Automatic adjustment method for equipment state time sequence data over-limit threshold under variable working conditions

By collecting equipment historical data in the operation and maintenance platform and using box diagram noise reduction processing, the equipment's over-limit threshold is automatically updated, which solves the problem of difficulty in updating over-limit thresholds in equipment operation and maintenance of metallurgical enterprises, and improves the online data analysis of equipment Accuracy and operation and maintenance efficiency.

CN119003979APending Publication Date: 2024-11-22SHANGHAI BAOSTEEL IND TECHNOLOGICAL SERVICE

Patent Information

Application Number
CN202410839454.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-26
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

In equipment operation and maintenance, due to equipment aging and environmental factors, it is difficult for the existing technology to effectively update the limit threshold of equipment online data, resulting in frequent false alarms, which increases the workload of the operation and maintenance platform and the repetition of false alarms. Appear.

Method used

The operation and maintenance platform collects equipment historical status timing data, uses box diagrams to perform noise reduction processing, calculates interquartiles and interquartile distance, sets an automatic update limit threshold, reduces the impact of noise, and automatically updates it to In the operation and maintenance platform, automatic adjustment of equipment exceeding limit threshold is realized.

Benefits of technology

It improves the model coverage rate of the intelligent operation and maintenance platform, reduces the false alarm rate, improves the accuracy of online data application and analysis of equipment, reduces the workload of manual maintenance, and realizes adaptive adjustment of exceeding the limit threshold.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119003979A_ABST
    Figure CN119003979A_ABST
Patent Text Reader

Abstract

The invention discloses a method for automatically adjusting an over-limit threshold value of equipment state time sequence data under variable working conditions, and the method comprises the steps: collecting the historical state time sequence data of equipment through an operation and maintenance platform to form a data sample, and carrying out the noise reduction of the data sample through a box diagram; respectively calculating a lower quartile Q1 and an upper quartile Q3 of the box diagram, and determining an upper limit value and a lower limit value of the box diagram for filtering noisy points; and setting an attention value notice of the data sample by using the data after denoising, setting the attention value notice as a recommended over-limit threshold value, automatically updating the attention value notice to the operation and maintenance platform, associating the attention value notice with equipment related data configured on the operation and maintenance platform, and automatically updating the parameter setting of the over-limit threshold value of the equipment. According to the method, the single-parameter overrun threshold values of the online data of all the devices can be automatically set in batches, the model coverage rate of the intelligent operation and maintenance platform is increased, the false alarm rate of the model is reduced, and the accuracy of online data application and analysis of the devices of the intelligent remote operation and maintenance platform is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of equipment online data application and analysis, and in particular to a method for automatically adjusting an over-limit threshold of equipment status time series data under variable working conditions. Background Art

[0002] Usually, metallurgical enterprises have equipment remote operation and maintenance platforms for equipment operation and maintenance. As the operation and maintenance platform runs, the number of connected devices and data items generated increases. The operation and maintenance platform reflects the abnormal status of the equipment through online data. It is necessary to set an over-limit threshold for the online data. Equipment that exceeds the threshold can generate an alarm, thereby reminding the alarm processing personnel of the operation and maintenance platform to pay attention to the relevant equipment in time, so that the abnormal equipment can be handled in the first time, reducing and avoiding production cuts and shutdowns caused by equipment failures.

[0003] The over-limit thresholds for online data are usually set by inspection personnel in related industries based on their experience. Inspection personnel update online data thresholds including vibration, pressure, flow, current, etc. from time to time based on their own experience and historical data trends of the equipment. Different types of equipment require inspection personnel with different professions to make judgments and settings.

[0004] Due to equipment aging and environmental factors, over-limit thresholds also need to be constantly updated and reset. Inspection personnel only have time to handle the maintenance of over-limit thresholds in their daily work. They need to manually observe the operating trends of each device with the naked eye, and it is difficult to set the frequency of regular updates and resets, which results in a huge workload.

[0005] At the same time, some over-limit thresholds that have been set according to inspection standards also have a large number of false alarms. Tens of thousands of false alarms are generated due to unreasonable over-limit threshold settings. After confirming that the equipment is normal, these false alarms need to be eliminated one by one on the operation and maintenance platform. The operation volume is also large, and the recurrence of some invalid alarms cannot be avoided. Therefore, it is necessary to continuously update the over-limit threshold settings to avoid the generation of invalid alarms, thereby reducing the workload of inspection personnel. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a method for automatically adjusting the over-limit threshold of equipment status time series data under variable working conditions, which can automatically set the single parameter over-limit threshold of all equipment online data in batches, improve the model coverage of the intelligent operation and maintenance platform, reduce the false alarm rate of the model, and improve the accuracy of online data application and analysis of equipment on the intelligent remote operation and maintenance platform.

