Intermittent turning tool service life monitoring and evaluating system based on multi-sensor fusion

Abstract

The invention relates to the technical field of numerical control machine tool cutter state monitoring, and discloses an intermittent turning cutter service life monitoring and evaluating system based on multi-sensor fusion. The detection and evaluation system is composed of a data acquisition module, a noise data preprocessing module, a noise feature analysis and mode recognition module, a tool life prediction module and a multi-source data fusion module. According to the method, noise, vibration and temperature data are integrated, a weighted average algorithm is adopted, noise interference of a single sensor is reduced, the accuracy of cutter state evaluation is improved, filtering, smoothing and signal enhancement technologies are used, high-frequency noise is removed, data fluctuation is smoothed, useful signals are enhanced, data quality is improved, and data of all the sensors are standardized; the weight is calculated according to the reliability and the precision, weighted average fusion is carried out, fused high-quality data is generated, and the beneficial effect that errors caused by interference caused by noise in service life monitoring during cutting of the cutter are avoided is achieved.

Description

Technical Field

[0001] The invention relates to the technical field of tool state monitoring for numerically controlled machine tools, and in particular to a tool life monitoring and evaluation system for intermittent turning based on multi-sensor fusion. Background Art

[0002] With the continuous advancement of modern industrial informatization and intelligence, CNC machine tools are increasingly used in industrial production, and the accuracy of the tool plays a decisive role in the comprehensive performance of CNC machine tools and the quality of the workpiece. The wear of CNC machine tool tools will lead to a decrease in the dimensional accuracy of the machined surface, an increase in roughness, and a reduction in the quality of the workpiece. It may also cause the workpiece to be scrapped and increase production costs. If the performance of CNC machine tool tools can be monitored or predicted, the tools can be replaced or maintained in time to ensure the quality of the workpiece. Therefore, the research on the performance monitoring of CNC machine tool tools is very meaningful.

[0003] The prior art publication number is CN 116967844 A, which is a tool condition monitoring and life prediction system for CNC machine tools and its use method. Through temperature, noise, vibration frequency and actual use screen, this type of multi-source data is integrated to analyze its working status, and compared with the accumulated process knowledge base to judge abnormal conditions. According to the real-time results of monitoring and analysis, it can evaluate and predict the tool life, facilitate tool replacement, perform health online detection and maintenance, and improve the accuracy of CNC machine tool tool condition monitoring.

[0004] However, the effectiveness of the aforementioned system is highly dependent on the quality of data collected by the data acquisition module. In the actual data acquisition process, there is a huge amount of noise when the tool is cutting, and it is difficult for the analysis module to accurately evaluate the tool status. Even if the existing technology has noise cleaning processing, the cleaned data cannot be thoroughly cleaned and there are factual errors. This directly leads to the inability to detect when the tool is close to the critical damage value, and finally it breaks due to excessive cutting. Summary of the invention

[0005] Technical issues solved:

[0006] In view of the shortcomings of the existing technology, the present invention provides an intermittent turning tool life monitoring and evaluation system based on multi-sensor fusion, which has the advantages of not causing errors in tool life monitoring due to noise interference during tool cutting, and can accurately detect the critical life value of the tool during cutting to avoid economic losses, thereby solving the problems of the above-mentioned technology.

[0007] Technical solution:

[0008] To achieve the above object, the present invention provides the following technical solutions: a tool life monitoring and evaluation system for intermittent turning based on multi-sensor fusion, the detection and evaluation system consisting of a data acquisition module, a noise data preprocessing module, a noise feature analysis and pattern recognition module, a tool life prediction module and a multi-source data fusion module;

[0009] The data acquisition module is used to collect noise, vibration and temperature data of the tool during intermittent turning in real time;

[0010] The noise data preprocessing module is used to preprocess the collected noise data, extract effective information related to the tool state, use filtering and noise reduction technology to remove interference components, retain noise characteristics related to the tool state, support feature extraction algorithms, including spectrum analysis and wavelet transform, and extract key noise characteristics reflecting the tool state;

[0011] The noise feature analysis and pattern recognition module is used to conduct in-depth analysis of the pre-processed noise features and establish a pattern recognition model of the tool state based on historical data. By constructing a noise feature library containing different tool states: normal, hidden damage and damaged, as the basic data for pattern recognition, a convolutional neural network machine learning algorithm is used to train the tool state recognition model, match the real-time collected noise features with the patterns in the feature library, and identify the specific state of the current tool.

