective time, and context-aware, which significantly improves the robustness and operability of the prediction result in real industrial environment. BRIEF DESCRIPTION OF DRAWINGS

[0028] The present application will be further explained in conjunction with the accompanying drawings and examples:

[0029] Figure 1 is a flow chart of a numerical control machine tool spindle predictive maintenance system of the present application. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail in conjunction with specific examples.

[0031] Example 1:

[0032] Please refer to Figure 1The application provides a numerical control machine tool spindle predictive maintenance system, comprising: a multi-source data acquisition and preprocessing module, which is used for acquiring a spindle driving motor real-time power and a plurality of temperature signals, processing the spindle driving motor real-time power to estimate a real-time cutting torque sequence and a real-time spindle total power sequence, and processing the plurality of temperature signals to obtain a plurality of real-time temperature sequences;

[0033] a basic degradation state evaluation module, which is used for calculating a mechanical cumulative damage quantitative index based on the real-time cutting torque sequence, calculating a cumulative thermal damage index based on the plurality of real-time temperature sequences, and fusing the mechanical cumulative damage quantitative index and the cumulative thermal damage index to generate a basic degradation index;

[0034] a working condition variation index calculation module, which is used for identifying a power mutation event based on the real-time spindle total power sequence and constructing a probability distribution, calculating an information entropy of the probability distribution, and normalizing the information entropy to generate a working condition variation index;

[0035] an adaptive life prediction compensation module, which is used for determining a basic degradation rate based on the basic degradation index to estimate a basic remaining service life, generating a prediction credibility and a prediction effective time scale based on the working condition variation index, and combining the basic remaining service life, the prediction credibility and the prediction effective time scale to generate a composite life prediction result.

[0036] In the embodiment, through the precise cooperation of the four internal modules, deep insight and accurate prediction of the spindle health state are realized; the input and preprocessing unit of the system, i.e., the multi-source data acquisition and preprocessing module, continuously captures running data from key parts of the numerical control machine tool spindle, including driving motor real-time power and a plurality of temperature signals distributed at heat-sensitive positions such as front and rear bearings of the spindle ; the core task of the module is not only data acquisition, but also analysis of the original signal through a physical model, according to the principle of motor engineering, the cutting net power is separated from the total power by excluding the no-load power , for example, by collecting the power of the spindle when it is not cutting, i.e., the no-load running power, taking it as a no-load power baseline, subtracting the baseline value or the no-load power model value based on the speed from the total power, the real-time cutting torque sequence which is important for mechanical damage evaluation is accurately estimated, and the real-time spindle total power sequence and the plurality of real-time temperature sequences are provided to the upper module; these preprocessed and synchronized time series data constitute the data cornerstone of the entire prediction system;

[0037] Based on the preprocessed data, the basic degradation state evaluation module receives the cutting torque and temperature data to construct a basic degradation index that can reflect the irreversible damage accumulation of the spindle At the same time, the working condition variation index calculation module processes the total power data in parallel, with the aim of quantifying the complexity and variability of the current processing conditions and generating a dynamic working condition variation index. ; The adaptive life prediction compensation module is the decision-making unit of the system, integrating the outputs of the first two modules: not only based on The historical trend of the remaining life of the foundation is used to predict the Dynamic compensation and adjustment are performed on the prediction's credibility and effective time range. This architecture enables the system to output a composite life prediction result that includes a predicted value, a confidence interval, and an effective time scale, providing production managers with a decision-making basis that is far richer, more reliable, and more practical than traditional methods, thereby improving the effectiveness of predictive maintenance in a flexible manufacturing environment.

[0038] Example 2

[0039] The generation process of basic degradation index is:

[0040] Normalize the quantitative index of mechanical cumulative damage and the cumulative thermal damage index;

[0041] The normalized mechanical cumulative damage quantitative index and the cumulative thermal damage index are weighted and fused to obtain the basic degradation index.

[0042] The calculation process of the quantitative index of mechanical cumulative damage is as follows:

[0043] Perform Fourier transform on the real-time cutting torque sequence within the time window to obtain the power spectrum density. Calculate the integral of the power spectrum density within the preset core frequency band to obtain an integral result. The integral result is accumulated over time to obtain a quantitative index of mechanical cumulative damage.