[0007] In order to solve the above technical problems, the method for automatically adjusting the threshold value of the equipment status time series data under variable working conditions of the present invention comprises the following steps: Step 1: Use the operation and maintenance platform to collect the equipment historical status time series data to form data samples, and use the box plot to reduce the noise of the data samples; Step 2: Calculate the lower quartile Q1 and upper quartile Q3 of the box plot respectively, determine the maximum and minimum values ​​of the inner and outer limits of the box plot, and use the four boundary lines of the box plot to judge the noise points. The noise points beyond the boundary lines are filtered out and do not participate in the subsequent threshold calculation. Among them, the maximum value of the inner limit = Q3 + 1.5IQR; Inner limit minimum = Q3 - 1.5IQR; Outer limit maximum = Q3 + 3IQR; Outer limit minimum = Q3 - 3IQR; IQR is interquartile range, IQR=Q3-Q1; Step 3: Use the data after removing noise from the box plot to set the notice value for judging abnormal equipment status. According to expert experience and test results, the multiple of the interquartile range (IQR) is determined to be 2.5, so the notice value is notice = Q3 + 2.5IQR. Step 4: Set the notice value to the recommended over-limit threshold, and automatically update it to the operation and maintenance platform, associate it with the device-related data configured on the operation and maintenance platform, and automatically update the over-limit threshold parameter settings of the device.

[0008] Furthermore, the frequency of automatic update of the device's over-limit threshold parameter settings is 4 times / year. Since the automatic adjustment method of the over-limit threshold of the equipment status time series data under variable working conditions of the present invention adopts the above technical scheme, that is, the method collects the equipment historical status time series data through the operation and maintenance platform to form a data sample, and uses the box plot to perform noise reduction processing on the data sample; respectively calculates the lower quartile Q1 and the upper quartile Q3 of the box plot, determines the maximum and minimum values ​​of the inner and outer limits of the box plot, and is used to filter out noise points; uses the data after noise points to set the notice value notice for judging the abnormal equipment status, sets the notice value notice to the recommended over-limit threshold, and automatically updates it to the operation and maintenance platform, associates it with the equipment-related data configured on the operation and maintenance platform, and automatically updates the over-limit threshold parameter setting of the equipment. The method can automatically set the single parameter over-limit threshold of all equipment online data in batches, improve the model coverage of the intelligent operation and maintenance platform, reduce the false alarm rate of the model, and improve the accuracy of the application and analysis of the equipment online data of the intelligent remote operation and maintenance platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments: Figure 1 Schematic diagram of the box plot in this method. DETAILED DESCRIPTION

[0010] The method for automatically adjusting the threshold value of the equipment status time series data exceeding the limit under variable working conditions of the present invention comprises the following steps: Step 1: Use the operation and maintenance platform to collect the equipment historical status time series data to form data samples, and use the box plot to reduce the noise of the data samples; Step 2: Calculate the lower quartile Q1 and upper quartile Q3 of the box plot respectively, determine the maximum and minimum values ​​of the inner and outer limits of the box plot, and use the four boundary lines of the box plot to judge the noise points. The noise points beyond the boundary lines are filtered out and do not participate in the subsequent threshold calculation. Among them, the maximum value of the inner limit = Q3 + 1.5IQR; Inner limit minimum = Q3 - 1.5IQR; Outer limit maximum = Q3 + 3IQR; Outer limit minimum = Q3 - 3IQR; IQR is interquartile range, IQR=Q3-Q1; Step 3: Use the data after removing noise from the box plot to set the notice value for judging abnormal equipment status. According to expert experience and test results, the multiple of the interquartile range (IQR) is determined to be 2.5, so the notice value is notice = Q3 + 2.5IQR. Step 4: Set the notice value to the recommended over-limit threshold, and automatically update it to the operation and maintenance platform, associate it with the device-related data configured on the operation and maintenance platform, and automatically update the over-limit threshold parameter settings of the device.

[0011] Preferably, the frequency of automatic update of the over-limit threshold parameter setting of the equipment is 4 times / year. Due to the variable operating conditions of the equipment, it is necessary to update the over-limit threshold within a certain period of time to continuously make the over-limit threshold meet the latest equipment status feedback. Therefore, the calculation and update of the over-limit threshold must be timely. This method uses the method of calculating and updating 4 times a year to automatically and regularly calculate and adjust the over-limit threshold to keep it within the effective setting range. By updating the over-limit threshold setting of a certain device four times, it was found that the maximum rate of change can reach 15%, proving that the application can achieve adaptive adjustment of the over-limit threshold. The over-limit threshold can be dynamically updated in an adaptive manner, effectively reducing false alarms caused by changes in the data itself.