[0012] The tool life prediction module predicts the remaining life of the tool based on the noise data and the tool state identification results. According to the noise characteristics and historical data, a tool life curve model is established to describe the relationship between the tool state change and the use time. A time series analysis algorithm is used to predict the remaining life of the tool according to the current noise characteristics. The warning threshold and critical threshold of the tool life are set. When the predicted life is close to or below the threshold, an alarm prompt is issued;

[0013] The multi-source data fusion module fuses the noise data with other sensor data: vibration and temperature, adopts a weighted average data fusion algorithm, integrates multiple sensor data, and improves the accuracy of tool status evaluation.

[0014] Preferably, the noise in the data acquisition module is calculated as:

[0015]

[0016] Where: SPL is the noise sound pressure level, in decibels, P ref is the reference sound pressure, usually taken as 20 micropascals; log 10 represents a logarithmic function with base 10; P represents the measured sound pressure;

[0017] The vibration calculation in the data acquisition module is:

[0018]

[0019] Among them: a RMS is the effective value of vibration acceleration, in meters per square second; a(t) is the instantaneous acceleration, in meters per square second.

[0020] Preferably, the temperature data in the data acquisition module is calculated as:

[0021]

[0022] Where: T avg is the average temperature in degrees Celsius; T i is the temperature value at the i-th sampling moment, in degrees Celsius.

[0023] Preferably, the filtering calculation in the noise data preprocessing module is:

[0024]

[0025] Where: H(f) is the filter response at frequency f, f c is the cutoff frequency;

[0026] The noise reduction calculation in the noise data preprocessing module is:

[0027]

[0028] Where: y[n] is the denoised signal, x[n] is the original signal, and M is the window size of the moving average.

[0029] Preferably, the noise data and spectrum analysis in the processing module are calculated as follows:

[0030]

[0031] Where: X(f) is the spectrum at frequency f, x(t) is the time domain signal, and the fast Fourier transform algorithm is used to efficiently calculate the spectrum;

[0032] The wavelet transform calculation in the noise data preprocessing module is:

[0033]

[0034] Where CWT(a,b) is the wavelet coefficient at scale a and displacement b, x(t) is the original signal, ψ(t) is the wavelet function, and ψ * (t) is its complex conjugate. Wavelet transform provides the frequency information of the signal at different scales and positions.

[0035] Preferably, the input layer of the convolutional neural network in the noise feature analysis and identification module is the preprocessed noise feature data. Suppose the input data is X, and its dimension is (nH, nW, nC), where nH is the height: time axis, nW is the width: frequency axis, and nC is the number of channels: multiple channels of the spectrum. The convolution layer is calculated as:

[0036]

[0037] Where: Z i,j,k is the output of the kth convolution kernel at position (i, j); W m,n,c,k is the weight of the kth convolution kernel at position m,n,c,k; f H and f W is the height and width of the convolution kernel;

[0038] The activation function is calculated as:

[0039] A i,j,k =max(0,Z i,j,k )

[0040] Among them: A i,j,k is the output after activation.

[0041] Preferably, the convolutional neural network machine learning algorithm further includes a pooling layer, which is used to reduce the size of the feature map, reduce the amount of calculation and the number of parameters, and is calculated as:

[0042]

[0043] Where: P i,j,k is the output after pooling; s is the pooling stride;

[0044] The fully connected layer flattens the output of the pooling layer and connects it to the output layer. Let the input of the fully connected layer be X and the output be Y. The calculation formula is:

[0045]

[0046] Where: Y i is the output of the i-th output neuron; W j,i is the weight connecting the jth input neuron and the ith output neuron; n fc is the number of neurons in the fully connected layer;

[0047] The convolutional neural network machine learning algorithm also includes a loss function, which is used to measure the difference between the model's predicted value and the actual value. The cross entropy loss function is used for classification problems:

[0048]

[0049] Where: L is the loss value; C is the number of categories; p i,j is the probability that the model predicts that the i-th sample belongs to the j-th category.

[0050] Preferably, the time series module in the tool life prediction module is:

[0051]

[0052] Where: y t is the observed value at time t; φ i is the autoregressive coefficient, which represents the past observations y t-i For the current observation y t The influence of t is the error term at time t;

[0053] Using this model to predict future time points, the prediction formula is expressed as:

[0054]

[0055] in: is the predicted value at time t+h; y t+h-i is the observed value at the past time point t+hi.

