Tool state indication method and indication device for multi-tool machining of variable pitch screws

Abstract

The present application relates to the technical field of tool state detection, and particularly to a tool state indication method and an indication device for multi-tool machining of variable pitch screws. The method includes: determining a vibration fluctuation coefficient according to the dispersion degree of all vibration data of each tool in any direction in each time period, and the difference of all vibration data of each tool in any direction between each time period and the immediately preceding time period; for the vertical direction among all directions of each tool, determining a frequency domain fluctuation coefficient according to the difference and distribution of the fluctuation conditions of all vibration data in the frequency domain between each time period and the first time period; determining a tool weight based on the correlation between the vibration fluctuation coefficient and the frequency domain fluctuation coefficient of each tool in all time periods, and monitoring the tool state in combination with the vibration data. The present application comprehensively analyzes the fluctuation conditions of vibration data in the time domain and the frequency domain to improve the accuracy of tool state detection.

Description

Technical Field

[0001] This application relates to the technical field of tool condition detection, and specifically relates to a tool condition indication method and an indication device for multi-tool machining of variable pitch screws. Background Art

[0002] With the continuous progress of manufacturing technology, the processing requirements for precision mechanical parts are getting higher and higher. As a special mechanical component, the variable pitch screw plays an important role in many application fields due to its non-uniform pitch design, such as in equipment like vacuum pumps. These devices have extremely high requirements for the machining accuracy of the screw. Therefore, during the machining process of the variable pitch screw, ensuring the good condition of the tool is crucial for guaranteeing the quality of the final product.

[0003] The existing technology establishes a model between the tool condition and the monitoring data through vibration data, so as to realize the monitoring of the tool condition. However, the model between the tool condition and the monitoring data in the existing technology is established under the conditions of fixed tools and fixed pitches. When the tool is replaced or the parameters are modified, tool compensation is required, thus introducing new errors. Therefore, this method is relatively complex and has poor flexibility. When the machining conditions or tools need to be frequently replaced, the accuracy is also poor. Moreover, as the machining progresses, the tool condition deteriorates. In addition to the mutability of the vibration data of the tool, there is also a certain gradual change characteristic as a whole. In traditional anomaly detection algorithms, this gradual change characteristic of the data will reduce the accuracy of tool condition detection. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide a tool condition indication method and an indication device for multi-tool machining of variable pitch screws. The specific technical solutions adopted are as follows:

[0005] In the first aspect, an embodiment of this application provides a tool condition indication method for multi-tool machining of variable pitch screws. The method includes the following steps:

[0006] During the multi-tool machining process of the variable pitch screw, obtain the vibration data in all directions of each tool at each acquisition moment within the preset duration before and including the current moment;

[0007] Divide the preset duration into multiple time periods. According to the dispersion degree of all vibration data in any direction of each tool in each time period, determine the vibration characteristic value of each tool in each time period, and combine the difference of all vibration data in any direction of each tool between each time period and its adjacent previous time period to determine the vibration fluctuation coefficient of each tool in each time period;

[0008] For the vertical direction in all directions of each tool, according to the differences in the frequency-domain fluctuation conditions of all vibration data between each time period and the first time period, determine the frequency-domain vibration values of each tool at each time period, and combine the distribution of all vibration data in the frequency domain in any direction of each tool at each time period to determine the frequency-domain fluctuation coefficient of each tool at each time period;

[0009] Based on the correlation between the vibration fluctuation coefficient and the frequency-domain fluctuation coefficient of each tool in all time periods, and combining the vibration fluctuation coefficient and the frequency-domain fluctuation coefficient, determine the tool weight of each tool at each time period;

[0010] Based on the vibration data of each tool in all directions at each acquisition moment and the tool weight, determine the feature vector of each tool at each acquisition moment. Based on the distance between the feature vectors of any two acquisition moments of each tool, determine the distance formula between the feature vectors of any two acquisition moments of each tool, and combine the feature vectors of each tool at all acquisition moments and the anomaly detection algorithm to monitor the tool state of the multi-tool machining of the variable pitch screw.

[0011] Preferably, the vibration characteristic value of each tool at each time period is the sum of the degrees of dispersion of the vibration data in all directions of each tool at each time period.