[0044] The calculation process of the cumulative heat-induced damage index is:

[0045] Based on multiple real-time temperature sequences and the preset spatial distances between temperature measurement points, the instantaneous thermal stress characterization quantity is calculated, and the instantaneous thermal stress characterization quantity is accumulated over time to obtain the cumulative thermal damage index.

[0046] In this embodiment, the basic degradation index The internal logic of the generation of is to quantify and mathematically integrate two key physical damage mechanisms. To achieve this goal, the basic degradation state assessment module first quantifies mechanical fatigue and thermal stress accumulation in parallel.

[0047] For mechanical damage, quantitative indicators of mechanical cumulative damage The calculation of dfs abandons the simple moment summation because it cannot reflect the key effect of cyclic stress frequency on fatigue life;

[0048] The technical motivation is that the fatigue damage of the material is closely related to the frequency and energy distribution of the stress cycle; the power spectral density analysis based on Fourier transform is introduced, aiming to more scientifically measure the accumulated energy of the stress component that contributes most to the damage from the frequency domain perspective; the calculation formula of the index at discrete time points is defined as: ; wherein, represents the mechanical cumulative damage index at the cutoff time point , with the physical dimension of the square of the moment of force multiplied by time, for example ; refers to the index of the cutoff discrete time point; is the power spectral density of the cutting moment sequence in the jth time window , which describes the distribution of moment energy with frequency f; is the preset core frequency band, which is determined by modal analysis or historical data spectrum analysis on the specific spindle, accurately locking the frequency range most related to key damage modes such as bearing wear and tool chatter, ensuring the pertinence and accuracy of the calculation;

[0049] A feasible historical data spectrum analysis method is as follows: collect the cutting moment or vibration signal time series of the spindle under the condition of performing typical heavy load machining or appearing known abnormalities such as vibration marks and abnormal noise; apply fast Fourier transform to the sequence data to obtain its power spectral density graph; identify the characteristic frequency band whose power spectral density value is more than one order of magnitude higher than the background noise under the above abnormal working conditions; for example, the upper and lower limits of the frequency range where 90% of the power spectrum peak energy is concentrated can be used as the boundaries of the core frequency band and ; by repeating this process on multiple historical failure samples and comprehensive analysis, a universal and robust core frequency band can be determined for a specific type of spindle;

[0050] Through this formula, the time-varying cutting moment signal is converted into a monotonically increasing quantitative index , whose growth rate directly reflects the severity of the mechanical load recently suffered by the spindle, and its cumulative value records the total mechanical fatigue damage of the spindle since it was put into operation;

[0051] For thermal-induced damage, the calculation of the cumulative thermal-induced damage index aims to quantitatively accumulate the continuous thermal stress effect caused by the spatial temperature inhomogeneity in time; the motivation is that the thermal damage of the spindle is accumulated by the continuous action of the thermal stress field generated by the temperature difference between different positions; therefore, the instantaneous stress level must be characterized first, and then integrated in the time dimension; the instantaneous thermal stress characterization quantity , as a discrete approximation of the spatial temperature gradient, is given by: ; where, is the instantaneous thermal stress indicator; and are the real-time temperatures of the first and second temperature measurement points, respectively, transmitted from module one; and are the real-time temperatures of the first and second temperature measurement points, respectively, transmitted from module one; is the known spatial linear distance between the two temperature measurement points;

[0052] To effectively capture the thermal gradient generated by the spindle operation, a minimum sensor layout scheme is recommended: at least two temperature sensors are installed on the front and rear bearing outer rings of the spindle, respectively; on each bearing, the two sensors should be oppositely arranged along the radial direction, for example, one at the 12 o'clock position and the other at the 6 o'clock position; such a layout not only can monitor the overall temperature rise of the bearing, but also can maximize the detection of local temperature differences caused by uneven force or poor lubrication, thereby providing high-quality input data for the calculation of the instantaneous thermal stress indicator ;

