A real-time monitoring system for faults of gas cabinet components

By collecting and analyzing the controller's historical and current current data, calculating standard predicted values, bias predicted values ​​and abnormality, the current offset misjudgment problem caused by degraded controller performance is solved, and the fault monitoring accuracy of gas cabinet controller is improved sex.

CN118777670BActive Publication Date: 2024-11-15德川电气有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411261866.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-10
Publication Date
2024-11-15
Estimated Expiration
2044-09-10

AI Technical Summary

Technical Problem

After using the controller for a period of time, the performance of the controller will decline, resulting in an upward or downward offset from the preset threshold, resulting in a misjudgment when judging whether the controller is abnormal by the preset fixed threshold.

Method used

The monitoring data acquisition module is used to obtain the controller's historical and current current data, and the final abnormality degree is calculated through standard predicted values, biased predicted values, predicted abnormality degree and trend abnormality degree, which is used to monitor the controller's faults in real time.

Benefits of technology

Improve the accuracy of online fault monitoring of gas cabinet controllers, avoid the problem of inapplicability of fixed thresholds due to changes in current range, and ensure the reliability and production efficiency of gas cabinets.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118777670B_ABST
    Figure CN118777670B_ABST
Patent Text Reader

Abstract

The present application relates to the field of data processing technology, and specifically to a real-time monitoring system for faults of gas cabinet component units, including: using the controller in the gas cabinet as a fault monitoring unit, obtaining the historical current data of the current collection date and the current data at the monitoring time; obtaining the predicted abnormality degree of the current data at the monitoring time of the current collection date based on the historical current data at the monitoring time of the current collection date; obtaining the trend abnormality degree based on the trend change of the current data between adjacent sampling times under the current collection date; obtaining the final abnormality degree based on the predicted abnormality degree and trend abnormality degree of the current data at the monitoring time of the current collection date; and performing controller fault monitoring based on the final abnormality degree of the current data at the monitoring time of the current collection date. The present application aims to solve the problem of errors existing when judging the abnormality of the gas cabinet controller by a fixed current threshold, and realize real-time monitoring of the range deviation of the working current caused by the aging of the controller.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a real-time monitoring system for failures of gas cabinet component units. Background Art

[0002] A gas cabinet is a type of gas cabinet used to store and supply gas, and is mainly used in laboratories, industrial production, medical and other fields. Its main function is to store compressed gas safely and conveniently, and to control the output flow and pressure of the gas through a controller as needed, to achieve the functions of pressure reduction, flow control, pressure monitoring, display and alarm, and to ensure that the gas cabinet can operate stably and efficiently under set conditions. In order to ensure the stable operation of the gas cabinet, it is necessary to monitor faults during the use of the gas cabinet.

[0003] Fault monitoring of gas cabinets is usually reflected in the real-time monitoring of different components in the gas cabinet. Gas cabinets include mixed gas proportioning cabinets, gas filling cabinets, special gas cabinets and other types. The components of each type are different. For example, the mixed gas proportioning cabinet includes controllers, flow regulators, mixers and other components; the gas filling cabinet includes integrated switches, controllers, explosion-proof valves and other components; the special gas cabinet includes pneumatic valves, pressure regulating valves, detectors, controllers and other components. Among them, the controller is one of the key components of the gas cabinet. Its role in the gas cabinet includes monitoring and controlling gas concentration, alarm, network detection, data recording and analysis, and control equipment.

[0004] After a period of use, the controller will face performance degradation and failure risks. Therefore, it is necessary to monitor the real-time faults of the controller and then calculate the remaining service life of the controller to ensure the reliability, safety and production efficiency of the gas cabinet in the automated production system. Among them, the stable monitoring of current in electrical performance monitoring is an important part of fault monitoring. The current is collected in real time through sensors or external monitoring systems, and the collected current is compared with the preset threshold to monitor the controller. However, after a period of use, the controller is affected by performance degradation, which makes the current range of the controller shift upward or downward compared to the preset threshold, resulting in misjudgment when judging whether the controller is abnormal through the preset fixed threshold. Summary of the invention

[0005] The present application provides a real-time monitoring system for faults of component units of a gas cabinet to solve the problem that the existing controller is affected by performance degradation after being used for a period of time, causing the collected current range to be offset upward or downward compared to the preset threshold, resulting in misjudgment when judging whether the controller is abnormal through the preset fixed threshold.