[0012] Preferably, the expression of the vibration fluctuation coefficient of each tool at each time period is: ; in the formula, represents the vibration fluctuation coefficient of tool j at time period i; represents the vibration characteristic value of tool j at time period i; , , respectively represent the differences in all vibration data of tool j in the lateral, vertical, and longitudinal directions in all directions between time period i and its adjacent previous time period; represents a preset constant greater than 0.

[0013] Preferably, the method for determining the frequency-domain vibration value of each tool at each time period is:

[0014] Take all the vibration data in the vertical direction of each tool at each time period as the input of the time-frequency conversion algorithm, output the spectrogram in the vertical direction of each tool at each time period, and use it as the spectrogram of each tool at each time period;

[0015] Extract all the peaks in the spectrogram of each tool at each time period, fit all the peaks to obtain a peak fitting curve, and take the difference between the integral area of the peak fitting curve of each tool at each time period and the integral area of the peak fitting curve of the first time period as the frequency-domain vibration value of each tool at each time period.

[0016] Preferably, the expression of the frequency-domain fluctuation coefficient of each tool at each time period is: ; In the formula, represents the frequency-domain fluctuation coefficient of tool j at time period i; , respectively represent the main frequencies in the spectrograms of tool j at time period i and time period i-k; represents the frequency-domain vibration value of tool j at time period i.

[0017] Preferably, the method for determining the tool weight of each tool at each time period is:

[0018] Calculate the normalized value of the product of the vibration fluctuation coefficient and the frequency-domain fluctuation coefficient of each tool at each time period;

[0019] Analyze the correlation between the vibration fluctuation coefficient and the frequency-domain fluctuation coefficient of each tool at all time periods;

[0020] Take the product of the normalized value and the correlation of each tool at each time period as the tool weight of each tool at each time period.

[0021] Preferably, the method for determining the feature vector of each tool at each acquisition moment is:

[0022] Take the result of multiplying the vibration data of each tool in all directions at each acquisition moment by the tool weight corresponding to the time period where the acquisition moment is located to form the feature vector of each tool at each acquisition moment.

[0023] Preferably, the distance formula between the feature vectors of any two acquisition moments of each tool is: ; In the formula, is the distance between the feature vectors of tool j at acquisition moment and acquisition moment ; is the Euclidean distance between the feature vectors of tool j at acquisition moment and acquisition moment , , are respectively the tool weights of tool j at acquisition moment and acquisition moment .

[0024] Preferably, the monitoring of the tool state in multi-tool machining of a variable pitch screw includes:

[0025] Take the feature vectors of each tool at the current moment and all the acquisition moments before it as the input of the LOF anomaly detection algorithm, where the distance formula between the feature vectors of any two acquisition moments of each tool is used as the distance formula in the LOF anomaly detection algorithm, and output the anomaly score values of each tool at the current moment and all the acquisition moments before it;

[0026] Take the anomaly score values of each tool at the current moment and all the acquisition moments before it as the input of the threshold segmentation algorithm, and output the segmentation threshold;

[0027] If the anomaly score value at the current moment is greater than the segmentation threshold, the tool state is abnormal; otherwise, the tool state is normal.

[0028] In a second aspect, an embodiment of the present application further provides a tool state indication device for multi-tool machining of a variable pitch screw, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the tool state indication method for multi-tool machining of a variable pitch screw described in any one of the above are implemented.

[0029] The present application has at least the following beneficial effects:

[0030] Based on the time-series change of the vibration data of the tool in different directions, the present application constructs a vibration fluctuation coefficient. The beneficial effect is that the vibration fluctuation coefficient reflects the vibration change of the tool in the time domain, enhances the sensitivity of capturing the change trend of the vibration data, and thus enhances the sensitivity of identifying the abnormal tool state; then, according to the fluctuation of the vibration data of the tool in the vertical direction in the frequency domain, the present application constructs a frequency domain fluctuation coefficient. The beneficial effect is that the frequency domain fluctuation coefficient reflects the vibration change of the tool in the frequency domain, further enhances the accuracy of tool anomaly identification, and can more accurately reflect the change of the tool state; by comprehensively combining the vibration fluctuation coefficient and the frequency domain fluctuation coefficient, the present application constructs a tool weight. The beneficial effect is that the tool weight reflects the wear degree of the tool in each period, can weaken the influence of the gradual change feature of the vibration data on the accuracy of anomaly monitoring, and improve the accuracy of anomaly detection. By comprehensively analyzing the fluctuation of the vibration data in the time domain and the frequency domain, the present application assigns weights to the vibration data at each acquisition moment, and combines the anomaly detection algorithm to monitor the wear state of the tool, weakens the influence of the normal gradual change feature of the vibration data of the tool on the anomaly detection, and improves the accuracy of tool state detection. Description of the Drawings