[0053] Then, the instantaneous quantity is time-accumulated to obtain the cumulative thermal-induced damage index: ; where, is the cumulative thermal-induced damage index at the cutoff time point ; through these two serial calculation steps, the basic degradation state evaluation module converts multiple discrete temperature readings into a cumulative thermal damage index that is consistent in mathematical form with the mechanical damage index ; is the jth discrete time point;

[0054] To generate a unified basic degradation index , the module uses a linear weighting method for fusion; this process begins with maximum and minimum value normalization processing of and to eliminate dimensional differences and map them to a unified interval; perform weighted fusion: ; where, and are the normalized indices; the weight coefficients and satisfy , and their values can be set based on expert experience for a specific machine tool or determined through data-driven methods such as principal component analysis;

[0055] When principal component analysis is used to determine the weights, the steps are as follows: obtain at least a number of complete maintenance cycles of normalized mechanical cumulative damage indices and cumulative thermal-induced damage indices Time series data; take these two series as variables and construct a The input matrix is ​​n, where n is the number of time points; principal component analysis is performed on the matrix to calculate the load vector of the first principal component, which is recorded as The two load values ​​reflect the contribution of the original variables to the first principal component, that is, the overall degradation trend of the system; the weight coefficient is normalized according to the absolute size of the load value, and the formula is: , This method bases the determination of weights on an objective analysis of the inherent relevance of the data itself.

[0056] This fusion step generates a unified, monotonically increasing health metric This indicator comprehensively reflects the cumulative effects of mechanical and thermal damage, providing a reliable degradation trajectory basis for subsequent life prediction.

[0057] Example 3

[0058] The generation process of the operating condition variation index is:

[0059] The differential sequence of the real-time spindle total power sequence is calculated, and the amplitude of the power mutation event is identified based on the adaptive threshold. Within the analysis time window, the amplitude of the power mutation event is counted and the amplitude probability distribution is constructed. Based on the amplitude probability distribution, the power mutation information entropy is calculated and normalized to obtain the operating condition variation index.

[0060] In this embodiment, the operating condition variation index calculation module aims to solve the technical problem that existing technologies cannot distinguish between normal operating condition fluctuations and actual performance degradation. By introducing Shannon entropy, a core tool of information theory, the complexity of the operating condition is quantitatively characterized.

[0061] The generation of the working condition variation index follows a strict derivation chain; by calculating the real-time total spindle power sequence Difference To capture instantaneous changes in power and use adaptive threshold Identify meaningful operating condition change events from random noise; this threshold is calculated dynamically, for example, within an analysis window Built-in setting ,in is the standard deviation of the power sequence within the window, Usually 3 is used, based on the 3-sigma rule in statistics, which ensures the statistical significance of event identification; in a sliding analysis window The amplitudes of all identified power mutation events are collected and divided into N discrete intervals. The event frequency of each interval is counted to construct the probability distribution of power mutation amplitude. ; wherein, the interval number N can be based on the data-driven Sturges rule, M is the total number of mutation events within the window, and is adaptively determined; by substituting this probability distribution into the definition of Shannon entropy, the power mutation information entropy : ; wherein, The unit of is bit, and its value directly reflects the complexity of the working condition; in order to obtain a standardized dimensionless index, the module performs maximum and minimum value normalization processing on the information entropy , and derives the final working condition variation index : ; wherein, and are the historical entropy value extremes learned and updated by the system during long-term operation, which ensures the dynamic adaptability of the index;

[0062] Through the above steps, the working condition variation index calculation module realizes the quantitative representation of the working condition complexity, and generates the working condition variation index ; The index provides a key input for the adaptive prediction ability of the system, so that the system can clearly judge whether it is in a stable processing stage or a severe fluctuation stage.

[0063] Embodiment 4

[0064] The generation process of the composite life prediction result is:

[0065] fitting the historical basic degradation index sequence, estimating the current degradation rate, calculating the basic remaining useful life based on the current degradation rate and the preset failure threshold, combining the basic remaining useful life, the prediction confidence and the prediction effective time scale to generate the composite life prediction result;

[0066] The generation process of the prediction confidence is:

[0067] The prediction confidence is calculated based on the working condition variation index and the preset confidence adjustment factor;

[0068] The generation process of the prediction effective time scale is:

[0069] The prediction effective time scale is calculated based on the working condition variation index and the preset time scale decay coefficient by using an exponential decay model.

