CN116662060B

CN116662060B - Data processing method and system of sensor signal acquisition processing system

Info

Publication number: CN116662060B
Application number: CN202310946997.9A
Authority: CN
Inventors: 周平江; 姚伟; 巩鑫; 冯斌
Original assignee: Shenzhen Chuangyin Technology Co ltd
Current assignee: Shenzhen Chuangyin Technology Co ltd
Priority date: 2023-07-31
Filing date: 2023-07-31
Publication date: 2024-02-06
Anticipated expiration: 2043-07-31
Also published as: CN116662060A

Abstract

The application provides a data processing method and a system of a sensor signal acquisition processing system, wherein a first system fault location network is subjected to function layer function updating according to training abnormal working data to obtain a second system fault location network, so that when target abnormal working data is acquired, the second system fault location network can be subjected to fault location on the target abnormal working data according to the second system fault location network after network knowledge learning, the first system fault location network can determine the number of system fault location units according to the number of control activities of sensor control activities of the training abnormal working data, and therefore, the function layer function updating can be directly carried out in the architecture of the first system fault location network according to the abnormal working data of different sensor control activities, the expansibility and pertinence in the network knowledge learning process are improved, the accuracy in the fault location process is improved, and the sensor signal acquisition processing efficiency is ensured.

Description

Data processing method and system of sensor signal acquisition processing system

Technical Field

The application relates to the technical field of sensors and data processing, in particular to a data processing method and system of a sensor signal acquisition processing system.

Background

With the increasing demand for electricity, power distribution systems are moving toward high capacity and ultra-high voltage, and corresponding electrical devices are moving toward this direction. On the other hand, with the rapid development of sensor technology, digital communication technology, computer software and hardware technology and power electronics technology, favorable conditions are provided for the development of the power distribution system in an intelligent direction. Various types of power transformers are used as one of electric equipment, are eyes of a power system, and play important roles in the power system, such as electric energy metering and relay protection. In a specific application scene, signal data acquired by various branch equipment power transformers are required to be transmitted to a central control system in real time through a network, and real-time monitoring and predictive maintenance are carried out on the working environment, the product condition and the equipment running condition.

However, in the control process of the sensor signal acquisition and processing, various possible faults may occur, and for the operation and maintenance end, it is necessary to timely check the existing faults and perform configuration of the operation and maintenance scheme so as to timely recover the normal acquisition and processing of the sensor signal. In the related art, the knowledge learning of the network model can be generally performed through an artificial intelligent algorithm to perform fault analysis on abnormal working data of the sensor signal acquisition processing system, however, the expansibility and pertinence of the related art in the network knowledge learning process are poor, so that the accuracy in the fault positioning process cannot be well ensured in the practical application process.

Disclosure of Invention

In order to at least overcome the above-mentioned shortcomings in the prior art, an object of an embodiment of the present application is to provide a data processing method and system of a sensor signal acquisition processing system.

According to an aspect of the embodiments of the present application, there is provided a data processing method of a sensor signal acquisition processing system, including:

acquiring target abnormal working data to be subjected to fault positioning, wherein the target abnormal working data are abnormal working data of system scheduling software;

invoking a second system fault location network, wherein the second system fault location network comprises a plurality of system fault location units, the system fault location units consist of a plurality of interdependent system fault location units, and the number of the system fault location units is determined by the number of control activities of the first system fault location network based on the sensor control activities of the abnormal working data of the system scheduling software in a network knowledge learning process;

loading the target abnormal working data into the second system fault location network, performing fault location on the loaded abnormal working data by a preceding system fault location unit aiming at each system fault location unit dependent on the plurality of system fault location units, performing fault location on the abnormal working data generated by heuristic search on the fault location data according to the preceding system fault location unit by a following system fault location unit, and generating target fault location data according to a plurality of fault location data of the plurality of system fault location units, wherein each sensor control activity of the plurality of fault location data represents the confidence degree of each candidate fault tag in the plurality of candidate fault tags.

In a possible implementation manner of the first aspect, the performing, by the preceding system fault location unit, fault location on the loaded abnormal working data for each of the plurality of system fault location units, and performing, by the following system fault location unit, fault location on the abnormal working data generated by performing heuristic search according to the fault location data of the preceding system fault location unit, includes:

aiming at each phase-dependent system fault locating unit in the plurality of system fault locating units, carrying out fault locating on the target abnormal working data according to a first system fault locating unit to generate first fault locating data, wherein the first system fault locating unit is a forward system fault locating unit in each phase-dependent system fault locating unit;

performing fault location on heuristic fault location data generated by performing heuristic search on the first fault location data according to a second system fault location unit to generate second fault location data, wherein the second system fault location unit is a backward system fault location unit in the system fault location units depending on each phase, and the heuristic fault location data is part of control data of the target abnormal working data.