[0031] To more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0032] Figure 1 It is a flowchart of the steps of a tool state indication method for multi-tool machining of a variable pitch screw provided by an embodiment of the present application;

[0033] Figure 2 It is a schematic diagram of the vibration fluctuation coefficient extraction process provided by an embodiment of the present application;

[0034] Figure 3 It is a schematic diagram of the tool weight acquisition process provided by an embodiment of the present application;

[0035] Figure 4 It is a schematic diagram of the tool state detection process provided by an embodiment of the present application. Detailed implementation manners

[0036] In order to further elaborate on the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, details the specific implementation manners, structures, features and effects of the tool state indication method and indication device for multi-tool machining of a variable pitch screw according to the present application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0037] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present application belongs.

[0038] The following specifically describes the specific solutions of the tool state indication method and indication device for multi-tool machining of a variable pitch screw provided by the present application in conjunction with the accompanying drawings.

[0039] Please refer to Figure 1 , which shows a flowchart of the steps of a tool state indication method for multi-tool machining of a variable pitch screw provided by an embodiment of the present application. The method includes the following steps:

[0040] Step S1: During the multi-tool machining of the variable pitch screw, obtain the vibration data in all directions of each tool at each acquisition moment within a preset time period before and at the current moment.

[0041] During the multi-tool machining process of a variable pitch screw, acceleration sensors are installed at the rear half of each tool directly above each turning tool. During the operation of each tool, vibration data in all directions of each tool at each acquisition moment within a preset duration before and including the current moment is collected, that is, acceleration data. The data acquisition frequency is set to f, and all directions include the lateral, vertical, and longitudinal directions. The lateral direction is the direction in which the tool moves left and right around the machine tool spindle on the machine tool. The vertical direction is the direction in which the tool moves back and forth along the machine tool spindle on the machine tool. The longitudinal direction represents the direction in which the tool moves up and down perpendicular to the machine tool workbench on the machine tool.

[0042] It should be noted that the values of the preset duration and the data acquisition frequency f are both set manually. In this embodiment, the value of the preset duration is 1 s, and the value of the data acquisition frequency f is 1 kHz. Implementers can also set them according to specific situations, and this embodiment does not make special restrictions.

[0043] Step S2: Divide the preset duration into multiple time periods. According to the dispersion degree of all vibration data in any direction of each tool in each time period, determine the vibration characteristic value of each tool in each time period. And in combination with the difference of all vibration data in any direction of each tool between each time period and its adjacent previous time period, determine the vibration fluctuation coefficient of each tool in each time period.

[0044] The prior art establishes a tool condition monitoring model under fixed machining conditions and fixed tools to achieve the indication of tool conditions. However, due to the complexity during actual operation, there are various changes in machining conditions or tools. Therefore, it is impossible to conduct experiments and establish models in each mode. Usually, a tool compensation amount is constructed to achieve adaptability to changes in machining conditions or tools. At the same time, new errors are inevitably introduced. However, in the multi-tool machining of variable pitch screws, since multiple tools need to be used simultaneously when machining the same part and different pitches also need to be changed, a tool compensation amount needs to be introduced for the condition monitoring of each tool. That is equivalent to introducing multiple different error amounts when machining the same part, thus having a greater impact on the part quality.

[0045] To avoid the problems of constructing a fixed tool condition monitoring model and introducing multiple error amounts, according to the phenomenon that the change of tool conditions will cause vibration during the multi-tool machining of variable pitch screws, the tool conditions are monitored by detecting the outliers in the vibration data. However, due to the wear during the tool operation process, the tool conditions show a deteriorating trend, which in turn leads to a certain gradual change characteristic in the vibration situation. In traditional anomaly detection algorithms, the fluctuation trend of the data is not considered, so this gradual change characteristic will interfere with the determination of outliers, and further lead to the inability to accurately determine the tool conditions.

[0046] Therefore, in order to accurately monitor the status of the tool, the sensitivity of identifying abnormal status of the tool is enhanced by analyzing the time series changes of the vibration data. Specifically:

[0047] During variable pitch screw multi-tool processing, when the tool is in good condition, the tool surface is smooth and the friction is relatively small, so that the vibration of the tool is relatively small; as the tool is used, there is slight wear, at this time the friction of the tool increases relatively, and the vibration also increases relatively, but the amplitude of the change is small; as the tool wear increases, the defects on the tool surface become larger and larger, thereby increasing the friction of the tool and making the vibration of the tool larger and larger.