[0070] In this embodiment, the function of the adaptive life prediction compensation module is to generate a composite and intelligent life prediction result; the process starts with an estimation of a basic remaining useful life; the module fits the historical basic degradation index sequence by using a basic degradation model, estimates the current degradation rate ;

[0071] One widely used and effective baseline degradation model is the exponential degradation model, mathematically expressed as ; where parameter can be fitted by applying nonlinear least squares method to historical data series. Once the model parameters are determined, the current degradation rate can be analytically calculated by taking the first derivative of the model, i.e. . This model can well describe the physical process of many devices accelerating degradation as they approach their end-of-life;

[0072] Based on this rate, the baseline RUL is calculated as ; where is a pre-defined failure threshold, which is derived from statistical analysis of historical data to determine the critical value of when the main shaft fails to function or its machining accuracy is out of specification;

[0073] The functional failure can be quantitatively defined, for example, the total vibration value of the main shaft bearing exceeds the alarm limit value specified in the ISO10816 standard for this type of machine, or the surface roughness Ra value of the machined standard test piece exceeds the upper limit specified in the drawing for three consecutive times; in the historical database, all time points that meet the above failure definition are screened out, and the corresponding baseline degradation index value at that time is recorded; the values of of these failure points form a statistical sample set; to establish a sensitive and reliable threshold, the 90th percentile of the sample set is taken as the failure threshold ; this method ensures that the prediction model will not be too conservative due to individual extreme cases of early failure;

[0074] This is not directly output, but is used as a baseline prediction value, which is adjusted by the module using the working condition variation index , which is achieved by generating two key dynamic compensation factors: prediction reliability and prediction effective time scale ;

[0075] The generation of prediction reliability , the technical motivation is that when the working condition of the machine tool changes dramatically, the uncertainty of degradation assessment increases, at this time the prediction reliability is low, so the mechanism must be used to convey this uncertainty to the user; the generation formula is ; where is the final output of the prediction reliability between [0, 1]; It is a credibility adjustment factor greater than 0, which is used to adjust the sensitivity of credibility. Its value can be determined by backtesting and optimizing on historical data containing false alarms caused by sudden changes in working conditions, and selecting a value that can effectively suppress false alarms. value;

[0076] The optimization process is to set an objective function, such as the F1 score, which takes into account both the precision and recall of predictions, on a historical validation dataset containing normal working conditions and sudden changes in working conditions; and iteratively adjust the F1 score within a preset range, such as 0.1 to 2.0, with a step size of 0.1. , calculate the F1 score of the system prediction result under each value; select the value that maximizes the F1 score The value is used as the best credibility adjustment factor;

[0077] The credibility of the prediction This enables the system to dynamically adjust the confidence interval of the prediction results, thereby providing effective risk warnings;

[0078] Prediction effective time scale The technical motivation for the generation of RUL is to overcome the defect of traditional prediction systems that output RUL values ​​without time limits, so that the system can intelligently adjust its prediction horizon according to the clarity of the current situation. This implementation adopts an exponential decay model to construct this relationship: ;in, is the effective time scale of the forecast after dynamic adjustment; It is the longest prediction time allowed by the system and is a priori value set in combination with the factory maintenance cycle and equipment characteristics; is the time scale attenuation coefficient. This dimensionless constant can be preset according to the severity of the equipment or maintenance strategy. Values ​​correspond to more conservative forecasting strategies;

[0079] In order to make The value setting is more operational and can introduce a factory equipment risk assessment score. This score is usually assessed by the maintenance department based on factors such as the importance of the equipment on the production line, maintenance cost and difficulty, and ranges from 1 to 10. The value can be associated with the RAS, for example, by setting it via an empirical formula: ,in is a baseline coefficient (e.g., 2.0), and RAS is the risk score for the device; thus, for devices with high severity (RAS>5), The value will increase accordingly, causing the prediction effective time scale to decay faster when the operating conditions change, thereby triggering more timely short-term warnings;

[0080] The prediction effective time scale enables the system to provide long-term prediction in stable working conditions and automatically switch to short-term warning mode in volatile working conditions;

[0081] The module outputs a composite prediction result composed of the basic residual life , prediction reliability , and prediction effective time scale . The multi-dimensional and adaptive prediction result provides more operationally meaningful information containing risk assessment for maintenance decision, thereby significantly improving the accuracy and reliability of predictive maintenance.