[0006] A gas cabinet component unit fault real-time monitoring system of the present application adopts the following technical solution:

[0007] An embodiment of the present application provides a real-time monitoring system for faults of gas cabinet components, the system comprising the following modules:

[0008] Monitoring data acquisition module: used to collect the controller monitoring current to obtain the historical current data of the current acquisition date and the current data of the current acquisition date at the monitoring time;

[0009] Abnormal data acquisition module: used to obtain the standard prediction value of the current data on the current collection date at the monitoring time according to the historical current data on the current collection date at the monitoring time; obtain the bias prediction value of the current data on the current collection date at the monitoring time according to the number of repetitions of the current data in the historical current data on the current collection date at the monitoring time; obtain the predicted abnormality degree of the current data on the current collection date at the monitoring time according to the standard prediction value and the bias prediction value of the current data on the current collection date at the monitoring time; obtain the trend abnormality degree of the current data on the current collection date at the monitoring time according to the trend change of the current data on the current collection date at the monitoring time and the current data at the adjacent sampling times; obtain the final abnormality degree of the current data on the current collection date at the monitoring time according to the predicted abnormality degree and trend abnormality degree of the current data on the current collection date at the monitoring time;

[0010] Fault monitoring module: used to perform fault monitoring of the controller on the current collection date at the monitoring time according to the final abnormality degree of the current data on the current collection date at the monitoring time.

[0011] Furthermore, the method of obtaining the standard predicted value of the current data at the monitoring time on the current collection date includes:

[0012] The monitoring time of the current collection date is recorded as Time, the current collection date Standard forecast value at time The calculation method is:

[0013]

[0014] in, Indicates the number of historical current data collected on the current date. Tiandi Current data at the moment, represents the number of days between historical current data collected on the current date, and D represents the sum of the number of days between historical current data collected on the current date. Indicates the total number of days of historical current data collected on the current date.

[0015] Further, the step of obtaining the bias prediction value of the current data at the monitoring time on the current collection date includes:

[0016] The current data sequence of each day in the historical current data of the current collection date is recorded as one-dimensional current data, and the current data of all dimensions are mapped to a two-dimensional discrete space, and a number of current data points are obtained in the two-dimensional discrete space. The number of current data points contained in each current data after mapping is recorded as the number of repetitions of each current data; the horizontal axis of the two-dimensional discrete space is the sampling time of the day, and the vertical axis is the current data at each sampling time. The unit length of the horizontal axis is a sampling interval, and the unit length of the vertical axis is the sampling unit of the ammeter;

[0017] The current data at the start sampling time of the controller's first use date is used as the seed point of regional growth, and the current data points that meet the regional growth rule in the two-dimensional discrete space are merged into the seed point area. When there are no current data points that meet the regional growth rule in the two-dimensional discrete space, the area where the seed point grows is recorded as the main area of ​​all historical current data on the current acquisition date;

[0018] The bias prediction value of the current data on the current collection date at the monitoring time is obtained based on the main areas of all historical current data on the current collection date.

[0019] Furthermore, the specific method of obtaining the current data points that satisfy the regional growth rule includes:

[0020] Seed points are used to perform regional growth on current data points in a two-dimensional discrete space. During the regional growth process, if all current data points grown from the seed points are in the first region formed by the seed points, the current data point with the maximum or minimum value of each horizontal axis or each vertical axis in the first region is selected as a current data point a, the unit length of the horizontal axis and the unit length of the vertical axis centered on the current data point a are recorded as the growth range of the current data point a, and all current data points in the growth range of the current data point a that are not in the first region are recorded as current data points that meet the regional growth rule.