In a possible implementation manner of the first aspect, the plurality of system fault location units respectively correspond to a control tag of the sensor control activity;

the heuristic fault location data comprise sensor control activities of a first control tag associated with the first fault location data, wherein the first control tag is a control tag corresponding to the first system fault location unit;

when the fault location loss value of the fault location data of each system fault location unit is obtained in the network knowledge learning flow of the first system fault location network, the influence coefficient of the control tag corresponding to the system fault location unit is larger than that of other control tags.

In a possible implementation manner of the first aspect, the network knowledge learning procedure of the second system fault location network includes:

acquiring a plurality of pieces of training abnormal working data, wherein the plurality of pieces of training abnormal working data are abnormal working data of system scheduling software;

according to the first system fault location network, the plurality of training abnormal working data are loaded into the first system fault location network, according to the plurality of training abnormal working data, function layer function updating is carried out on the first system fault location network, a second system fault location network is generated, the first system fault location network is used for taking the control activity number of active sensor control activities of the plurality of training abnormal working data as the number of system fault location units in the first system fault location network, and different system fault location units are used for carrying out fault location on different system working nodes of the abnormal working data.

In a possible implementation manner of the first aspect, each training abnormal working data defines a priori fault location marking data, and the fault location marking data characterizes target fault location data of the training abnormal working data;

before said taking the number of active sensor control activities of the plurality of training abnormal operation data as the number of system fault location units in the first system fault location network, the method further comprises:

and analyzing the fault positioning labeling data of the plurality of training abnormal working data to generate the control activity quantity of the sensor control activities of the plurality of training abnormal working data.

In a possible implementation manner of the first aspect, the method further includes:

acquiring sensor control linkage data among target system working nodes with node dependency relations in a plurality of target system working nodes corresponding to the plurality of training abnormal working data in a network knowledge learning process of the first system fault positioning network, wherein the target system working nodes are system working nodes in which sensor control activities of target control labels in the plurality of training abnormal working data are located;

When each system fault locating unit in the plurality of system fault locating units carries out heuristic search on abnormal working data, the heuristic search is carried out on the abnormal working data based on the relation between sensor control linkage data and set heuristic search paths between target system working nodes with node dependency relations in the plurality of target system working nodes.

acquiring airspace data of a plurality of target system working nodes corresponding to the plurality of training abnormal working data in a network knowledge learning process of the first system fault positioning network;

when each system fault locating unit in the plurality of system fault locating units carries out heuristic search on abnormal working data, the first system fault locating unit carries out heuristic search on target abnormal working data based on the relation between sensor control linkage data of a first target system working node and a second target system working node and a set heuristic search path and time-space domain data of the first target system working node or the second target system working node, wherein the first target system working node is a system working node where a sensor control activity of a first control tag corresponding to the first system fault locating unit is located, and the second target system working node is a system working node where a sensor control activity of a second control tag corresponding to the second system fault locating unit is located.

In a possible implementation manner of the first aspect, the first system fault location network and the second system fault location network each include a plurality of modal data location networks, where the plurality of modal data location networks are respectively used for acquiring modal collection data of abnormal working data based on different data modalities and performing fault location on the abnormal working data;

the number of the plurality of system fault locating units is determined by a first system fault locating network based on the number of control activities of sensor control activities of abnormal working data of the system scheduling software in a network knowledge learning process, and the method comprises the following steps:

in a network knowledge learning process, the first system fault location network takes the control activity quantity of active sensor control activities of abnormal working data of the system scheduling software as the quantity of system fault location units in each modal data location network;

the system fault locating unit for each phase dependence in the plurality of system fault locating units, the preceding system fault locating unit performing fault locating on the loaded abnormal working data, the following system fault locating unit performing fault locating on the abnormal working data generated by heuristic searching according to the fault locating data of the preceding system fault locating unit, generating target fault locating data according to the plurality of fault locating data of the plurality of system fault locating units, including:

Acquiring at least one mode acquisition data of target abnormal working data by the plurality of mode data positioning networks according to corresponding data modes respectively, performing fault positioning on the acquired data by a plurality of system fault positioning units in each mode data positioning network, and generating target fault positioning data according to the fault positioning data of the plurality of mode data positioning networks;

the fault location data of each of the plurality of modal data location networks includes modal fault location data of a plurality of system fault location units;

the generating the target fault location data according to the fault location data of the plurality of modal data location networks includes:

and carrying out feature fusion on the modal fault locating data of the system fault locating units of the modal data locating networks according to the influence coefficients corresponding to the modal data locating networks and the influence coefficients corresponding to the system fault locating units in the modal data locating networks to generate target fault locating data, wherein the influence coefficients corresponding to the modal data locating networks and the influence coefficients corresponding to the system fault locating units in the modal data locating networks are determined based on a regression modeling algorithm.

determining a target fault label corresponding to each sensor control activity based on the target fault positioning data;

acquiring a fault restoration knowledge graph related to the target fault label;

performing software repair processing on the sensor signal acquisition processing system based on the fault repair knowledge graph and the control behavior of the corresponding sensor control activity;

the step of performing software repair processing on the sensor signal acquisition processing system based on the fault repair knowledge graph and the control behavior of the corresponding sensor control activity comprises the following steps:

aiming at each to-be-matched fault restoration knowledge point in the fault restoration knowledge map, acquiring a tree node sequence and/or a tree node execution example sequence corresponding to a knowledge tree of the to-be-matched fault restoration knowledge point, wherein the tree node sequence is loaded in a sensor control service of the sensor signal acquisition and processing system, the tree node execution example sequence is loaded in a scheduling software service of the sensor signal acquisition and processing system, and a restoration scheduling relationship exists between the tree node and the tree node execution example;

Obtaining matching characteristics of control behaviors of the fault restoration knowledge points to be matched and corresponding sensor control activities according to the tree node sequence, the tree node execution instance and the restoration scheduling relationship or according to the tree node sequence;

determining matching characteristics of the tree node sequence according to the matching characteristics of the at least one to-be-matched fault restoration knowledge point and the fault restoration characteristic vector of the at least one to-be-matched fault restoration knowledge point, so as to determine a tree node execution example sequence to be restored according to the matching characteristics of the tree node sequence and the restoration scheduling relation, and restore the scheduling software service;

determining a tree node to be repaired or updating the tree node according to the matching characteristic aiming at the tree node sequence;

and repairing the scheduling software service according to the tree node to be repaired and/or the updating tree node.

The sensor control service comprises a sensor local tree node, a sensor cooperation tree node and a sensor operation logic node, wherein the sensor operation logic node comprises at least one sensor local tree node, and the sensor local tree node comprises at least one sensor cooperation tree node;

Obtaining the matching characteristic of the control behavior of the fault restoration knowledge point to be matched and the corresponding sensor control activity according to the tree node sequence, the tree node execution instance and the restoration scheduling relationship or according to the tree node sequence comprises at least one of the following:

aiming at the sensor local tree node, determining the sensor local tree node corresponding to each software function instance of the scheduling software service in the sensor control service according to the repair scheduling relationship, and determining a comparison value of the number of the sensor local tree node compared with the number of all software function instances corresponding to the control behaviors of the corresponding sensor control activities in the scheduling software service;

determining sensor cooperation tree nodes of all software function examples corresponding to the sensor control service in the scheduling software service according to the repair scheduling relationship, and determining a comparison value of the number of cooperation fields of the sensor cooperation tree nodes compared with the number of cooperation fields of all sensor cooperation tree nodes corresponding to the control behavior of the corresponding sensor control activity in the scheduling software service;

determining sensor local tree nodes corresponding to all software function examples of the scheduling software service in the sensor control service according to the repair scheduling relationship, determining sensor cooperation tree nodes corresponding to the sensor local tree nodes in the scheduling software service, and determining comparison values of the number of the sensor cooperation tree nodes compared with the number of all the sensor cooperation tree nodes corresponding to the control behaviors of the corresponding sensor control activities in the scheduling software service;

Determining a sensor cooperation tree node associated with a sensor local tree node in the sensor control service, determining a sensor cooperation tree node associated with a control element in the sensor control service in the past time domain compared with a first comparison value of all sensor cooperation tree nodes in the scheduling software service in the time domain, comparing with a second comparison value of all sensor cooperation tree nodes corresponding to control actions of corresponding sensor control activities in the scheduling software service, and determining a change value of the first comparison value compared with the second comparison value.

According to one aspect of the embodiments of the present application, there is provided a data processing system of a sensor signal acquisition processing system, the data processing system of the sensor signal acquisition processing system including a processor and a machine-readable storage medium having stored therein machine-executable instructions loaded and executed by the processor to implement a data processing method of the sensor signal acquisition processing system in any one of the foregoing possible implementations.