[0056] Preferably, the tool life curve model of the tool life prediction module is calculated as:

[0057] S(t)=S0e -kt

[0058] Where: S(t) is the tool state at time t; k is a decay coefficient, which indicates the rate of change of the tool state over time;

[0059] Based on the life curve model, the remaining life of the tool remaining Calculated by the following formula:

[0060]

[0061] Where: S threshold is the set tool status threshold, indicating the critical point of tool failure;

[0062] The set warning threshold and critical threshold of tool life are expressed as:

[0063] Warning Threshold=α×Total Life;

[0064] Critical Threshold=β×Total Life;

[0065] Among them: α and β are constants less than 1, usually α>β.

[0066] Preferably, the weighted average data fusion algorithm in the multi-source data fusion module is:

[0067] Suppose there are three sensor data: noise data N, vibration data V and temperature data T, and the weight of each sensor data is w N 、w V and w T , and the weight sum is 1:

[0068] w N +w V +w T =1

[0069] The weighted average fused data F is expressed as:

[0070] F=w N N+w V ·V+w T ·T

[0071] Where: F is the fused data; w N 、w V and w T are the weights of noise data, vibration data, and temperature data respectively;

[0072] Weight w N 、w V and w T The reliability and accuracy of each sensor data are calculated. The reliability of each sensor is R N , R V and R T , the weight is calculated according to the following formula:

[0073]

[0074]

[0075] Where: R N , R V and R T are the reliability coefficients of the noise sensor, vibration sensor and temperature sensor respectively.

[0076] Compared with the prior art, the present invention provides a tool life monitoring and evaluation system for intermittent turning based on multi-sensor fusion, which has the following beneficial effects:

[0077] 1. The present invention adopts a weighted average algorithm to reduce the noise interference of a single sensor and improve the accuracy of tool state evaluation by integrating noise, vibration and temperature data. It uses filtering, smoothing and signal enhancement techniques to remove high-frequency noise, smooth data fluctuations, enhance useful signals, improve data quality, standardize the data of each sensor, calculate weights according to reliability and accuracy, perform weighted average fusion, generate fused high-quality data, use time series models such as ARIMA, predict the remaining life of the tool based on historical data, establish the relationship between state changes and usage time, set early warning and critical thresholds, monitor the predicted life, and issue alarms in time to avoid production interruptions and equipment damage caused by failure, thereby achieving the beneficial effect of not causing errors in tool life monitoring due to interference caused by noise during tool cutting.

[0078] 2. The present invention establishes a tool life curve model based on historical data to describe the relationship between tool state changes and usage time. According to the current fused data, a time series analysis algorithm is used to predict the remaining life of the tool, and a warning threshold and a critical threshold of the tool life are set to ensure timely alarm prompts. By real-time monitoring of the tool state and predicted life, an alarm prompt is issued in time to ensure that the operator can take timely measures to avoid production interruptions and economic losses caused by tool failure. When the predicted life is close to or lower than the set threshold, an alarm prompt is immediately issued, achieving the beneficial effect of being able to accurately discover the critical life value during tool cutting to avoid economic losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 It is a schematic diagram of the system of the present invention; DETAILED DESCRIPTION

[0080] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0081] See also Figure 1 ,Tool life monitoring and evaluation system for intermittent turning based on multi-sensor fusion,The detection and evaluation system consists of data acquisition module, noise data preprocessing module, noise feature analysis and pattern recognition module, tool life prediction module and multi-source data fusion module;

[0082] The data acquisition module is used to collect the noise, vibration and temperature data of the tool during intermittent turning in real time;

[0083] The noise data preprocessing module is used to preprocess the collected noise data, extract the effective information related to the tool status, use filtering and noise reduction technology to remove interference components, retain the noise characteristics related to the tool status, and support feature extraction algorithms, including spectrum analysis and wavelet transform, to extract key noise characteristics reflecting the tool status;

[0084] The noise feature analysis and pattern recognition module is used to conduct in-depth analysis of the pre-processed noise features and establish a pattern recognition model for the tool status based on historical data. By building a noise feature library containing different tool states: normal, hidden damage and damaged, as the basic data for pattern recognition, the convolutional neural network machine learning algorithm is used to train the tool status recognition model, match the real-time collected noise features with the patterns in the feature library, and identify the specific status of the current tool.