[0082] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, which should be covered by the scope of the claims of the present application.

Claims

1. A predictive maintenance system for a CNC machine tool spindle, characterized in that: include: a multi-source data acquisition and preprocessing module, configured to obtain the real-time power of the spindle drive motor and multiple temperature signals, process the real-time power of the spindle drive motor to estimate a real-time cutting torque sequence and a real-time total spindle power sequence, and process the multiple temperature signals to obtain a multiple real-time temperature sequence; a basic degradation state assessment module, configured to calculate a mechanical cumulative damage quantitative index based on the real-time cutting torque sequence, calculate a cumulative thermal damage index based on the multiple real-time temperature sequences, and fuse the mechanical cumulative damage quantitative index and the cumulative thermal damage index to generate a basic degradation index; a working condition variation index calculation module, configured to identify power mutation events based on the real-time spindle total power sequence and construct a probability distribution, calculate the information entropy of the probability distribution, and normalize the information entropy to generate a working condition variation index; An adaptive life prediction compensation module is used to determine a basic degradation rate based on the basic degradation index to estimate a basic remaining service life, generate a prediction credibility and a prediction effective time scale based on the operating condition variation index, and generate a composite life prediction result by combining the basic remaining service life, the prediction credibility and the prediction effective time scale.

2. A CNC machine tool spindle predictive maintenance system according to claim 1, characterized in that: The generation process of the basic degradation index is as follows: Normalizing the mechanical cumulative damage quantification index and the cumulative thermal damage index; The normalized mechanical cumulative damage quantification index and the cumulative thermal damage index are weightedly fused to obtain the basic degradation index.

3. A predictive maintenance system for a CNC machine tool spindle according to claim 2, characterized in that: The calculation process of the mechanical cumulative damage quantitative index is as follows: The real-time cutting torque sequence within the time window is subjected to Fourier transform to obtain a power spectral density, the integral of the power spectral density within a preset core frequency band is calculated to obtain an integral result, and the integral result is accumulated over time to obtain the mechanical cumulative damage quantitative index.

4. A predictive maintenance system for a CNC machine tool spindle according to claim 2, characterized in that: The calculation process of the cumulative heat-induced damage index is: Based on the multi-channel real-time temperature sequence and the preset spatial distance between the temperature measurement points, the instantaneous thermal stress characterization quantity is calculated, and the instantaneous thermal stress characterization quantity is accumulated over time to obtain the cumulative heat-induced damage index.

5. The predictive maintenance system for a CNC machine tool spindle according to claim 1, characterized in that: The generation process of the operating condition variation index is as follows: The differential sequence of the real-time spindle total power sequence is calculated, and the amplitude of the power mutation event is identified based on an adaptive threshold. Within the analysis time window, the amplitude of the power mutation event is counted and an amplitude probability distribution is constructed. Based on the amplitude probability distribution, the power mutation information entropy is calculated, and the power mutation information entropy is normalized to obtain the operating condition variation index.

6. The predictive maintenance system for a CNC machine tool spindle according to claim 1, characterized in that: The generation process of the composite life prediction result is as follows: The basic degradation indicator sequence of history is fitted to estimate the current degradation rate, and based on the current degradation rate and a preset failure threshold, the basic remaining useful life is calculated. The basic remaining useful life, the prediction credibility, and the prediction effective time scale are combined to generate the composite life prediction result.

7. The predictive maintenance system for a CNC machine tool spindle according to claim 1, characterized in that: The generation process of the prediction credibility is: The prediction credibility is obtained by performing calculation based on the operating condition variation index and a preset credibility adjustment factor.

8. The predictive maintenance system for a CNC machine tool spindle according to claim 1, characterized in that: The generation process of the prediction effective time scale is: An exponential decay model is used to calculate based on the operating condition variation index and a preset time scale decay coefficient to obtain the predicted effective time scale.

Images