[0048] When processing the workpiece, the change trends of the tool in each movement direction are not exactly the same. The vertical direction is the forward direction of the tool and is also the main direction of the cutting force. As the tool wears, the cutting force will gradually increase, resulting in drastic changes in the vibration data in the forward direction; the longitudinal direction is the direction perpendicular to the workpiece surface. As the tool wear increases, the contact area and pressure between the tool and the workpiece will change, resulting in relatively drastic changes in the vibration in the longitudinal direction; and in the transverse direction, due to the relatively small cutting force, the vibration changes are relatively gentle.

[0049] Therefore, by analyzing the fluctuation of all vibration data in the same direction of the tool and the difference in vibration data in different directions, the wear of the tool can be judged, specifically:

[0050] (1) In order to analyze the changes in tool vibration data within a local time and better understand the wear of the tool, the preset time is evenly divided into multiple time periods.

[0051] (2) Further, according to the discrete degree of all vibration data of each tool in any direction in each time period, the vibration characteristic value of each tool in each time period is determined to judge the comprehensive fluctuation of vibration data in different directions, so as to identify whether the tool state is abnormal, specifically:

[0052] The discreteness of all vibration data in any direction of each tool in each time period is analyzed, and the cumulative sum of the discreteness in all directions is used as the vibration characteristic value of each tool in each time period, which is used to characterize the comprehensive fluctuation of the tool vibration data in each time period. The greater the discreteness of the vibration data, the greater the vibration characteristic value, indicating that the more intense the tool vibration is and the more serious the tool wear may be.

[0053] It should be noted that there are many methods to measure the dispersion degree of a set of data. In this embodiment, the variance of all vibration data in any direction of each tool at each time period is calculated to measure the dispersion degree of all vibration data in any direction of each tool at each time period. Implementers can also use other methods to measure the dispersion degree of data, such as the coefficient of variation or standard deviation. Regarding the selection of methods for measuring the dispersion degree of a set of data, this embodiment does not make special restrictions.

[0054] (3) Further, according to the vibration characteristic values of each tool at each time period, and combining the differences in all vibration data in any direction of each tool between each time period and its adjacent previous time period, the vibration fluctuation coefficient of each tool at each time period is determined to judge the difference degree of vibration between different directions, so as to judge the wear condition of the tool. Specifically:

[0055] The vibration fluctuation coefficient of tool j at time period i The expression is: ; In the formula, represents the vibration fluctuation coefficient of tool j at time period i; represents the vibration characteristic value of tool j at time period i; , , respectively represent the differences in all vibration data of tool j in the lateral, vertical, and longitudinal directions among all directions between time period i and its adjacent previous time period; represents a preset constant greater than 0, which is used to prevent the denominator from being 0, where The value of is set manually. In this embodiment, The value of is 0.01. On the premise of ensuring that the denominator is not 0 and does not overly affect the calculation result, implementers can also set it according to specific situations by themselves. This embodiment does not make special restrictions.

[0056] It should be noted that there are many methods to measure the differences between data groups. In this embodiment, the Euclidean distance of all vibration data of tool j in the lateral, vertical, and longitudinal directions between time period i and its adjacent previous time period is calculated to measure the differences in vibration data in the same direction between different time periods. Implementers can also use other methods to measure the differences between data groups, such as the Mahalanobis distance and Manhattan distance. Regarding the selection of methods for measuring the differences between data groups, this embodiment does not make special restrictions.

[0057] Among them, the calculation steps of the Euclidean distance are well-known technologies, and the specific calculation process will not be elaborated here.

[0058] Furthermore, it can be understood from the vibration fluctuation coefficients of each tool in each time period that when the i-th time period is in the working state of tool wear, the vibration fluctuations in all directions in the i-th time period are relatively large, and the more severe the tool wear, the more severe the vibration fluctuations. The greater the fluctuation difference between the i-th time period and the previous time period, that is 、 、 are all larger, indicating that the tool wear is more severe. At the same time, due to the change in vibration caused by tool wear, the degree of vibration change in the lateral and longitudinal directions is less than that in the vertical direction of the tool. Therefore is larger, and the more severe the tool wear, the larger the vibration characteristic value, then the vibration fluctuation coefficient is larger; on the contrary, when there is no tool wear or the wear is small, the vibration fluctuations in all directions in the i-th time period are small, and the fluctuation difference between the i-th time period and the previous time period is small, that is 、 、 are all smaller, and the vibration characteristic value is smaller, then the vibration fluctuation coefficient is smaller, indicating that the tool wear is more minor.