[0021] Further, the step of obtaining the bias prediction value of the current data at the monitoring time on the current collection date includes:

[0022] The monitoring time of the current collection date is recorded as At this moment, the main area of ​​the historical current data of the current collection date is in the first The total number of repetitions of the current data at time is recorded as The main area for obtaining historical current data on the current acquisition date is in The maximum current data at the moment and current data minimum , the maximum value of the current data and current data minimum The current data with the largest difference from the seed point is recorded as the current acquisition date in Offset current data at time , will Offset current data at time The number of repetitions is recorded as ;

[0023] The current collection date is Biased prediction value at time The calculation method is:

[0024]

[0025] in, In the main area indicating the historical current data of the current collection date, The vertical axis range of the time except the offset current data Current data, V means the current collection date is The average number of repetitions of each current data in the vertical axis range of the main area at time, In the main area indicating the historical current data of the current collection date, The vertical axis range of time, Indicates that the current collection date is The vertical axis range of the main area at time excluding the offset current data Outside The number of repetitions of the current data.

[0026] Furthermore, the method for obtaining the vertical axis range includes:

[0027] The main area of ​​the historical current data of the current collection date is in the The maximum current data at the moment and current data minimum The interval formed The main area of ​​the historical current data recorded as the current collection date is in the The vertical axis range of time.

[0028] Furthermore, obtaining the predicted abnormality degree of the current data at the monitoring time on the current collection date includes:

[0029] The monitoring time of the current collection date is recorded as At this moment, the current collection date is The current data at the time is recorded as , the current collection date is The predicted abnormality at the moment The calculation method is:

[0030]

[0031] in, Indicates that the current collection date is The standard forecast value at time, Indicates that the current collection date is The biased prediction value at the moment; represents the absolute value function, express function.

[0032] Furthermore, the step of obtaining the trend abnormality degree of the current data at the monitoring time on the current collection date includes:

[0033] The monitoring time of the current collection date is recorded as At this moment, the current collection date is The degree of abnormality of the trend of the current data at the moment The calculation method is:

[0034]

[0035] in, Indicates the number of days from the current collection date. Time to The slope of the current data change at that moment, Indicates the number of days from the current collection date. Time to The changing slope of the current data at each moment; represents the absolute value function, express function.

[0036] Furthermore, obtaining the final abnormality degree of the current data at the monitoring time on the current collection date includes:

[0037] The monitoring time of the current collection date is recorded as At this moment, the current collection date is The final abnormality level of the current data at the moment The calculation method is:

[0038]

[0039] in, Indicates that the current collection date is The predicted abnormality of the current data at the moment, Indicates that the current collection date is The degree of abnormality of the trend of the current data at the moment.

[0040] Further, the fault monitoring of the controller on the current collection date at the monitoring time according to the final abnormality degree of the current data on the current collection date at the monitoring time includes:

[0041] Preset abnormality threshold , if the current collection date is The final abnormality level of the current data at the moment satisfy When the controller does not operate and calculates the The final abnormality of the moment ; If the current collection date is The final abnormality level of the current data at the moment satisfy When the controller does not operate and calculates the The final abnormality of the moment ; and so on, until there is a natural number , so that the current collection date is The final abnormality level of the current data at the moment satisfy , the current collection date is in The current data at that moment is recorded as abnormal data and the controller is stopped.

[0042] The beneficial effects of the technical solution of the present application are as follows: the present application collects the historical current data of the controller and the current data of the current collection date, uses the historical current data of the current collection date to obtain the standard predicted value of the current data at the monitoring time of the current collection date, and uses the initial working current data of the controller as the main reference to judge the difference between the monitoring data and the standard data at the current sampling time; uses the historical current data of the current collection date to obtain the biased predicted value of the current data at the monitoring time of the current collection date based on the number of repetitions of each current data, thereby solving the problem of current data offset caused by aging of the controller during use; obtains the predicted abnormality degree of the current data at the monitoring time of the current collection date according to the standard predicted value and the biased predicted value, obtains the trend abnormality degree of the current data at the monitoring time of the current collection date according to the change trend of the current data within the current collection date, thereby solving the fault caused by the sudden change of the current data of the controller; the trend abnormality degree is combined with the predicted abnormality degree to obtain the final abnormality degree of the current collection date at the monitoring time, which is used for real-time fault monitoring of the controller, thereby solving the problem that the current range of the controller changes with the working time and the preset fixed threshold is no longer applicable, thereby improving the accuracy of online fault monitoring of the gas cabinet controller. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 This is a structural block diagram of a real-time monitoring system for faults of gas cabinet components in this application;

[0045] Figure 2 This is a schematic diagram of current data points that satisfy the region growing rule in a two-dimensional discrete space of this application. DETAILED DESCRIPTION

[0046] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a real-time monitoring system for faults of gas cabinet components proposed in the present application in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in an embodiment may be combined in any suitable form.

