Operation and maintenance state monitoring method of cutting equipment

Abstract

The invention relates to the technical field of data processing, in particular to a method for monitoring the operation and maintenance state of cutting equipment, and the method comprises the steps: obtaining an operation parameter sequence collected by a sensor during the operation of the cutting equipment; constructing a sliding window, traversing the operation parameter sequence, and in the traversing process, adaptively adjusting the window length by comparing the dispersion degree of data in adjacent windows; and carrying out anomaly detection on data in the current window by utilizing an improved Z-Score algorithm, taking the length of the current window as a sliding step length after a detection result is obtained so as to continuously obtain a new window to repeat an anomaly detection process, and monitoring the operation and maintenance state of the equipment based on the detection result. According to the invention, the accuracy of the anomaly detection result can be improved, and the accurate monitoring of the operation and maintenance state of the cutting equipment is realized.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and more specifically, to a method for monitoring the operation and maintenance status of a cutting device. Background Art

[0002] In the manufacturing industry, especially in the production of automotive parts, cutting equipment plays a vital role and is often used for complex cutting, grinding, milling and other operations in the processing process. With the development of industrialization and automation, modern cutting equipment needs to have higher precision and stability to ensure production efficiency and product quality. However, during long-term operation, the equipment may have problems such as wear, aging or failure. If it is not discovered in time, it will lead to equipment shutdown, production delays, and even affect the quality of the final product. Therefore, it is very important to monitor the operation and maintenance status of the cutting equipment.

[0003] Z-Score (standard score) is a statistic that measures the degree to which a data point deviates from its average value. It is often used for anomaly detection in monitoring systems. It identifies anomalies by calculating the deviation between the current equipment status data and historical data. In equipment status monitoring, data is usually presented in the form of a time series, indicating the continuous operation status of the equipment over a period of time. The operating parameters of the cutting equipment will fluctuate over time. Therefore, in order to more accurately detect anomalies in the equipment operating parameters, sliding window technology came into being.

[0004] However, when monitoring the operation and maintenance status of cutting equipment, the fixed-length sliding window is difficult to capture local data features due to the real-time changes in parameters such as load and speed during the cutting process. At the same time, changes in working conditions (such as material switching and tool wear) cause data distribution shifts, making it impossible for the global mean to accurately reflect the current status, which in turn affects the anomaly detection results based on the Z-Score algorithm, resulting in deviations or even false alarms, thereby reducing the monitoring effect. Summary of the invention

[0005] In order to solve the problem that the traditional fixed window and global mean cannot adapt to the dynamic changes of working conditions, resulting in large deviations in abnormal detection results, and thus affecting the monitoring effect of the equipment operation and maintenance status, the present invention provides a method for monitoring the operation and maintenance status of cutting equipment. The method includes: Acquire an operating parameter sequence collected by a sensor when the cutting device is running; Construct a sliding window and traverse the running parameter sequence. During the traversal process, the window length is adaptively adjusted by comparing the discreteness of the data in adjacent windows. The improved Z-Score algorithm is used to detect anomalies in the data in the current window. When the detection results are obtained, the length of the current window is used as the sliding step to continuously obtain new windows to repeat the anomaly detection process, and the operation and maintenance status of the equipment is monitored based on the detection results. The process of using the improved Z-Score algorithm for anomaly detection is as follows: Calculate the sensitivity coefficient of the current window, which represents the system's sensitivity to data changes in the current window; Get the data mean in the current window, use the sensitivity coefficient to weight the average difference between the data in the current window and the data mean, correct the data mean based on the weighted value, the correction value is positively correlated with the weighted value, and use the correction value as the mean when the Z-Score algorithm standardizes the data in the current window to obtain the anomaly detection results of each data in the current window.

[0006] The present invention can adapt to the dynamic changes of equipment operating parameters by constructing an adaptive mechanism of sliding windows, so as to accurately capture the local characteristics of the data and provide a data basis for anomaly detection. The process of correcting the data mean can reduce the influence of extreme values ​​on the data mean, thereby improving the accuracy of the Z-Score algorithm's detection results of equipment operating parameter anomalies and realizing accurate monitoring of the equipment operation and maintenance status.

[0007] Preferably, by comparing the discreteness of data in adjacent windows, the window length is adaptively adjusted to satisfy the following relationship: ; In the formula, For the The length of the window; is the preset reference length; For the The degree of dispersion of data within a window; For the The degree of dispersion of data within a window; is the floor rounding function; is the window ordinal number.

[0008] The present invention realizes window length adaptation through discrete degree comparison, taking into account both stability and response speed.