[0085] The tool life prediction module predicts the remaining life of the tool based on the noise data and tool state identification results. According to the noise characteristics and historical data, a tool life curve model is established to describe the relationship between tool state changes and usage time. The time series analysis algorithm is used to predict the remaining life of the tool based on the current noise characteristics. The warning threshold and critical threshold of the tool life are set. When the predicted life is close to or below the threshold, an alarm prompt is issued;

[0086] The multi-source data fusion module fuses the noise data with other sensor data: vibration and temperature, and adopts a weighted average data fusion algorithm to integrate multiple sensor data to improve the accuracy of tool status assessment.

[0087] Specifically, the noise in the data acquisition module is calculated as:

[0088]

[0089] Where: SPL is the noise sound pressure level, in decibels, P ref is the reference sound pressure, usually taken as 20 micropascals; log 10 represents a logarithmic function with base 10; P represents the measured sound pressure;

[0090] The vibration calculation in the data acquisition module is:

[0091]

[0092] Among them: a RMS is the effective value of vibration acceleration, in meters per square second; a(t) is the instantaneous acceleration, in meters per square second.

[0093] Specifically, the temperature data in the data acquisition module is calculated as:

[0094]

[0095] Where: T avg is the average temperature in degrees Celsius; T i is the temperature value at the i-th sampling moment, in degrees Celsius.

[0096] Specifically, the filtering calculation in the noise data preprocessing module is:

[0097]

[0098] Where: H(f) is the filter response at frequency f, f c is the cutoff frequency;

[0099] The noise reduction calculation in the noise data preprocessing module is:

[0100]

[0101] Where: y[n] is the denoised signal, x[n] is the original signal, and M is the window size of the moving average.

[0102] Specifically, the noise data and spectrum analysis calculation in the processing module are:

[0103]

[0104] Where: X(f) is the spectrum at frequency f, x(t) is the time domain signal, and the fast Fourier transform algorithm is used to efficiently calculate the spectrum;

[0105] The wavelet transform calculation in the noise data preprocessing module is:

[0106]

[0107] Where CWT(a,b) is the wavelet coefficient at scale a and displacement b, x(t) is the original signal, ψ(t) is the wavelet function, and ψ * (t) is its complex conjugate. Wavelet transform provides the frequency information of the signal at different scales and positions.

[0108] Specifically, the input layer of the convolutional neural network in the noise feature analysis and identification module is the preprocessed noise feature data. Let the input data be X, whose dimension is (nH, nW, nC), where nH is the height: time axis, nW is the width: frequency axis, and nC is the number of channels: multiple channels of the spectrum. The convolution layer is calculated as:

[0109]

[0110] Where: Z i,j,k is the output of the kth convolution kernel at position (i, j); W m,n,c,kis the weight of the kth convolution kernel at position m,n,c,k; f H and f W is the height and width of the convolution kernel;

[0111] The activation function is calculated as:

[0112] A i,j,k =max(0,Z i,j,k )

[0113] Among them: A i,j,k is the output after activation.

[0114] Specifically, the convolutional neural network machine learning algorithm also includes a pooling layer, which is used to reduce the size of the feature map, reduce the amount of calculation and the number of parameters, and is calculated as:

[0115]

[0116] Where: P i,j,k is the output after pooling; s is the pooling stride;

[0117] The fully connected layer flattens the output of the pooling layer and connects it to the output layer. Let the input of the fully connected layer be X and the output be Y. The calculation formula is:

[0118]

[0119] Where: Y i is the output of the i-th output neuron; W j,i is the weight connecting the jth input neuron and the ith output neuron; n fc is the number of neurons in the fully connected layer;

[0120] The convolutional neural network machine learning algorithm also includes a loss function, which is used to measure the difference between the model's predicted value and the actual value. The cross entropy loss function is used for classification problems:

[0121]

[0122] Where: L is the loss value; C is the number of categories; p i,j is the probability that the model predicts that the i-th sample belongs to the j-th category.

[0123] Specifically, the time series module in the tool life prediction module is:

[0124]

[0125] Where: y t is the observed value at time t; φ i is the autoregressive coefficient, which represents the past observations y t-i For the current observation yt The influence of t is the error term at time t;

[0126] Using this model to predict future time points, the prediction formula is expressed as:

[0127]

[0128] in: is the predicted value at time t+h; y t+h-i is the observed value at the past time point t+hi.