[0047] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0048] The specific scheme of a real-time monitoring system for faults of gas cabinet components provided by the present application is described in detail below with reference to the accompanying drawings.

[0049] See also Figure 1 , which shows a structural block diagram of a gas cabinet component unit fault real-time monitoring system provided by an embodiment of the present application, the system includes the following modules:

[0050] Monitoring data acquisition module 101: The acquisition controller monitors the current to obtain the historical current data of the current acquisition date and the current data of the current acquisition date at the monitoring time.

[0051] It should be noted that this embodiment builds an abnormality detection model for judging controller failure by monitoring the current of the controller, so it is necessary to collect the current data of the controller first; this embodiment installs an ammeter at the power input interface of the controller of the production system to obtain all current readings from the start sampling time to the end sampling time in the production process, and the current readings recorded at every sampling interval are recorded as the current data of a sampling time. The sampling interval of this embodiment is selected as one minute for description; the current data of all sampling times in a historical day constitute the current data sequence of the historical day, and the current data of all dates from the first use date of the controller to the day before the current sampling date constitute the historical current data of the current collection date.

[0052] Furthermore, the acquisition controller records all current data from the start sampling moment to the current sampling moment on the current acquisition date as the current data on the current acquisition date, records the moment before the current sampling moment as the monitoring moment, and records the current data at the moment before the current sampling moment as the current data on the current acquisition date at the monitoring moment.

[0053] At this point, all historical current data on the current collection date and current data on the current collection date at the monitoring time have been acquired.

[0054] The abnormal data acquisition module 102 is used to acquire the precise abnormal degree of the current data on the current collection date at the monitoring time according to the current data on the current collection date and the historical current data on the current collection date.

[0055] (1) Based on the historical current data of the current collection date at the monitoring time, a standard prediction value of the current data of the current collection date at the monitoring time is obtained; based on the number of repetitions of the current data in the historical current data of the current collection date at the monitoring time, a bias prediction value of the current data of the current collection date at the monitoring time is obtained; based on the standard prediction value and bias prediction value of the current data of the current collection date at the monitoring time, a predicted abnormality degree of the current data of the current collection date at the monitoring time is obtained.

[0056] It should be noted that during the production process of the gas cabinet, the controller has certain similarities in its operations at the same sampling time every day. The numerical mean of the historical current data of the current collection date at the monitoring time is used as the predicted result of the monitoring time of the current collection date. The greater the difference between the current data at the monitoring time and the predicted result at the monitoring time, the greater the difference between the current data at the monitoring time and the predicted result at the monitoring time, which indicates that the actual sampling value at the monitoring time may be abnormal.

[0057] It should be further explained that during long-term use, the controller may have aging or loss of components, which may cause an increase in the resistance inside the controller, causing the working current of the controller after loss to deviate from the preset working current. Therefore, when judging whether the controller is faulty based on a preset threshold constructed according to the preset working current, the offset current may easily exceed the threshold and cause a normal controller to be misjudged as a faulty controller. Therefore, this embodiment uses the interval between the collection date of historical current data and the current collection date as a weight to obtain the standard predicted value of the current data on the current collection date at the monitoring time, which is used to construct an abnormality detection model.

[0058] Specifically, in this embodiment, the monitoring time of the current collection date is recorded as the first At that time, the current collection date Standard forecast value at time The calculation method is:

[0059]

[0060] in, Indicates the current collection date. The standard forecast value at time, Indicates the number of historical current data collected on the current date. Tiandi Current data at the moment, Indicates the number of days between historical current data collected on the current date. The number of days between historical current data collected on the current date refers to the number of complete days included in the historical current data collected on the current date. It represents the total number of days of the historical current data of the current collection date, and D represents the sum of the interval days corresponding to all collection dates in the historical current data of the current collection date. Indicates that the number of days from the historical current data to the current collection date is used as the first day of the calculation of the current collection date. The weight of the standard prediction value at the moment, the farther the distance, the greater the weight, so that the result of the standard prediction value is closer to the preset working current of the controller.