[0059] Preferably, the schematic diagram of the vibration fluctuation coefficient extraction process provided in this embodiment is as Figure 2 shown.

[0060] Step S3: For the vertical direction among all directions of each tool, determine the frequency-domain vibration value of each tool in each time period according to the difference in the frequency-domain fluctuation of all vibration data between each time period and the first time period, and combine the distribution of all vibration data in the frequency domain in any direction of each tool in each time period to determine the frequency-domain fluctuation coefficient of each tool in each time period.

[0061] The vibration fluctuation coefficient can identify the time-series change of vibration data and enhance the ability to identify tool abnormalities. However, due to the complexity of actual operation, in addition to tool abnormalities that can cause vibration fluctuations, other factors can also affect vibration data, such as changes in production parameters or production environment. If only relying on the vibration fluctuation coefficient to identify the tool state, it will cause certain misjudgments. Therefore, it is necessary to further judge by combining the frequency-domain information of vibration.

[0062] In the processing of variable pitch screws, when there is no tool wear or the wear is relatively minor, due to the relatively smooth tool surface, the friction force is small, and the cutting edge is also relatively sharp, so the main frequency of the vibration data in the frequency domain is small, and the amplitude fluctuation in the entire frequency domain is also relatively minor. As the tool wear degree increases, the friction force between the tool surface and the workpiece becomes larger and larger, resulting in the phenomenon of the main frequency shifting backward in the frequency domain of the vibration data. At the same time, the amplitude fluctuation in the entire frequency domain is also more obvious, that is, the frequency spectrum fluctuation is more chaotic.

[0063] When the tool has no wear or has relatively light wear, due to the small difference in the tool state, the fluctuation of the tool vibration is relatively small. Therefore, the main frequency of each segment does not strictly follow the changing trend of the main frequency moving backward, and there will be certain fluctuations. However, as the degree of tool wear increases, the fluctuation of the tool vibration becomes more and more obvious, resulting in the main frequency change in the corresponding frequency domain more strictly following the characteristic of the main frequency moving backward. Therefore, the more severe the tool wear, the greater the average difference in the main frequency change.

[0064] Therefore, by analyzing the fluctuation and distribution of vibration data in the frequency domain, the wear state of the tool can be further accurately judged. Specifically:

[0065] (1) For the vertical direction among all directions of each tool, according to the difference in the fluctuation of all vibration data in the frequency domain between each time period and the first time period, the frequency-domain vibration value of each tool at each time period is determined to judge the fluctuation of the vibration data in the frequency domain, so as to accurately judge the wear state of the tool. Specifically:

[0066] During the processing of the variable pitch screw, since the vertical direction of the tool is the most sensitive to tool wear, all vibration data in the vertical direction of each tool at each time period are used as the input of the time-frequency conversion algorithm, and the spectrogram in the vertical direction of each tool at each time period is output, and it is used as the spectrogram of each tool at each time period;

[0067] It should be noted that there are many commonly used time-frequency conversion algorithms. In this embodiment, the discrete Fourier transform is used to convert the time-domain data to the frequency domain. Implementers can also use other time-frequency conversion algorithms such as wavelet transform. Regarding the selection of the time-frequency conversion algorithm, no special limitation is made in this embodiment.

[0068] Furthermore, all peaks in the spectrogram of each tool at each time period are extracted, and all peaks are fitted to obtain a peak fitting curve. The difference between the integral area of the peak fitting curve of each tool at each time period and the integral area of the peak fitting curve of the first time period is used as the frequency-domain vibration value of each tool at each time period.

[0069] Among them, the first time period refers to the time period that is ranked first in chronological order after the preset time period division.

[0070] It should be understood that there are many methods to measure the difference between data. In this embodiment, the absolute value of the difference between the integral area of the peak fitting curve of each tool at each time period and the integral area of the peak fitting curve of the first time period is calculated to measure the difference in the integral area of the peak fitting curve between each time period and the first time period. Implementers can also use other methods to measure the difference between data such as ratios. Regarding the selection of the method to measure the difference between data, no special limitation is made in this embodiment.