[0009] Preferably, the method for obtaining the discrete degree includes: Get the interquartile range and median of the data in any window, calculate the difference between the interquartile range and the median, and get the degree of dispersion of the data in any window.

[0010] The calculation method of the discrete degree provided by the present invention can quickly capture the change of distribution shape when the working condition changes (such as material switching, load adjustment) cause the data distribution to shift, and provide a reliable basis for adjusting the window length.

[0011] Preferably, calculating the difference between the interquartile range and the median to obtain the degree of dispersion of the data in any window includes: Calculate the ratio of the interquartile range of the data in any window to the median of the data in the window to obtain the degree of dispersion of the data in the window.

[0012] Preferably, calculating the difference between the interquartile range and the median to obtain the degree of dispersion of the data in any window also includes: Calculate the interquartile range of the data in any window and the absolute value of the difference between it and the median of the data in the window to obtain the degree of dispersion of the data in the window.

[0013] Preferably, the sensitivity coefficient satisfies the following relationship: ; In the formula, For the The sensitivity coefficient of the window; For the The median of the data in the window; For the The median deviation of the data in the window; For the The variance of the data within the window; is the normalization function; is the maximum value of the variance of the data in each historical window.

[0014] The present invention sets a larger sensitivity coefficient for windows with concentrated and stable data distribution, which can improve the system's sensitivity to data with small fluctuations, and sets a smaller sensitivity coefficient for windows with dispersed and large fluctuations, which can avoid false alarms caused by large fluctuations in the data itself.

[0015] Preferably, correcting the data mean based on the weighted value comprises: The weighted value and the data mean are summed to obtain a corrected value of the data mean.

[0016] Preferably, obtaining the anomaly detection result of each data in the current window includes: Calculate the difference between any data in the current window and the calibration value, and use the ratio of the difference to the standard deviation of the data in the target window as the standardized value of any data; The anomaly detection result of any data is obtained based on the comparison result between the standardized value and the preset anomaly threshold.

[0017] Preferably, obtaining an anomaly detection result of any data based on a comparison result of the standardized value and a preset anomaly threshold value includes: When the normalized value of any data is greater than the abnormal threshold, the data is determined to be abnormal data; when the normalized value of any data is less than or equal to the abnormal threshold, the data is determined to be normal data.

[0018] Preferably, the operating parameter sequence is one of a temperature sequence, a pressure sequence and a rotation speed sequence.

[0019] The type of operating parameter sequence selected by the present invention can reflect the operation and maintenance status of the equipment, thereby providing a reliable data basis for monitoring the operation and maintenance status of the equipment.

[0020] The present invention has the following effects: The present invention realizes the multi-scale feature capture of equipment operating parameters by dynamically adjusting the sliding window length. When data fluctuations intensify, the window automatically shrinks to improve real-time response capabilities and ensure timely detection of sudden anomalies; in the data stable stage, the window expands to enhance statistical stability and reduce random noise interference. Combined with the mean correction algorithm based on the sensitivity coefficient, the system can adaptively compensate for the biasing effect of extreme values ​​on the mean, so that the corrected mean is closer to the actual data distribution. This dual optimization mechanism significantly improves the robustness of the Z-Score algorithm to changes in operating conditions, improves the accuracy of anomaly detection results, and provides reliable technical support for predictive maintenance of equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 It is a schematic flow chart of the steps of a method for monitoring the operation and maintenance status of a cutting device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] 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 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0023] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0024] Reference Figure 1 , a method for monitoring the operation and maintenance status of a cutting device, comprising steps S1 to S3, specifically as follows: S1: Acquire the operating parameter sequence collected by the sensor when the cutting equipment is running.

[0025] In an exemplary embodiment of the present invention, the operating parameter sequence is one of a temperature sequence, a pressure sequence and a rotation speed sequence.

[0026] It should be noted that, during the operation of the cutting equipment, the equipment temperature, the internal pressure of the equipment and the equipment speed can reflect the operating status of the equipment, therefore, the equipment operation and maintenance status can be detected by analyzing an operating parameter of the equipment. Among them, the operating parameter selected by the present invention is the equipment speed.

[0027] Optionally, when the equipment temperature is used to monitor the operation and maintenance status of the equipment, a temperature sensor may be installed in the cutting equipment to collect temperature data during the operation of the equipment; When the internal pressure of the equipment is used to monitor the operation and maintenance status of the equipment, a pressure sensor can be installed in the cutting equipment to collect pressure data during the operation of the equipment; When the equipment speed is used to monitor the operation and maintenance status of the equipment, a magnetoelectric speed sensor, a Hall-type speed sensor, etc. can be installed in the cutting equipment to collect the speed data during the operation of the equipment.