[0061] It should be further explained that when calculating the standard prediction value, the current data close to the initial operation of the controller occupies a larger weight. Therefore, if the current of the controller is offset, the standard prediction value at all sampling moments of each day will have a larger deviation compared to the actual current data of the current collection date. Therefore, this embodiment obtains the biased prediction value of the current collection date at the monitoring time based on the distribution of all historical current data of the current collection date at the monitoring time.

[0062] Specifically, this embodiment records the current data sequence of each day in the historical current data of the current collection date as current data of one dimension, maps the current data of all dimensions into a two-dimensional discrete space, and obtains a number of current data points in the two-dimensional discrete space. The horizontal axis of the two-dimensional discrete space is the sampling time of the day, and the vertical axis is the current data at each sampling time. The unit length of the horizontal axis is a sampling interval, and the unit length of the vertical axis is the sampling unit of the ammeter. It should be noted that at a sampling time, there may be multiple days of the same current data in all the historical current data of the current collection date, so there may be multiple repeated current data points for one current data in the two-dimensional discrete space, and the number of current data points of each current data after mapping is recorded as the number of repetitions of each current data.

[0063] Furthermore, the current data point at the start sampling time of the controller's first use date is used as a seed point for regional growth, and the current data points that meet the regional growth rules in the two-dimensional discrete space are merged into the seed point area; the current data points that meet the regional growth rules are obtained by using the seed points to perform regional growth on the current data points in the two-dimensional discrete space. During the regional growth process, if all the current data points grown by the seed points are in the first area formed by the seed points, the current data point with the maximum or minimum value of each horizontal axis or each vertical axis in the first area is selected as a current data point a, and the unit length of the horizontal axis and the unit length of the vertical axis centered on the current data point a are recorded as the growth range of the current data point a, and all current data points in the growth range of the current data point a that are not in the first area are recorded as current data points that meet the regional growth rules; the current data points that meet the regional growth rules are as follows: Figure 2 As shown in the figure, the black circle is the current data point, the gray circle is the current data point a, and the white circle is the current data point b. In the figure, the current data point a is a current data point that is at a maximum or minimum value on the horizontal axis or the vertical axis in the first region, and the current data point b is a current data point that is within the growth range of the current data point a and is not in the first region. Then the current data point b is a current data point that satisfies the regional growth rule.

[0064] Furthermore, the monitoring time of the current collection date is recorded as At this moment, the main area of ​​the historical current data of the current collection date is in the first The total number of repetitions of the current data at time is recorded as ; The main area for obtaining historical current data for the current acquisition date is in The maximum current data at the moment and current data minimum , the maximum value of the current data and current data minimum The interval formed The main area of ​​the historical current data recorded as the current collection date is in the Time vertical axis range , calculate the total number of repetitions With vertical axis range The ratio V, V is used to characterize the mean value of the number of repetitions of each current data in the vertical axis range.

[0065] Secondly, the maximum value of the current data and current data minimum The current data with the largest difference from the seed point is recorded as the current acquisition date in Offset current data at time ; Offset current data at time The number of repetitions is recorded as .

[0066] Further, the current acquisition date is obtained according to the number of repetitions of the current data at the monitoring time. Biased prediction value at time The calculation method is:

[0067]

[0068] in, Indicates that the current collection date is The biased prediction value at the moment, V represents the current collection date at The average number of repetitions of each current data in the vertical axis range of the main area at time, In the main area indicating the historical current data of the current collection date, The total number of repetitions of the current data point at time, Indicates The offset current data at time, Indicates The number of repetitions of the offset current data at time instant; In the main area indicating the historical current data of the current collection date, The vertical axis range of time, In the main area indicating the historical current data of the current collection date, The vertical axis range of the time except the offset current data Current data, In the main area indicating the historical current data of the current collection date, The vertical axis range of the time except the offset current data The number of repetitions of the current data.

[0069] Indicates that the current collection date is The average number of repetitions of each current data in the vertical axis range of the main area at time In the mean and Offset current data at time Number of repetitions The proportion of the sum, the larger the value, the higher the current collection date. At the time, the proportion of offset current data in the vertical axis range of the main area is small, that is, the current data of the controller has not been offset during use, so the current collection date is The biased prediction value at the moment depends on the current collection date in Current data of main areas at the moment; The larger the value, the more repetitions of the offset current data, which means that the current data is collected on the current date. When the current offset occurs in the vertical axis range of the main area at the time, the current acquisition date is The biased prediction value at the moment depends mainly on the current collection date in Offset current data at time instant.

