Intelligent operation and maintenance control method and related equipment based on edge computing

Abstract

The present invention belongs to the field of edge computing technology, and specifically relates to an intelligent operation and maintenance control method based on edge computing and related equipment. The method includes: collecting data of IoT products through an IoT gateway, the data including parameter data; performing feature operations on the parameter data according to a preset feature library and feature trigger conditions, and calculating a feature data set; performing warning calculations on the feature data set according to a preset fault warning model and warning trigger conditions, and obtaining warning result data; inputting the warning result data into a preset fault diagnosis model to obtain a fault diagnosis result, and the fault diagnosis result includes a faulty component and a fault cause. The present invention can realize the rapid import and operation of data collection for IoT products and user diagnostic analysis capabilities, and supports user-defined analysis functions. It supports intelligent operation and maintenance functions such as fault prediction, fault diagnosis, and health assessment of equipment or systems.

Description

Technical Field

[0001] The present invention belongs to the field of edge computing technology, and specifically relates to an intelligent operation and maintenance control method based on edge computing and related equipment. Background Art

[0002] At present, the data collection and calculation processing of equipment is aimed at solving the equipment operation control, and its purpose is to ensure that the equipment operates according to the set logic and requirements. There is a lack of effective solutions for the impact of other equipment and environment on the equipment, and the prevention, diagnosis, maintenance and repair of equipment failures.

[0003] Existing devices such as data acquisition gateways mainly solve the problem of data acquisition of equipment, and do not have the ability to analyze large amounts of data and perform complex calculations. Existing edge computing gateways mainly provide CPU and memory hardware resources such as data calculation and data storage, and lack the ability for users to quickly support intelligent maintenance of equipment such as fault warning, fault diagnosis, and performance evaluation based on their own knowledge and capabilities.

[0004] Some existing edge computing devices can use machine learning methods to analyze and warn of some anomalies and failures that have already occurred. However, such devices have the defect of being based on hindsight learning and can only analyze and predict data that has already occurred, but cannot predict and evaluate potential problems that have not yet occurred.

[0005] Some of the existing central platforms can realize the analysis and calculation of collected data, and carry out problems such as fault prediction, fault diagnosis, and health assessment of some equipment related to equipment maintenance. However, due to the high frequency and large amount of equipment data, all data are transmitted to the central platform through the network for calculation and processing, which has network bandwidth limitations and processing delays. It cannot solve the problem of not missing any equipment data features and timely calculation and processing. Summary of the invention

[0006] The present invention aims to provide an intelligent operation and maintenance control method and related equipment based on edge computing, in order to solve the technical problems that existing equipment data cannot be analyzed in a timely manner and equipment failures cannot be predicted.

[0007] An intelligent operation and maintenance control method based on edge computing, comprising:

[0008] Collecting data of IoT products through an IoT gateway, wherein the data includes parameter data;

[0009] According to the preset feature library and feature trigger conditions, feature operations are performed on the parameter data to calculate and obtain a feature data set;

[0010] According to the preset fault warning model and warning triggering conditions, the warning calculation is performed on the characteristic data set to obtain warning result data;

[0011] The early warning result data is input into a preset fault diagnosis model to obtain a fault diagnosis result, which includes a faulty component and a fault cause.

[0012] As a preferred solution, the data also includes fault warning data, and the fault warning data is directly input into the fault diagnosis model to obtain a fault diagnosis result.

[0013] As a preferred solution, before performing feature operations on the parameter data according to the preset feature library and feature trigger conditions to obtain the feature data set, the method further includes:

[0014] The parameter data is cleaned by using a user-defined preset cleaning rule to perform at least one of data anomaly elimination, missing value interpolation, smoothing and noise reduction, or filtering.

[0015] As a preferred solution, it also includes marking the cleaned parameter data with data labels:

[0016] Obtain a preset tag start calculation expression, calculate whether the parameter data is established, if not, continue to obtain the next tag start calculation expression, until all tag start calculation expressions are calculated, if established, proceed to the next step;

[0017] Get the tag end calculation expression corresponding to the current tag start calculation expression, calculate whether the parameter data is true, if not, continue to get the next tag start calculation expression until all tag start calculation expressions are calculated, if true, mark the parameter data as the data tag corresponding to the current tag start calculation expression.