[0071] In addition, it should be understood that there are many common fitting methods. In this embodiment, the non - linear least - squares fitting method is adopted to fit all peak data. Implementers can also adopt other fitting methods such as polynomial function fitting. Regarding the selection of the fitting method, this embodiment does not make special restrictions.

[0072] Among them, the discrete Fourier transform is a well - known technology in the field of signal processing, and the specific process of converting the time - domain signal to the frequency - domain will not be elaborated here; moreover, the non - linear least - squares fitting method, the process of peak extraction, and the process of calculating the integral area according to the fitting curve are all well - known technologies, and the specific implementation process will not be elaborated here.

[0073] From the frequency - domain vibration values of each tool in each time period, it can be understood that if the tool wear is more severe in the current time period, the difference between the fluctuation situation of this time period in the frequency domain and that of the first time period is greater. Therefore, the difference between the integral area of the peak fitting curve of the tool in this time period and that of the peak fitting curve of the first time period is greater, that is, the frequency - domain vibration value of the tool is greater; conversely, if the tool wear is more severe in the current time period, the difference between the fluctuation situation of this time period in the frequency domain and that of the first time period is greater. Therefore, the difference between the integral area of the peak fitting curve of the tool in this time period and that of the peak fitting curve of the first time period is greater, that is, the frequency - domain vibration value of the tool is greater.

[0074] Based on the frequency - domain vibration values of each tool in each time period, and in combination with the distribution of all vibration data of each tool in any direction in the frequency domain in each time period, determine the frequency - domain fluctuation coefficient of each tool in each time period to further determine the tool wear state, specifically as follows:

[0075] The frequency - domain fluctuation coefficient of tool j in time period i The expression is: ; where 、 respectively represent the main frequencies in the spectrograms of tool j in time period i and time period i - k; represents the frequency - domain vibration value of tool j in time period i.

[0076] From the frequency - domain fluctuation coefficients of each tool in each time period, it can be understood that if the tool wear in the current time period is more severe, the main frequency in the frequency domain of this time period is greater, and the more severe the tool wear, the more strictly it follows the characteristic of the main frequency shifting backward, and the more the main frequency shifts backward, that is , and is smaller. In addition, the more severe the tool wear, the greater the frequency - domain vibration value, and the greater the frequency - domain fluctuation coefficient; conversely, if the tool wear in the current time period is relatively slight, the main frequency in the frequency domain of this time period is smaller, and the more slight the tool wear, the main frequency is smaller, The larger the value is, and the smaller the vibration value in the frequency domain is, the smaller the fluctuation coefficient in the frequency domain is.

[0077] Step S4: Based on the correlation between the vibration fluctuation coefficient and the frequency domain fluctuation coefficient of each tool in all periods, and in combination with the vibration fluctuation coefficient and the frequency domain fluctuation coefficient, determine the tool weight of each tool in each period.

[0078] In the original anomaly detection algorithm, the change trend of data is not considered, and the weights of each data point are the same. Only the value of the data point itself is used to detect anomaly points, resulting in difficulty in accurately detecting anomaly points when facing a data set with a gradual change feature. Therefore, to avoid this problem, the change trend of tool vibration data is considered, and the vibration fluctuation coefficient and the frequency domain fluctuation coefficient are calculated according to the change characteristics of tool vibration data in the time domain and the frequency domain. When performing anomaly detection, according to the time domain and frequency domain changes of vibration data within each vibration segment, tool weights are assigned to each period, and then anomaly value detection is performed in combination with the tool weights. Specifically:

[0079] By analyzing the correlation between the vibration fluctuation coefficient and the frequency domain fluctuation coefficient of each tool in all periods, and in combination with the vibration fluctuation coefficient and the frequency domain fluctuation coefficient, determine the tool weight of each tool in each period to weaken the influence of the gradual change feature of vibration data on vibration data anomaly detection, thereby improving the accuracy of anomaly detection and more accurately monitoring the wear state of the tool. Specifically:

[0080] Calculate the normalized value of the product of the vibration fluctuation coefficient and the frequency domain fluctuation coefficient of each tool in each period;

[0081] Furthermore, analyze the correlation between the vibration fluctuation coefficient and the frequency domain fluctuation coefficient of each tool in all periods;

[0082] It should be noted that there are many methods to measure the correlation between data groups. In this embodiment, the Pearson correlation coefficient between the vibration fluctuation coefficient and the frequency domain fluctuation coefficient of each tool in all periods is calculated to measure the correlation between the vibration fluctuation coefficient and the frequency domain fluctuation coefficient. Implementers can also use other methods to measure the correlation between data groups, such as the Spearman correlation coefficient or the Kendall rank correlation coefficient. Regarding the selection of methods to measure the correlation between data groups, this embodiment does not make special restrictions.