[0070] At this point, the standard prediction value and bias prediction value of the current collection date at the monitoring time are obtained.

[0071] Furthermore, an abnormality detection model at the monitoring time is constructed based on the standard prediction value and bias prediction value of the current acquisition date at the monitoring time, and the current data at the monitoring time is input into the abnormality detection model to obtain the predicted abnormality degree at the monitoring time. The predicted abnormality at the moment The calculation method is:

[0072]

[0073] in, Indicates that the current collection date is The predicted abnormality at the moment, Indicates that the current collection date is Current data at the moment, Indicates that the current collection date is The standard forecast value at time, Indicates that the current collection date is The biased prediction value at the moment; represents the absolute value function, express function. Indicates that according to the current collection date The standard prediction value and bias prediction value at the time are used to construct the evaluation value of the anomaly detection model at the monitoring time. The current collection date is If the current data at the time is significantly different from the evaluation value, it means that the current collection date is The current data at a given moment may be abnormal.

[0074] At this point, the predicted abnormality level at the monitoring time on the current collection date is obtained.

[0075] (2) The trend abnormality degree of the current data on the current collection date at the monitoring time is obtained based on the trend change of the current data on the current collection date at the monitoring time and the current data at the adjacent sampling times. The final abnormality degree of the current data on the current collection date at the monitoring time is obtained based on the predicted abnormality degree and trend abnormality degree of the current data on the current collection date at the monitoring time.

[0076] It should be noted that the fault of the current data of the controller is not only related to the abnormality when the historical current data is greatly different, but also related to the change of the adjacent current data on the current collection date. If the controller fails and the components inside the controller are burned to form a short circuit, the current data collected at this time will change suddenly compared with the current data collected at the adjacent sampling time, and the greater the degree of sudden change at the sampling time, the greater the possibility of the controller failure. Therefore, this embodiment obtains the trend abnormality degree of the current data at the monitoring time of the current collection date by comparing the change difference between the current data at the monitoring time of the current collection date and the current data at the adjacent sampling time.

[0077] Specifically, the current collection date is The degree of abnormality of the trend of the current data at the moment The calculation method is:

[0078]

[0079] in, Indicates that the current collection date is The degree of abnormality of the trend of the current data at the moment, Indicates the number of days from the current collection date. Time to The slope of the current data change at that moment, Indicates the number of days from the current collection date. Time to The changing slope of the current data at each moment; represents the absolute value function, express function. The larger the value, the more likely the current collection date is. The greater the change in the current data at the moment, the more likely it is that the controller is abnormal. The current collection date is The greater the abnormality of the trend of the current data at the moment; The larger the value, the more likely the current collection date is. The current data at the moment may be an extreme point, and the more likely it is caused by controller abnormality, the more likely it is that the current collection date is The greater the abnormality of the trend of the current data at the moment.

[0080] At this point, the degree of trend abnormality of the current data at the monitoring time on the current collection date is obtained.

[0081] Further, the final abnormality degree of the current data on the current collection date at the monitoring time is obtained according to the predicted abnormality degree and trend abnormality degree of the current data on the current collection date at the monitoring time. The final abnormality level of the current data at the moment The calculation method is:

[0082]

[0083] in, Indicates that the current collection date is The final abnormality level of the current data at the moment, Indicates that the current collection date is The predicted abnormality of the current data at the moment, Indicates that the current collection date is The degree of abnormality of the trend of the current data at the moment.

[0084] At this point, the final abnormality level of the current data at the monitoring time on the current collection date is obtained.

[0085] Fault monitoring module 103: used to perform fault monitoring of the controller on the current collection date at the monitoring time according to the final abnormality degree of the current data on the current collection date at the monitoring time.