[0018] As a preferred solution, when performing warning calculation on the feature data set according to the preset fault warning model and warning trigger conditions, a historical feature data set is also obtained, and the feature data set and the historical feature data set are used together to perform warning calculation through the fault warning model and warning trigger conditions to obtain warning result data.

[0019] As a preferred solution, when performing feature operations on the parameter data or performing early warning calculations on the feature data set, independent operation scheduling engines are used for scheduling. When scheduling, the operation scheduling engine dynamically adjusts and allocates computing tasks based on trigger conditions and available task pool resources to achieve efficient and dynamic scheduling of operations.

[0020] As a preferred scheme, after the fault diagnosis model obtains the early warning result data or fault alarm data, it extracts the corresponding preset fault judgment calculation model and operation input parameters, and the operation input parameters include at least one of the collected data, the feature data set, the historical feature data set or the historical fault record data, and calculates the operation input parameters to obtain the faulty component and the cause of the fault.

[0021] An intelligent operation and maintenance control system based on edge computing, comprising:

[0022] A data collection module, used to collect data of IoT products through an IoT gateway, wherein the data includes parameter data;

[0023] A feature extraction module is used to perform feature operations on the parameter data according to a preset feature library and feature trigger conditions, and calculate a feature data set;

[0024] An early warning calculation module is used to perform early warning calculation on the characteristic data set according to a preset fault early warning model and early warning triggering conditions to obtain early warning result data;

[0025] The fault diagnosis module is used to input the warning result data into a preset fault diagnosis model to obtain a fault diagnosis result, wherein the fault diagnosis result includes a faulty component and a fault cause.

[0026] A computer device includes a memory and a processor, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the steps of the above-mentioned intelligent operation and maintenance control method based on edge computing.

[0027] A storage medium storing computer-readable instructions, which, when executed by one or more processors, causes the one or more processors to execute the steps of the above-mentioned intelligent operation and maintenance control method based on edge computing.

[0028] The positive progress of the present invention is that the present invention adopts the intelligent operation and maintenance control method based on edge computing and related equipment, which can realize the data collection of IoT products, the rapid import and operation of user diagnosis and analysis capabilities, and support user-defined analysis functions. It supports intelligent operation and maintenance functions such as fault prediction, fault diagnosis and health assessment of equipment or systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A schematic diagram of a method flow of the present invention;

[0030] Figure 2 A calculation process diagram of the data tag of the present invention;

[0031] Figure 3A schematic diagram of a method flow of a feature operation scheduling engine of the present invention;

[0032] Figure 4 A schematic diagram of a method flow of the early warning operation scheduling engine of the present invention;

[0033] Figure 5 The present invention is a flow chart of a method for fault diagnosis. DETAILED DESCRIPTION

[0034] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further described below with reference to specific illustrations.

[0035] Reference Figure 1 , an intelligent operation and maintenance control method based on edge computing, comprising the following steps:

[0036] S001, data collection: collect data of IoT products through the IoT gateway, including parameter data.

[0037] The present invention can connect to the Internet of Things gateway through the Internet of Things protocol to collect data such as various parameters, fault alarm information, etc. of Internet of Things products, such as Internet of Things devices or Internet of Things systems. The Internet of Things protocol includes MQTT, CoAP or DDS, etc. The collected data can be divided into parameter data and fault alarm data according to different data types.

[0038] S002, data cleaning: using user-defined preset cleaning rules to perform at least one of the following cleaning processes on the parameter data: data anomaly elimination, missing value interpolation, smoothing and noise reduction, or filtering.

[0039] This step is an optional step that is not necessary. The data cleaning of the present invention supports the user to define a cleaning algorithm or rule to perform cleaning processing on the collected parameter data.

[0040] S003, data label: mark the cleaned parameter data with data labels.