[0083] Among them, the calculation steps of the Pearson correlation coefficient are well-known technologies, and the specific calculation process will not be elaborated here.

[0084] Furthermore, take the product of the normalized value and the correlation of each tool in each period as the tool weight of each tool in each period.

[0085] Further, it can be understood from the tool weights of each tool at each time period that when the vibration change is caused by tool wear, it will cause the vibration data to maintain a consistent change trend in the time domain and the frequency domain, that is, there is a large correlation between the vibration fluctuation coefficient and the frequency domain fluctuation coefficient of the tool. Moreover, the more severe the tool wear, the larger the vibration fluctuation coefficient and the frequency domain fluctuation coefficient. Therefore, the more severe the tool wear, the greater the tool weight; conversely, when the vibration change is not caused by tool wear or the tool wear is less severe, the corresponding tool weight is smaller.

[0086] Preferably, the schematic diagram of the process for obtaining the tool weight provided in this embodiment is as Figure 3 shown.

[0087] Step S5: Multiply the vibration data of each tool in all directions at each acquisition moment by the tool weight corresponding to the time period where the acquisition moment is located, and form the feature vector of each tool at each acquisition moment. Based on the distance between the feature vectors of any two acquisition moments of each tool, determine the distance formula between the feature vectors of any two acquisition moments of each tool, and combine the feature vectors of each tool at all acquisition moments and the anomaly detection algorithm to monitor the tool state of the multi-tool machining of the variable pitch screw.

[0088] When monitoring the tool state during the multi-tool machining of the variable pitch screw, the vibration data of each tool in all directions at each acquisition moment are used to form the vibration vector of each tool at each acquisition moment. Each vibration vector must belong to one of all time periods. Multiply the vibration vector of each tool at each acquisition moment by the product of the tool weight of the time period where the acquisition moment is located, and use it as the feature vector of each tool at each acquisition moment.

[0089] Further, based on the distance between the feature vectors of any two acquisition moments of each tool, determine the distance formula between the feature vectors of any two acquisition moments of each tool. Specifically:

[0090] The distance formula between the feature vectors of any two acquisition moments of each tool is: ; where is the distance between the feature vectors of tool j at acquisition moment and acquisition moment ; is the Euclidean distance between the feature vectors of tool j at acquisition moment and acquisition moment ; , are the tool weights of tool j at acquisition moment and acquisition moment respectively.

[0091] Take the feature vectors of each tool at the current moment and all the acquisition moments before it as the input of the LOF anomaly detection algorithm. Among them, take the distance formula between the feature vectors of any two acquisition moments of each tool as the distance formula in the LOF anomaly detection algorithm, and output the anomaly score values of each tool at the current moment and all the acquisition moments before it.

[0092] Take the anomaly score values of each tool at the current moment and all the acquisition moments before it as the input of the threshold segmentation algorithm, and output the segmentation threshold.

[0093] If the anomaly score value at the current moment is greater than the segmentation threshold, the tool state is abnormal, and a tool state prompt is given to indicate that the tool is severely worn. Otherwise, if the anomaly score value at the current moment is less than or equal to the segmentation threshold, the tool state is normal.

[0094] Among them, the LOF anomaly detection algorithm is a well-known technology, and its specific principle will not be elaborated here.

[0095] Preferably, the schematic diagram of the tool state detection process provided in this embodiment is as Figure 4 shown.

[0096] Based on the same inventive concept as the above method, the embodiment of the present application also provides a tool state indication device for multi-tool machining of variable pitch screws, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the tool state indication method for multi-tool machining of variable pitch screws described in any one of the above.

[0097] It should be noted that: the above sequence of embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above describes specific embodiments of this specification. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0098] Each embodiment in this specification is described in a progressive manner. The same or similar parts between each embodiment can be referred to each other, and the key points of each embodiment are to illustrate the differences from other embodiments.