[0086] Preset abnormality threshold , this embodiment uses If the current collection date is The final abnormality level of the current data at the moment satisfy When the current collection date is The current data at the moment does not have any abnormality, the controller does not operate and calculates the The final abnormality of the moment ; If the current collection date is The final abnormality level of the current data at the moment satisfy When the current collection date is The current data at the moment does not have any abnormality, the controller does not operate and calculates the The final abnormality of the moment ; and so on, until there is a natural number , so that the current collection date is The final abnormality level of the current data at the moment satisfy , indicating that the current collection date is The current data at the moment is abnormal, and the current collection date is The current data at that moment is recorded as abnormal data and the controller is stopped, and the abnormal state of the controller is uploaded to the company's production management system.

[0087] It should be noted that the above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

[0088] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A real-time monitoring system for failure of gas cabinet components, characterized in that: The system includes the following modules: Monitoring data acquisition module: used to collect the controller monitoring current to obtain the historical current data of the current acquisition date and the current data of the current acquisition date at the monitoring time; Abnormal data acquisition module: used to obtain the standard prediction value of the current data on the current collection date at the monitoring time according to the historical current data on the current collection date at the monitoring time; obtain the bias prediction value of the current data on the current collection date at the monitoring time according to the number of repetitions of the current data in the historical current data on the current collection date at the monitoring time; obtain the predicted abnormality degree of the current data on the current collection date at the monitoring time according to the standard prediction value and the bias prediction value of the current data on the current collection date at the monitoring time; obtain the trend abnormality degree of the current data on the current collection date at the monitoring time according to the trend change of the current data on the current collection date at the monitoring time and the current data at the adjacent sampling times; obtain the final abnormality degree of the current data on the current collection date at the monitoring time according to the predicted abnormality degree and trend abnormality degree of the current data on the current collection date at the monitoring time; Wherein, obtaining the bias prediction value of the current data at the monitoring time on the current collection date includes: recording the current data sequence of each day in the historical current data on the current collection date as current data of one dimension, mapping the current data of all dimensions into a two-dimensional discrete space, obtaining a number of current data points in the two-dimensional discrete space, and recording the number of current data points contained in each current data after mapping as the number of repetitions of each current data; the horizontal axis of the two-dimensional discrete space is the sampling time of the day, the vertical axis is the current data at each sampling time, the unit length of the horizontal axis is a sampling interval, and the unit length of the vertical axis is the sampling unit of the ammeter; The current data at the start sampling time of the controller's first use date is used as the seed point of regional growth, and the current data points that meet the regional growth rule in the two-dimensional discrete space are merged into the seed point area. When there are no current data points that meet the regional growth rule in the two-dimensional discrete space, the area where the seed point grows is recorded as the main area of ​​all historical current data on the current acquisition date; Obtain the bias prediction value of the current data at the monitoring time on the current collection date according to the main areas of all historical current data on the current collection date; Wherein, obtaining the bias prediction value of the current data at the monitoring time on the current collection date further includes: recording the monitoring time of the current collection date as the t-th time, recording the total number of repetitions of the current data at the t-th time in the main area of ​​the historical current data of the current collection date as ρ″ 0,t , obtain the maximum current data I of the main area of ​​the historical current data of the current collection date at the tth moment max,t and current data minimum I min,t , the maximum current data I max,t and current data minimum I min,t The current data with the largest difference from the current data of the seed point is recorded as the offset current data I′ at the tth moment of the current acquisition date. 0,t , the offset current data I′ at time t 0,t The number of repetitions is denoted as ρ′ 0,t ; The biased prediction value I of the current collection date at time t 0,t,2 The calculation method is: Among them, I m, t represents the mth current data in the vertical axis range at the tth time in the main area of ​​the historical current data of the current acquisition date, excluding the offset current data; V represents the mean number of repetitions of each current data in the vertical axis range of the main area at the tth time on the current acquisition date; M 0,t The vertical axis range at time t in the main area of ​​the historical current data of the current acquisition date, ρ 0,t,m Indicates the current data I′ in the vertical axis range of the main area at the time t of the current acquisition date minus the offset current data 0,t The number of repetitions of the mth current data outside; Fault monitoring module: used to perform fault monitoring of the controller on the current collection date at the monitoring time according to the final abnormality degree of the current data on the current collection date at the monitoring time.