[0041] This step is optional and non-essential. The data labeling algorithm in this step is implemented by a user-defined labeling algorithm or rule, which performs calculations by associating acquisition parameters, faults or feature data, and sets data labels according to the calculation results, making data labels more accurate and easy to operate.

[0042] The data labeling algorithm of this step consists of the label start calculation expression, the label end calculation expression, and the label definition.

[0043] In some embodiments, specifically including:

[0044] Get a preset tag to start calculating the expression, and calculate whether the parameter data is valid. If not, continue to get the next tag to start calculating the expression until all the tag start calculating expressions are calculated. If they are valid, proceed to the next step.

[0045] Get the tag end calculation expression corresponding to the current tag start calculation expression, calculate whether the parameter data is true, if not, continue to get the next tag start calculation expression until all tag start calculation expressions are calculated, if true, mark the parameter data as the data tag corresponding to the current tag start calculation expression.

[0046] For example, data label A is "motor start-up phase", the label for judging motor start-up is pre-set to start calculating expression S1, and the label for judging motor start-up end calculating expression S2 is pre-set to end calculating expression S2. According to the above steps, refer to Figure 2 , calculate the parameter data to determine whether it is true. If it is true, assign data label A "motor starting stage" to the parameter data collected during this period.

[0047] The data labels of this step can be used in subsequent feature extraction steps and fault warning steps. According to the data labels of the parameter data, the corresponding feature calculation rules or warning calculation rules can be directly obtained, so that the feature calculation or warning calculation of the parameter data is more targeted, more accurate and more efficient.

[0048] S004, feature extraction: according to the preset feature library and feature trigger conditions, feature operations are performed on the parameter data to calculate and obtain a feature data set.

[0049] The feature extraction in this step is based on a preset feature library. According to the preset feature trigger conditions, multiple feature data are calculated and the multiple feature data form a feature data set. The feature library in this step stores multiple feature calculation rules and corresponding feature trigger conditions. The feature trigger condition is a calculation expression. If the expression calculation is valid, it is considered that the feature trigger condition is met.

[0050] In some embodiments, when performing feature operations, a feature operation scheduling engine is used for scheduling. The feature operation scheduling engine is controlled according to feature trigger conditions and resources. When scheduling, the feature operation scheduling engine dynamically adjusts and allocates computing tasks according to feature trigger conditions and available resources in the task pool to achieve efficient and dynamic scheduling of operations. Figure 3After the feature calculation starts, the feature operation scheduling engine calculates whether the resources are sufficient. If not, it needs to wait for the feature calculation. If so, it determines whether the feature trigger condition is satisfied. If so, it performs calculation according to the feature calculation rules to obtain a feature data, which includes feature parameters and feature values. If not, it performs feature calculation for the next parameter data.

[0051] S005, fault warning: According to the preset fault warning model and warning triggering conditions, the feature data set is calculated to obtain the warning result data.

[0052] The fault warning of this step is based on a preset fault warning model. According to the preset warning trigger conditions, multiple fault warning data are calculated, and the multiple fault warning data form warning result data. The fault warning model of this step stores multiple warning calculation rules and corresponding warning trigger conditions. The warning trigger condition is a calculation expression. If the expression calculation is established, it is considered that the warning trigger condition is met.

[0053] In some embodiments, when performing early warning calculation on a feature data set according to a preset fault early warning model and early warning triggering conditions, a historical feature data set is also obtained, and the feature data set and the historical feature data set are used together to perform early warning calculation through the fault early warning model and early warning triggering conditions to obtain early warning result data. That is, this step is to perform a fault early warning calculation on the total feature data set.

[0054] In some embodiments, when performing fault warning calculations, a warning calculation scheduling engine is used for scheduling. The warning calculation scheduling engine is controlled according to the warning trigger conditions and resources. When scheduling, the warning calculation scheduling engine dynamically adjusts and allocates computing tasks according to the warning trigger conditions and the availability of task pool resources to achieve efficient and dynamic scheduling of calculations. Figure 4 After the fault warning calculation starts, the warning operation scheduling engine calculates whether the resources are sufficient. If not, it needs to wait for the fault warning calculation. If it is satisfied, it determines whether the warning trigger condition is satisfied. If it is satisfied, it performs calculation according to the warning calculation rule to obtain a fault warning data, which includes the warning component and the warning code. If not, the fault warning calculation of the next feature data is performed.