According to an aspect of embodiments of the present application, there is provided a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The computer instructions are read from a computer-readable storage medium by a processor of a computer device, and executed by the processor, cause the computer device to perform the methods provided in the various alternative implementations of the above aspects.

In the technical schemes provided by some embodiments of the present application, the function layer function update is performed on the first system fault location network according to the training abnormal working data to obtain the second system fault location network, so that when the target abnormal working data is obtained, the second system fault location network after learning the network knowledge can perform fault location on the target abnormal working data, the first system fault location network can determine the number of system fault location units according to the number of control activities of the sensor control activities of the training abnormal working data, and therefore, the function layer function update can be directly performed on the abnormal working data of different sensor control activities in the architecture of the first system fault location network, thereby improving the expansibility and pertinence in the network knowledge learning process, and further improving the accuracy in the fault location process.

Drawings

For a clearer description of the technical solutions of the embodiments of the present application, reference will be made to the accompanying drawings, which are needed to be activated, for the sake of simplicity, and it should be understood that the following drawings only illustrate some embodiments of the present application, and therefore should not be considered as limiting the scope, and that it is possible for a person skilled in the art to extract other relevant drawings in combination with these drawings without the inventive effort.

Fig. 1 is a schematic flow chart of a data processing method of a sensor signal acquisition processing system according to an embodiment of the present application;

FIG. 2 is a second flow chart of a data processing method of the sensor signal acquisition processing system according to the embodiment of the present application;

fig. 3 is a schematic block diagram of a data processing system of a sensor signal acquisition processing system for implementing the data processing method of the sensor signal acquisition processing system according to an embodiment of the present application.

Detailed Description

The following description is presented to enable one of ordinary skill in the art to make and use the application and is provided in the context of a particular application and its requirements. It will be apparent to those having ordinary skill in the art that various changes can be made to the disclosed embodiments and that the general principles defined herein may be applied to other embodiments and applications without departing from the principles and scope of the present application. Thus, the present application is not limited to the embodiments described, but is to be accorded the widest scope consistent with the claims.

Fig. 1 is a flowchart of a data processing method of a sensor signal acquisition processing system according to an embodiment of the present application, and the data processing method of the sensor signal acquisition processing system is described in detail below.

Step 101, acquiring a plurality of pieces of training abnormal working data collected in a sensor signal acquisition control process of a sensor signal acquisition processing system, wherein each piece of training abnormal working data defines fault positioning marking data in a priori, and the fault positioning marking data represents target fault positioning data of the training abnormal working data.

The plurality of training abnormal working data are abnormal working data of system scheduling software, the function layer function of the first system fault location network can be updated according to the plurality of training abnormal working data, a second system fault location network is generated, and the generated second system fault location network can perform fault location on the abnormal working data of the system scheduling software. The abnormal working data of the system scheduling software can be understood as abnormal working data generated by the sensor signal acquisition processing system in the control process of the scheduling sensor, and for example, the abnormal working data can include, but is not limited to, environment sensing control operation of the sensor signal acquisition processing system for each measuring object, cooperative control operation between a controlled sensor and other sensors, signal conversion operation for measuring signals, signal amplification operation, signal filtering operation, signal calibration operation, sensor instruction issuing operation, sensor state switch operation, sensor data uploading control operation and the like.

In some alternative embodiments, the plurality of training abnormal working data may be collected in advance, and the plurality of training abnormal working data may be obtained when the network knowledge learning of the system fault location network is required. Each piece of training abnormal working data can be defined and configured with fault location marking data representing target fault location data, and the target fault location data refers to actual fault location data of the training abnormal working data, so that whether the fault location of the training abnormal working data is accurate or not can be judged in a network knowledge learning flow, whether function layer function updating is needed to be continuously carried out or not can be judged, and the target fault location data can be obtained or is close to the target fault location data when a second system fault location network obtained through training carries out fault location on the training abnormal working data.

Step 102, loading the plurality of training abnormal working data into the first system fault location network according to the first system fault location network.

After the plurality of training abnormal working data are acquired, the first system fault locating network can be updated with function layer functions according to the plurality of training abnormal working data and a second system fault locating network is generated, so that the acquired target abnormal working data can be accurately located according to the second system fault locating network.