[0129] Specifically, the tool life curve model of the tool life prediction module is calculated as:

[0130] S(t)=S0e -kt

[0131] Where: S(t) is the tool state at time t; k is a decay coefficient, which indicates the rate of change of the tool state over time;

[0132] Based on the life curve model, the remaining life of the tool remaining Calculated by the following formula:

[0133]

[0134] Where: S threshold is the set tool status threshold, indicating the critical point of tool failure;

[0135] The set warning threshold and critical threshold of tool life are expressed as:

[0136] Warning Threshold=α×Total Life;

[0137] Critical Threshold=β×Total Life;

[0138] Among them: α and β are constants less than 1, usually α>β.

[0139] Specifically, the weighted average data fusion algorithm in the multi-source data fusion module is:

[0140] Suppose there are three sensor data: noise data N, vibration data V and temperature data T, and the weight of each sensor data is w N 、w V and w T , and the weight sum is 1:

[0141] w N +w V +w T=1

[0142] The weighted average fused data F is expressed as:

[0143] F=w N N+w V ·V+w T ·T

[0144] Where: F is the fused data; w N 、w V and w T are the weights of noise data, vibration data, and temperature data respectively;

[0145] Weight w N 、w V and w T The reliability and accuracy of each sensor data are calculated. The reliability of each sensor is R N , R V and R T , the weight is calculated according to the following formula:

[0146]

[0147] Where: R N , R V and R T are the reliability coefficients of the noise sensor, vibration sensor and temperature sensor respectively.

[0148] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and alterations may be made to the embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. The intermittent turning tool life monitoring and evaluation system based on multi-sensor fusion is characterized by: The detection and evaluation system is composed of a data acquisition module, a noise data preprocessing module, a noise feature analysis and pattern recognition module, a tool life prediction module and a multi-source data fusion module; The data acquisition module is used to collect noise, vibration and temperature data of the tool during intermittent turning in real time; The noise data preprocessing module is used to preprocess the collected noise data, extract effective information related to the tool state, use filtering and noise reduction technology to remove interference components, retain noise characteristics related to the tool state, support feature extraction algorithms, including spectrum analysis and wavelet transform, and extract key noise characteristics reflecting the tool state; The noise feature analysis and pattern recognition module is used to conduct in-depth analysis of the pre-processed noise features and establish a pattern recognition model of the tool state based on historical data. By constructing a noise feature library containing different tool states: normal, hidden damage and damaged, as the basic data for pattern recognition, a convolutional neural network machine learning algorithm is used to train the tool state recognition model, match the real-time collected noise features with the patterns in the feature library, and identify the specific state of the current tool. The tool life prediction module predicts the remaining life of the tool based on the noise data and the tool state identification results. According to the noise characteristics and historical data, a tool life curve model is established to describe the relationship between the tool state change and the use time. A time series analysis algorithm is used to predict the remaining life of the tool according to the current noise characteristics. The warning threshold and critical threshold of the tool life are set. When the predicted life is close to or below the threshold, an alarm prompt is issued; The multi-source data fusion module fuses the noise data with other sensor data: vibration and temperature, adopts a weighted average data fusion algorithm, integrates multiple sensor data, and improves the accuracy of tool status evaluation.

2. The intermittent turning tool life monitoring and evaluation system based on multi-sensor fusion according to claim 1 is characterized in that: The noise in the data acquisition module is calculated as: Where: SPL is the noise sound pressure level, in decibels, P ref is the reference sound pressure, usually taken as 20 micropascals; log 10 represents a logarithmic function with base 10; P represents the measured sound pressure; The vibration calculation in the data acquisition module is: Among them: a RMS is the effective value of vibration acceleration, in meters per square second; a(t) is the instantaneous acceleration, in meters per square second.

3. The intermittent turning tool life monitoring and evaluation system based on multi-sensor fusion according to claim 2 is characterized in that: The temperature data in the data acquisition module is calculated as: Where: T avg is the average temperature in degrees Celsius; T i is the temperature value at the i-th sampling moment, in degrees Celsius.

4. The intermittent turning tool life monitoring and evaluation system based on multi-sensor fusion according to claim 3 is characterized in that: The filtering calculation in the noise data preprocessing module is: Where: H(f) is the filter response at frequency f, f c is the cutoff frequency; The noise reduction calculation in the noise data preprocessing module is: Where: y[n] is the denoised signal, x[n] is the original signal, and M is the window size of the moving average.