[0055] S006, fault diagnosis: input the warning result data into a preset fault diagnosis model to obtain a fault diagnosis result, which includes a faulty component and a fault cause.

[0056] The fault diagnosis in this step is based on a preset fault diagnosis model. According to the fault diagnosis model, a plurality of fault diagnosis data can be obtained, and the plurality of fault diagnosis data form the final fault diagnosis result.

[0057] The fault diagnosis result includes multiple faulty components and their corresponding fault causes. The fault diagnosis result can be transmitted to relevant maintenance personnel for review by display or remote transmission.

[0058] In some embodiments, after the fault diagnosis model obtains the warning result data or the fault alarm data, it extracts the corresponding preset fault judgment calculation model and operation input parameters. The operation input parameters include at least one of the collected data, feature data sets, historical feature data sets or historical fault record data. The operation input parameters are calculated to obtain the faulty component and the cause of the fault.

[0059] Specifically, refer to Figure 5 When the fault diagnosis starts, the fault judgment calculation model in the fault diagnosis model is called according to the fault warning results or fault alarm data. The operation input parameters are extracted from a large number of data according to the fault judgment calculation model to perform fault judgment calculation. If the diagnosis is completed, the fault component and the fault cause code are finally obtained. If the diagnosis is not completed, the fault diagnosis steps are continued.

[0060] The present invention also provides an intelligent operation and maintenance management system based on edge computing, comprising:

[0061] The data collection module is used to collect data of IoT products through the IoT gateway, including parameter data;

[0062] The feature extraction module is used to perform feature operations on parameter data according to a preset feature library and feature trigger conditions to calculate a feature data set;

[0063] The early warning calculation module is used to perform early warning calculation on the feature data set according to the preset fault early warning model and early warning triggering conditions to obtain early warning result data;

[0064] The fault diagnosis module is used to input the warning result data into a preset fault diagnosis model to obtain the fault diagnosis result, which includes the fault component and the cause of the fault.

[0065] In one embodiment, the present invention also provides a computer device including a memory and a processor, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor executes the steps of the intelligent operation and maintenance control method based on edge computing in the above-mentioned embodiments.

[0066] In one embodiment, the present invention also provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps in the intelligent operation and maintenance control method based on edge computing in the above-mentioned embodiments.

[0067] A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium can include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc. Among them, the storage medium can be a non-volatile storage medium.

[0068] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An intelligent operation and maintenance control method based on edge computing, characterized in that: include: Collecting data of IoT products through an IoT gateway, wherein the data includes parameter data; According to the preset feature library and feature trigger conditions, feature operations are performed on the parameter data to calculate and obtain a feature data set; According to the preset fault warning model and warning triggering conditions, the warning calculation is performed on the characteristic data set to obtain warning result data; Inputting the warning result data into a preset fault diagnosis model to obtain a fault diagnosis result, wherein the fault diagnosis result includes a faulty component and a fault cause; Before performing feature operations on the parameter data according to the preset feature library and feature trigger conditions to obtain the feature data set, the method further includes: Using a user-defined preset cleaning rule to perform at least one of data anomaly elimination, missing value interpolation, smoothing and noise reduction, or filtering on the parameter data; It also includes marking data labels for the cleaned parameter data: Obtain a preset tag start calculation expression, calculate whether the parameter data is established, if not, continue to obtain the next tag start calculation expression, until all tag start calculation expressions are calculated, if established, proceed to the next step; Obtain the tag end calculation expression corresponding to the current tag start calculation expression, calculate whether the parameter data is true, if not, continue to obtain the next tag start calculation expression until all tag start calculation expressions are calculated, if true, mark the parameter data as the data tag corresponding to the current tag start calculation expression; The performing feature operation on the parameter data according to the preset feature library and feature trigger conditions to calculate the feature data set also includes: The feature library stores a plurality of feature calculation rules and corresponding feature trigger conditions, wherein the feature trigger condition is a calculation expression. If the expression is calculated to be true, it is considered that the feature trigger condition is satisfied; After the feature calculation starts, the feature operation scheduling engine calculates whether the resources are sufficient. If not, it needs to wait for the feature calculation. If so, it determines whether the feature trigger condition is met. If so, it performs the calculation according to the feature calculation rule to obtain a feature data. The feature data includes feature parameters and feature values. If not, the feature calculation of the next parameter data is performed.