[0099] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. A tool status indication method for variable pitch screw multi-tool machining, characterized in that: The method comprises the following steps: During variable pitch screw multi-tool machining, the vibration data of each tool in all directions at the current moment and at each acquisition moment within a preset time period before it is obtained; The preset time is divided into multiple time periods, and the vibration characteristic value of each tool in each time period is determined according to the discrete degree of all vibration data of each tool in any direction in each time period, and the vibration fluctuation coefficient of each tool in each time period is determined by combining the difference of all vibration data of each tool in any direction between each time period and its adjacent previous time period; For the vertical direction in all directions of each tool, the frequency domain vibration value of each tool in each time period is determined according to the difference in the frequency domain fluctuation of all vibration data between each time period and the first time period, and the frequency domain fluctuation coefficient of each tool in each time period is determined in combination with the distribution of all vibration data in the frequency domain in any direction of each tool in each time period; Calculate the normalized value of the product of the vibration fluctuation coefficient and the frequency domain fluctuation coefficient of each tool in each time period; Analyze the correlation between the vibration fluctuation coefficient and the frequency domain fluctuation coefficient of each tool at all time periods; The product of the normalized value and the correlation of each tool in each time period is used as the tool weight of each tool in each time period; The vibration data of each tool in all directions at each acquisition moment are multiplied by the tool weight corresponding to the time period of the acquisition moment to form the feature vector of each tool at each acquisition moment. Based on the distance between the feature vectors of any two acquisition moments of each tool, the distance formula between the feature vectors of any two acquisition moments of each tool is determined. The feature vector of each tool at the current moment and all previous acquisition moments is used as the input of the LOF anomaly detection algorithm, where the distance formula between the feature vectors of each tool at any two acquisition moments is used as the distance formula in the LOF anomaly detection algorithm, and the anomaly score value of each tool at the current moment and each previous acquisition moment is output; The abnormal score value of each tool at the current moment and all previous acquisition moments is used as the input of the threshold segmentation algorithm, and the segmentation threshold is output; If the abnormal score value at the current moment is greater than the segmentation threshold, the tool state is abnormal, otherwise, the tool state is normal.

2. The tool status indication method for variable pitch screw multi-tool machining according to claim 1, characterized in that: The vibration characteristic value of each tool in each time period is the cumulative sum of the discrete degrees of vibration data of each tool in all directions in each time period.

3. The tool status indication method for variable pitch screw multi-tool machining according to claim 1, characterized in that: The expression of the vibration fluctuation coefficient of each tool in each time period is: ; In the formula, represents the vibration fluctuation coefficient of tool j in time period i; represents the vibration characteristic value of tool j in time period i; , , Respectively represent the differences of all vibration data of tool j in the horizontal, vertical and longitudinal directions in all directions between time period i and its adjacent previous time period; Indicates a preset constant greater than 0.

4. The tool status indication method for variable pitch screw multi-tool machining according to claim 1, characterized in that: The method for determining the frequency domain vibration value of each tool in each time period is as follows: All vibration data in the vertical direction of each tool in each time period are used as the input of the time-frequency conversion algorithm, and the frequency spectrum of each tool in the vertical direction in each time period is output, and it is used as the frequency spectrum of each tool in each time period; All peaks in the spectrum diagram of each tool in each time period are extracted, and all peaks are fitted to obtain the peak fitting curve. The difference between the integral area of ​​the peak fitting curve of each tool in each time period and the integral area of ​​the peak fitting curve of the first time period is taken as the frequency domain vibration value of each tool in each time period.

5. The tool status indication method for variable pitch screw multi-tool machining according to claim 4, characterized in that: The expression of the frequency domain fluctuation coefficient of each tool in each time period is: ; In the formula, represents the frequency domain fluctuation coefficient of tool j in time period i; , represent the main frequencies in the spectrum of tool j in time period i and time period ik respectively; Represents the frequency domain vibration value of tool j in time period i.

6. The tool status indication method for variable pitch screw multi-tool machining according to claim 1, characterized in that: The distance formula between the feature vectors of each tool at any two acquisition moments is: ; In the formula, is the time when tool j is collected and collection time The distance between the feature vectors; is the tool j at the time of collection and collection time The Euclidean distance between the feature vectors, , They are respectively the collection time of tool j and collection time Tool weight.

7. A tool status indication device for variable pitch screw multi-tool machining, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the tool status indication method for variable pitch screw multi-tool machining as described in any one of claims 1-6 are implemented.

Images