[0028] It should be noted that when the device is just started, the stability of the data is poor. Therefore, the present invention collects the operating parameters of the device in real time at a fixed collection frequency, such as 1 Hz, 30 minutes after the device is started, so as to obtain the operating parameter sequence of the device.

[0029] S2: Build a sliding window and traverse the running parameter sequence. During the traversal process, the window length is adaptively adjusted by comparing the discreteness of the data in adjacent windows.

[0030] The degree of discreteness refers to the data that can reflect the discreteness of data distribution.

[0031] It should be noted that, since during the operation of the sliding window mechanism, only the complete data in the left adjacent window of the current window can be obtained, the present invention adaptively adjusts the window length by comparing the discreteness of the data in the adjacent windows, which means that the current window length is adaptively adjusted by comparing the discreteness of the data in the current window with the discreteness of the data in the left adjacent window.

[0032] In an exemplary embodiment of the present invention, the determination of the discreteness of data in any window can be achieved by the following steps: Get the interquartile range and median of the data in any window, calculate the difference between the interquartile range and the median, and get the degree of dispersion of the data in any window.

[0033] It should be noted that the interquartile range (IQR) is a statistic that describes the degree of data dispersion, which is equal to the difference between the upper quartile (Q3) and the lower quartile (Q1). The calculation process of the interquartile range of the data series is a prior art and will not be described in detail in this embodiment.

[0034] Optionally, when the difference between the interquartile range of the data in any window and the median of the data in the window is large, it means that the middle part of the data in the window (i.e., the part between Q1 and Q3) is far away from the median, which further indicates that the data in the window is not concentrated and the corresponding discreteness is relatively large.

[0035] In an exemplary embodiment of the present invention, the ratio of the interquartile range of the data in any window to the median of the data in the window may be calculated to obtain the degree of dispersion of the data in the window.

[0036] In another exemplary embodiment of the present invention, the absolute value of the difference between the interquartile range of the data in any window and the median of the data in the window may be calculated to obtain the degree of dispersion of the data in the window.

[0037] Next, the process of dynamically adjusting the sliding window length when the sliding window traverses the running parameter sequence is described in detail: For example, the window ordinal number of any window can be recorded as , and use any method to determine the discrete degree to calculate the discrete degree of the data in any window, as well as the left adjacent window of any window, that is, the first The discrete degree of data in a window is determined; then, the length of any window is dynamically adjusted by comparing the discrete degree of data in the window with the left adjacent window.

[0038] Specifically, by comparing the discreteness of data in adjacent windows, the window length is adaptively adjusted to satisfy the following relationship: ; In the formula, For the The length of the window; is a preset reference length. In this embodiment, the preset reference length is the length of the operating parameters collected at 15 consecutive sampling moments; For the The degree of dispersion of data within a window; For the The degree of dispersion of data within a window; is the floor rounding function; is the window ordinal number.

[0039] Among them, if The value of is greater than or equal to 1, indicating that The degree of dispersion within the window is less than or equal to the The degree of discreteness of the data in a window indicates that the data trend tends to be stable. At this time, increasing the window length can improve statistical stability. The value of is less than 1, indicating that The degree of dispersion within the window is greater than that of the The discreteness of the data in a window indicates that the discreteness of the data increases. At this time, reducing the window length can quickly capture data mutations to achieve rapid response.

[0040] At the same time, using the benchmark length as a reference value can ensure that the window length adjustment process has a reasonable starting benchmark, thereby avoiding drastic fluctuations in the window length. By rounding down, the window length is always kept as an integer, which not only ensures the smoothness of the adjustment, but also avoids excessive consumption of computing resources due to frequent jumps.

[0041] In particular, the adaptive length of the first window is the preset reference length.

[0042] S3: Use the improved Z-Score algorithm to perform anomaly detection on the data in the current window. When the detection result is obtained, the length of the current window is used as the sliding step to continuously obtain new windows to repeat the anomaly detection process, and monitor the operation and maintenance status of the equipment based on the detection results.

[0043] It should be noted that during the operation of the cutting equipment, changes in operating conditions (such as material switching, tool wear, etc.) will cause the distribution of equipment operating parameters to shift. This shift causes deviations in the Z-Score normalization results based on the global mean, which in turn affects the accuracy of anomaly detection. In addition, since the global mean is easily affected by extreme values, the effect of anomaly detection is further weakened, resulting in the inability to obtain accurate anomaly detection results. Therefore, the present invention improves the traditional Z-Score normalization process, and the specific improvements are: using the sensitivity coefficient to weight the average difference between the data in the current window and the data mean, and correcting the data mean based on the weighted value, so as to perform Z-Score normalization based on the corrected value of the data mean.