2. The intelligent operation and maintenance control method based on edge computing according to claim 1, characterized in that: The data also includes fault warning data, and the fault warning data is directly input into the fault diagnosis model to obtain a fault diagnosis result.

3. The intelligent operation and maintenance control method based on edge computing according to claim 1, characterized in that: According to the preset fault warning model and warning triggering conditions, when the warning calculation is performed on the feature data set, the historical feature data set is also obtained, and the feature data set and the historical feature data set are used together to perform warning calculation through the fault warning model and warning triggering conditions to obtain warning result data.

4. The intelligent operation and maintenance control method based on edge computing according to claim 1, characterized in that: When performing feature operations on the parameter data or performing early warning calculations on the feature data set, each independent operation scheduling engine is used for scheduling. When scheduling, the operation scheduling engine dynamically adjusts and allocates computing tasks according to trigger conditions and available resources in the task pool to achieve efficient and dynamic scheduling of operations.

5. The intelligent operation and maintenance control method based on edge computing according to claim 1 or 2, characterized in that: After the fault diagnosis model obtains the early warning result data or the fault alarm data, it extracts the corresponding preset fault judgment calculation model and operation input parameters, and the operation input parameters include at least one of the collected data, the feature data set, the historical feature data set or the historical fault record data. The operation input parameters are calculated to obtain the faulty component and the cause of the fault.

6. An intelligent operation and maintenance control system based on edge computing, characterized in that: include: A data collection module, used to collect data of IoT products through an IoT gateway, wherein the data includes parameter data; A feature extraction module is used to use a user-defined preset cleaning rule to perform at least one of data anomaly elimination, missing value interpolation, smoothing noise reduction or filtering on the parameter data; and is also used to mark data labels for the cleaned parameter data: obtain a preset label start calculation expression, calculate whether the parameter data is established, if not, continue to obtain the next label start calculation expression, until all label start calculation expressions are calculated, if established, proceed to the next step; obtain the label end calculation expression corresponding to the current label start calculation expression, calculate whether the parameter data is established, if not, continue to obtain the next label start calculation expression, until all label start calculation expressions are calculated, if established, mark the parameter data as the data label corresponding to the current label start calculation expression; It is also used to perform feature operations on the parameter data according to a preset feature library and feature trigger conditions to calculate a feature data set; the feature library stores multiple feature calculation rules and corresponding feature trigger conditions, and the feature trigger condition is a calculation expression. If the expression calculation is established, it is considered that the feature trigger condition is met; the feature extraction module is also used to determine whether the feature operation scheduling engine calculation resources are sufficient after the feature calculation starts. If not, it is necessary to wait for the feature calculation. If so, it is determined whether the feature trigger condition is satisfied. If so, the calculation is performed according to the feature calculation rule to obtain a feature data, and the feature data includes feature parameters and feature values. If not, the feature calculation of the next parameter data is performed; An early warning calculation module is used to perform early warning calculation on the characteristic data set according to a preset fault early warning model and early warning triggering conditions to obtain early warning result data; The fault diagnosis module is used to input the warning result data into a preset fault diagnosis model to obtain a fault diagnosis result, wherein the fault diagnosis result includes a faulty component and a fault cause.

7. A computer device, characterized in that: It includes a memory and a processor, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the steps of the intelligent operation and maintenance control method based on edge computing as claimed in any one of claims 1 to 5.

8. A storage medium storing computer-readable instructions, characterized in that: When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the intelligent operation and maintenance control method based on edge computing as described in any one of claims 1 to 5.

Images