2. According to claim 1, a real-time monitoring system for failure of gas cabinet components, characterized in that: The method of obtaining the standard prediction value of the current data at the monitoring time on the current acquisition date includes: The monitoring time of the current collection date is recorded as the tth time, and the standard prediction value I at the tth time of the current collection date is 0,t,1 The calculation method is: Among them, I d,t It represents the current data at the tth time on the dth day in the historical current data of the current collection date, d represents the number of days between the historical current data of the current collection date, D represents the sum of the number of days between all the historical current data of the current collection date, and D0 represents the total number of days of the historical current data of the current collection date.

3. A real-time monitoring system for failure of gas cabinet components according to claim 1, characterized in that: The specific method for obtaining the current data points that meet the regional growth rule includes: Seed points are used to perform regional growth on current data points in a two-dimensional discrete space. During the regional growth process, if all current data points grown from the seed points are in the first region formed by the seed points, the current data point with the maximum or minimum value of each horizontal axis or each vertical axis in the first region is selected as a current data point a, the unit length of the horizontal axis and the unit length of the vertical axis centered on the current data point a are recorded as the growth range of the current data point a, and all current data points in the growth range of the current data point a that are not in the first region are recorded as current data points that meet the regional growth rule.

4. According to claim 1, a real-time monitoring system for failure of gas cabinet components, characterized in that: Methods for obtaining the vertical axis range include: The maximum current data I of the main area of ​​the historical current data of the current collection date at the t moment max,t and current data minimum I min,t The interval [I min,t , I max,t ] is recorded as the vertical axis range at the tth time in the main area of ​​the historical current data on the current collection date.

5. A gas cabinet component unit fault real-time monitoring system according to claim 1, characterized in that: The obtaining of the predicted abnormality degree of the current data at the monitoring time on the current collection date includes: The monitoring time of the current collection date is recorded as the tth time, and the current data of the current collection date at the tth time is recorded as I 0,t , the predicted abnormality level N of the current acquisition date at time t 0,t The calculation method is: Among them, I 0,t,1 represents the standard prediction value of the current collection date at time t, I 0,t,2 represents the biased prediction value of the current collection date at time t; | | represents the absolute value function, and sigmoid{} represents the sigmoid function.

6. A gas cabinet component unit fault real-time monitoring system according to claim 1, characterized in that: The degree of abnormal trend of the current data at the monitoring time on the current acquisition date is obtained by: The monitoring time of the current collection date is recorded as the tth time, and the trend abnormality degree A of the current data of the current collection date at the tth time is 0,t The calculation method is: A 0,t =sigmoid{|k 0,t+1 ×(k 0,t-1 -k 0,t+1 )|} Among them, k 0,t-1 Indicates the change slope of the current data from the t-1th moment to the tth moment in the current acquisition date, k 0,t+1 It represents the slope of the current data from time t to time t+1 in the current acquisition date; | | represents the absolute value function, and si.gmoid{} represents the sigmoid function.

7. A gas cabinet component unit fault real-time monitoring system according to claim 1, characterized in that: The obtaining of the final abnormality degree of the current data at the monitoring time on the current collection date includes: The monitoring time of the current collection date is recorded as the tth time, and the final abnormality degree P of the current data of the current collection date at the tth time is 0,t The calculation method is: P 0,t =N 0,t ×A 0,t Among them, N 0,t Indicates the predicted abnormality of the current data at time t on the current collection date, A 0,t Indicates the degree of abnormal trend of the current data at time t on the current collection date.

8. A gas cabinet component unit fault real-time monitoring system according to claim 1, characterized in that: The fault monitoring of the controller at the monitoring time on the current collection date according to the final abnormality degree of the current data at the monitoring time on the current collection date includes: The abnormality threshold B is preset. If the final abnormality level P of the current data at the tth moment on the current collection date is 0,t Satisfy P 0,t When <B, the controller does not operate and calculates the final abnormality level P at time t+1 0,t+1 ; If the final abnormality level P of the current data at the time t+1 of the current acquisition date 0,t+1 Satisfy P 0,t+1 When <B, the controller does not operate and calculates the final abnormality level P at time t+2 0,t+2 ; and so on, until there is a natural number c, so that the final abnormality degree P of the current data at the t+c moment of the current collection date is 0,t+c Satisfy P 0,t+c ≥B, the current data at the time t+c of the current collection date is recorded as abnormal data and the controller is stopped.