5. The intermittent turning tool life monitoring and evaluation system based on multi-sensor fusion according to claim 4 is characterized in that: The noise data and spectrum analysis in the processing module are calculated as follows: Where: X(f) is the spectrum at frequency f, x(t) is the time domain signal, and the fast Fourier transform algorithm is used to efficiently calculate the spectrum; The wavelet transform calculation in the noise data preprocessing module is: Where CWT(a,b) is the wavelet coefficient at scale a and displacement b, x(t) is the original signal, ψ(t) is the wavelet function, and ψ * (t) is its complex conjugate. Wavelet transform provides the frequency information of the signal at different scales and positions.

6. The intermittent turning tool life monitoring and evaluation system based on multi-sensor fusion according to claim 5 is characterized in that: The input layer of the convolutional neural network in the noise feature analysis and identification module is the preprocessed noise feature data. Let the input data be X, whose dimension is (nH, nW, nC), where nH is the height: time axis, nW is the width: frequency axis, and nC is the number of channels: multiple channels of the spectrum. The convolutional layer is calculated as: Where: Z i,j,k is the output of the kth convolution kernel at position (i, j); W m,n,c,k is the weight of the kth convolution kernel at position m,n,c,k; f H and f W is the height and width of the convolution kernel; The activation function is calculated as: A j,j,k =max(0,Z i,j,k ) Among them: A i,j,k is the output after activation.

7. The intermittent turning tool life monitoring and evaluation system based on multi-sensor fusion according to claim 6 is characterized in that: The convolutional neural network machine learning algorithm also includes a pooling layer, which is used to reduce the size of the feature map, reduce the amount of calculation and the number of parameters, and is calculated as: Where: P i,j,k is the output after pooling; s is the pooling stride; The fully connected layer flattens the output of the pooling layer and connects it to the output layer. Let the input of the fully connected layer be X and the output be Y. The calculation formula is: Where: Y i is the output of the i-th output neuron; W j,i is the weight connecting the jth input neuron and the ith output neuron; n fc is the number of neurons in the fully connected layer; The convolutional neural network machine learning algorithm also includes a loss function, which is used to measure the difference between the model's predicted value and the actual value. The cross entropy loss function is used for classification problems: Where: L is the loss value; C is the number of categories; p i,j is the probability that the model predicts that the i-th sample belongs to the j-th category.

8. The intermittent turning tool life monitoring and evaluation system based on multi-sensor fusion according to claim 7 is characterized in that: The time series module in the tool life prediction module is: Where: y t is the observed value at time t; φ i is the autoregressive coefficient, which represents the past observations y t-i For the current observation y t The influence of t is the error term at time t; Using this model to predict future time points, the prediction formula is expressed as: in: is the predicted value at time t+h; y t+h-i is the observed value at the past time point t+hi.

9. The intermittent turning tool life monitoring and evaluation system based on multi-sensor fusion according to claim 8, characterized in that: The tool life prediction module tool life curve model is calculated as: S(t)=S0e -kt Where: S(t) is the tool state at time t; k is a decay coefficient, which indicates the rate of change of the tool state over time; Based on the life curve model, the remaining life of the tool remaining Calculated by the following formula: Where: S threshold is the set tool status threshold, indicating the critical point of tool failure; The set warning threshold and critical threshold of tool life are expressed as: Warning Threshold=α×Total Life; CriticalThreshold=β×Total Life; Among them: α and β are constants less than 1, usually α>β.

10. The intermittent turning tool life monitoring and evaluation system based on multi-sensor fusion according to claim 9, characterized in that: The weighted average data fusion algorithm in the multi-source data fusion module is: Suppose there are three sensor data: noise data N, vibration data V and temperature data T, and the weight of each sensor data is w N 、w V and w T , and the weight sum is 1: In N +in V +in T =1 The weighted average fused data F is expressed as: F=w N ·N+w V ·V+w T ·T Where: F is the fused data; w N 、w V and w T are the weights of noise data, vibration data, and temperature data respectively; Weight w N 、w V and w T The reliability and accuracy of each sensor data are calculated. The reliability of each sensor is R N , R V and R T , the weight is calculated according to the following formula: Where: R N , R V and R T are the reliability coefficients of the noise sensor, vibration sensor and temperature sensor respectively.

Images