[0044] Among them, the present invention only improves the determination of the data mean in the traditional Z-Score standardization process, and does not improve other contents, such as the standardization process.

[0045] Specifically, the process of using the improved Z-Score algorithm for anomaly detection includes the following steps: Step 1: Calculate the sensitivity coefficient of the current window. The sensitivity coefficient represents the system's sensitivity to data changes in the current window. Specifically, the sensitivity coefficient of any window satisfies the following relationship: ; In the formula, For the The sensitivity coefficient of the window; For the The median of the data in the window; For the The median deviation of the data in the window; For the The variance of the data within the window; is the normalization function; is the maximum value of the variance of the data in each historical window.

[0046] The median absolute deviation (MAD) is a statistic that measures the degree of data dispersion, and represents the median of the absolute deviations of the data points from the median. It should be noted that the determination process of the median deviation is a prior art, and this embodiment will not be described in detail.

[0047] Optional, when Larger, and When it is smaller, the The data in a window is relatively concentrated and stable. At this time, the system's sensitivity to data changes in the window should be enhanced, so that the boundary can be tightened, making it easier to detect subtle data changes. The corresponding sensitivity coefficient of the window is relatively large.

[0048] when Larger, and When it is smaller, the The relative fluctuation of the data in a window is large. At this time, the system's sensitivity to the changes in the data in the window should be reduced, so as to relax the judgment boundary and avoid false alarms caused by large fluctuations in the data itself.

[0049] Step 2: Obtain the data mean in the current window, use the sensitivity coefficient to weight the average difference between the data in the current window and the data mean, and correct the data mean based on the weighted value, where the correction value is positively correlated with the weighted value; In an exemplary embodiment of the present invention, the determination of the correction value of the data mean in any window can be achieved by the following steps: The weighted value and the data mean are summed to obtain a corrected value of the data mean.

[0050] Specifically, the correction value of the data mean in any window satisfies the following relationship: ; In the formula, For the Correction value of the data mean within a window; For the The mean of the data in the window; For the The sensitivity coefficient of the window; For the In the window The value of the data; For the The mean of the data in the window; For the The amount of data in a window; is the absolute value symbol.

[0051] in, Reflects the The larger the value is, the more the mean of the data in the window is affected by the extreme value, and the more correction is needed to reduce the impact of the extreme value.

[0052] when The larger the value, the better the system The more sensitive the data changes in a window, the The value of can more accurately reflect the impact of extreme values ​​on the data, so we can keep The value of remains unchanged, or is only adjusted to a small extent. In this way, the correction amplitude of the data mean in the window determined based on the influence of the extreme value on the data mean in the window can be closer to the actual data distribution, so as to more accurately correct the data mean in the window and ensure the accuracy of the correction value.

[0053] In another embodiment, the product of the mean value of the data in any window and the weighted value of the average difference between the data in the window and the mean value of the data in the window may be used as the correction value of the mean value of the data in the window.

[0054] Step 3: Use the correction value as the mean value when the Z-Score algorithm is used to standardize the data in the current window to obtain the anomaly detection results of each data in the current window.

[0055] In an exemplary embodiment of the present invention, the determination of the abnormality detection result of the data can be achieved by the following steps: (1) Calculate the difference between any data in the current window and the calibration value, and use the ratio of the difference to the standard deviation of the data in the target window as the standardized value of any data; Specifically, the standardized value of any data in any window satisfies the following relationship: ; In the formula, For the In the window The standardized value of the data; For the In the window The value of the data; For the Correction value of the data mean within a window; For the The standard deviation of the data within the window.

[0056] It should be noted that in the process of using the Z-Score algorithm for anomaly detection, if the standardized value of any data is larger, the data is more likely to be abnormal data, and then based on this feature, a fixed threshold can be set to filter out abnormal data.

[0057] (2) Obtain the anomaly detection result of any data based on the comparison result between the standardized value and the preset anomaly threshold.

[0058] In an exemplary embodiment of the present invention, the abnormality detection result may be determined based on a preset abnormality threshold through the following steps: When the normalized value of any data is greater than the abnormal threshold, the data is determined to be abnormal data; when the normalized value of any data is less than or equal to the abnormal threshold, the data is determined to be normal data.

[0059] Optionally, the abnormality threshold can be set to 1. When the normalized value of any data in any window is greater than 1, the data is determined to be abnormal data; when the normalized value of another data in the window is less than or equal to 1, the other data is determined to be normal data, thereby determining the abnormality detection results of each data in the window.