[0012] This method takes the actual value of the tension of the cold rolling mill's tension roller motor-loop tension and the actual value of the uncoiling tension of the uncoiling machine body as examples. The historical data trend fluctuates greatly and has many abnormal noise points, so it cannot be used as a reference for trend judgment. By looking at its distribution diagram, it can be clearly observed that the data distribution is relatively discrete. By calculating its standard deviation, the actual value of tension σ1=969.583 and the actual value of uncoiling tension σ2=882.39. The larger the standard deviation, the greater the fluctuation of the data, and the calculated mean has a certain error.

[0013] The above data distribution characteristics all show an asymmetric skewed state. In statistics, the difference between the mode and the mean can be used as one of the indicators of distribution skewness: if the mean is greater than the mode, it is called positive skewness (or right skewness); otherwise, it is called negative skewness (or left skewness). Data with this characteristic are suitable for noise reduction using box plots.

[0014] like Figure 1 As shown in the figure, in the box plot, there is a line in the middle of the box, which represents the median M of the data. The upper and lower bottoms of the box are the upper quartile Q3 and the lower quartile Q1 of the data, which means that the box contains 50% of the data. Therefore, the height of the box reflects the degree of fluctuation of the data to a certain extent, and the upper and lower edges represent the maximum and minimum values ​​of the group of data; there will be some data points outside the box, which can be understood as outliers in the data. Outliers outside the inner limit are moderately abnormal, and outliers outside the outer limit are extremely abnormal. The maximum / minimum values ​​of the inner limit and the maximum / minimum values ​​of the outer limit are used to judge the noise points.

[0015] The maximum value of the inner limit = Q3 + 1.5IQR, the minimum value of the inner limit = Q3 - 1.5IQR, the maximum value of the outer limit = Q3 +3IQR, and the minimum value of the outer limit = Q3 - 3IQR. If the value is greater than the maximum value of the inner limit and less than the maximum value of the outer limit, or less than the minimum value of the inner limit and greater than the minimum value of the outer limit, it is called a mild outlier. If it is greater than the maximum value of the outer limit or less than the minimum value of the outer limit, it is called an extreme outlier.

[0016] After using the box plot to denoise the data, the trend is smooth and evenly distributed. This trend can be used as a basis for setting the over-limit threshold. The distribution diagram is also more concentrated and more even than before denoising. The standard deviation after denoising is the actual value of tension σ1=428.375 and the actual value of unwinding tension σ2=769.096, which are both concentrated to a certain extent compared with before denoising.

[0017] The relatively suitable over-limit threshold is calculated by the mean and variance of the data after denoising. When the real-time data of the equipment exceeds the over-limit threshold for a certain number of times, the operation and maintenance platform will trigger an alarm, thereby effectively monitoring the trend of the equipment data.

[0018] This method can provide updated recommendations for over-limit thresholds in a timely and regular manner, and provide mathematically logical over-limit threshold recommendations based on historical data, thereby reducing the workload of manually setting over-limit thresholds and improving the efficiency of intelligent operation and maintenance of equipment.

Claims

1. A method for automatically adjusting the threshold of equipment status time series data exceeding the limit under variable working conditions. Features The steps include: Step 1: Use the operation and maintenance platform to collect the equipment historical status time series data to form data samples, and use the box plot to perform noise reduction on the data samples; Step 2: Calculate the lower quartile Q1 and upper quartile Q3 of the box plot respectively, determine the maximum and minimum values ​​of the inner and outer limits of the box plot, and use the four boundary lines of the box plot to judge the noise points. The noise points beyond the boundary lines are filtered out and do not participate in the subsequent threshold calculation. Among them, the maximum value of the inner limit = Q3 + 1.5IQR; Inner limit minimum = Q3 - 1.5IQR; Outer limit maximum = Q3 + 3IQR; Outer limit minimum = Q3 - 3IQR; IQR is interquartile range, IQR=Q3-Q1; Step 3: Use the data after removing noise from the box plot to set the notice value for judging abnormal equipment status. According to expert experience and test results, the multiple of the interquartile range (IQR) is determined to be 2.5, so the notice value is notice = Q3 +2.5IQR. Step 4: Set the notice value to the recommended over-limit threshold, and automatically update it to the operation and maintenance platform, associate it with the device-related data configured on the operation and maintenance platform, and automatically update the over-limit threshold parameter settings of the device.

2. According to the method for automatically adjusting the threshold value of the equipment status time series data exceeding the limit under variable working conditions in claim 1, Features: The frequency of automatic update of the device's over-limit threshold parameter settings is 4 times / year.

Citation Information

Patent Citations

  • Historical operation data-based equipment alarm threshold calculation method

    CN110415494A

  • Power consumption anomaly detection method based on depth auto-encoder Gaussian mixture model

    CN113902581A

Cited By

  • Sewage treatment method and system

    CN120736596A