[0060] Furthermore, after obtaining the abnormality detection results of each data in the current window, the length of the current window can be used as the sliding step length, and the method of adaptively adjusting the window length in step S2 can be used to determine the next window. Then, the improved Z-Score algorithm is used for abnormality detection to determine the abnormality detection results of each data in any window. Similarly, the abnormality detection results of each data in the operating parameter sequence are obtained, so that the operation and maintenance status of the cutting equipment can be monitored based on the abnormality detection results of each data, so as to facilitate subsequent maintenance management work.

[0061] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0062] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A method for monitoring the operation and maintenance status of a cutting device, characterized in that: include: Acquire an operating parameter sequence collected by a sensor when the cutting device is running; Construct a sliding window and traverse the running parameter sequence. During the traversal process, the window length is adaptively adjusted by comparing the discreteness of the data in adjacent windows. The improved Z-Score algorithm is used to detect anomalies in the data in the current window. When the detection results are obtained, the length of the current window is used as the sliding step to continuously obtain new windows to repeat the anomaly detection process, and the operation and maintenance status of the equipment is monitored based on the detection results. The process of using the improved Z-Score algorithm for anomaly detection is as follows: Calculate the sensitivity coefficient of the current window, which represents the system's sensitivity to data changes in the current window; Get the data mean in the current window, use the sensitivity coefficient to weight the average difference between the data in the current window and the data mean, correct the data mean based on the weighted value, the correction value is positively correlated with the weighted value, and use the correction value as the mean when the Z-Score algorithm standardizes the data in the current window to obtain the anomaly detection results of each data in the current window.

2. A method for monitoring the operation and maintenance status of a cutting device according to claim 1, characterized in that: By comparing the discreteness of data in adjacent windows, the window length is adaptively adjusted to satisfy the following relationship: ； In the formula, For the The length of the window; is the preset reference length; For the The degree of dispersion of data within a window; For the The degree of dispersion of data within a window; is the floor rounding function; is the window ordinal number.

3. The method for monitoring the operation and maintenance status of a cutting device according to claim 2, characterized in that: The method for obtaining the discrete degree includes: The interquartile range and median of the data in any window are obtained, and the difference between the interquartile range and the median is calculated to obtain the degree of dispersion of the data in any window.

4. The method for monitoring the operation and maintenance status of a cutting device according to claim 3, characterized in that: The calculating the difference between the interquartile range and the median to obtain the degree of dispersion of the data in any window includes: Calculate the ratio of the interquartile range of the data in any window to the median of the data in the window to obtain the degree of dispersion of the data in the window.

5. The method for monitoring the operation and maintenance status of a cutting device according to claim 3, characterized in that: The calculating the difference between the interquartile range and the median to obtain the degree of dispersion of the data in any window also includes: Calculate the interquartile range of the data in any window and the absolute value of the difference between it and the median of the data in the window to obtain the degree of dispersion of the data in the window.

6. The method for monitoring the operation and maintenance status of a cutting device according to claim 1, characterized in that: The sensitivity coefficient satisfies the following relationship: ； In the formula, For the The sensitivity coefficient of the window; For the The median of the data in the window; For the The median deviation of the data in the window; For the The variance of the data within the window; is the normalization function; is the maximum value of the variance of the data in each historical window.

7. A method for monitoring the operation and maintenance status of a cutting device according to claim 6, characterized in that: The step of correcting the data mean based on the weighted value comprises: A sum operation is performed on the weighted value and the data mean to obtain a corrected value of the data mean.

8. The method for monitoring the operation and maintenance status of a cutting device according to claim 1, characterized in that: The obtaining of the abnormality detection results of each data in the current window includes: Calculate the difference between any data in the current window and the correction value, and use the ratio of the difference to the standard deviation of the data in the target window as the standardized value of any data; According to the comparison result of the standardized value and the preset abnormal threshold, the abnormality detection result of any data is obtained.

9. A method for monitoring the operation and maintenance status of a cutting device according to claim 8, characterized in that: The obtaining the abnormality detection result of any data based on the comparison result of the standardized value and the preset abnormality threshold value includes: When the normalized value of any data is greater than the abnormal threshold, the data is determined to be abnormal data; when the normalized value of any data is less than or equal to the abnormal threshold, the data is determined to be normal data.

10. The method for monitoring the operation and maintenance status of a cutting device according to claim 1, characterized in that: The operating parameter sequence is one of a temperature sequence, a pressure sequence and a rotation speed sequence.

Images