In some alternative embodiments, the functional layer function weight information of the first system fault location network is initialization weight information, the plurality of training abnormal working data can be used as network learning data and network test data, and the functional layer function update is performed on the first system fault location network, that is, the functional layer function weight information of the first system fault location network is iteratively updated and tested through the training abnormal working data, so that the functional layer function weight information after multiple rounds of iterative update and test can generate more accurate fault location data when the target abnormal working data is subjected to fault location.

For example, the plurality of training abnormal working data can be loaded into the first system fault location network, each training abnormal working data can be subjected to fault location by the first system fault location network, and the fault location performance of the first system fault location network is determined according to the fault location data of the first system fault location network and the fault location marking data of the training abnormal working data, namely, the target fault location data of the training abnormal working data, so that the function layer function weight information of the first system fault location network can be updated through iteration, the fault location performance of the first system fault location network is improved, and the fault location accuracy of the second system fault location network which is generated subsequently is improved.

Step 103, the first system fault location network determines a number of system fault location units based on a control activity number of the sensor control activities of the plurality of training abnormal operation data.

The system fault locating units are used for locating faults of different system working nodes of abnormal working data. In some alternative embodiments, the training abnormal working data is sequentially subjected to fault location by a plurality of system fault location units, and for two system fault location units which are dependent on each other, after the previous system fault location unit performs fault location on the abnormal working data, heuristic search can be performed on the original abnormal working data, and the heuristic searched abnormal working data is loaded to the subsequent system fault location unit, so that the subsequent system fault location unit can continue to perform fault location according to the fault location data of the previous system fault location unit.

In some alternative embodiments, this step 103 may be: the first system fault location network uses the number of control activities of the active sensor control activities of the plurality of training abnormal operation data as the number of system fault location units in the first system fault location network. The control tags for sensor control activity may include passive control tags and active control tags. The corresponding sensor control activities are passive control tag sensor control activity and active sensor control activity, respectively. In other words, the sensor control activity in which the control tag is a passive control tag is the passive control tag sensor control activity, and the sensor control activity in which the control tag is an active control tag is the active sensor control activity. The first system fault location network determines a plurality of system fault location units that respectively correspond to one control tag of the sensor control activity, i.e., one control tag of each system fault location unit that is concerned with the fault location sensor control activity.

In some alternative embodiments, the first system fault location network may directly obtain the number of control activities of the active sensor control activities of the training abnormal working data, or may obtain the number of control activities of the sensor control activities, and subtract one from the number of control activities of the sensor control activities to generate the number of control activities of the active sensor control activities.

In some alternative embodiments, the target fault location data in the fault location labeling data of the plurality of training abnormal working data, that is, the control tag representing the control activity of each sensor of the training abnormal working data, then the control tag data of the control activity of the sensor of the plurality of training abnormal working data may be obtained according to the fault location labeling data of the plurality of training abnormal working data. Therefore, the first system fault location network before the step 103 may further analyze the fault location labeling data of the plurality of training abnormal working data to generate the control activity number of the sensor control activities of the plurality of training abnormal working data, thereby executing the step 103, and determining the number of the system fault location units according to the control fault location labeling data number. For example, the first system fault location network may count the number of control activities of the sensor control activities in the target fault location data in the fault location labeling data, or may determine only the number of control activities of the active sensor control activities.

Step 104, a plurality of system fault locating units in the first system fault locating network sequentially locate faults of each piece of training abnormal working data, and generate fault locating data of each system fault locating unit on the training abnormal working data.

After the number of the system fault locating units is determined, the first system fault locating network can sequentially conduct fault locating on the training abnormal working data according to the plurality of system fault locating units to generate fault locating data. For example, for each system fault location unit, the loaded abnormal working data may be feature coded according to a feature coding function to generate an abnormal working extraction vector. And then, extracting vectors according to the abnormal work, and carrying out fault location decision on each sensor control activity in the abnormal work data to generate fault location data.

In some alternative embodiments, the system fault location unit may include at least one fault location subunit, with different fault location subunits having different depths. In the network knowledge learning process of the first system fault location network, for each system fault location unit, the first system fault location network may further obtain, as the system fault location unit, a fault location subunit corresponding to the number of abnormal working data based on the number of abnormal working data of the plurality of training abnormal working data. Therefore, a proper fault location subunit can be selected for carrying out function layer function updating according to the abnormal working data quantity of the training abnormal working data.

After the first system fault location network determines the basic network layer (fault location subunit) of each system fault location unit, the first system fault location network may sequentially perform fault location on the training abnormal working data according to the plurality of system fault location units. For example, for two system fault locating units that depend on each other, the first system fault locating network may perform fault locating on the target abnormal working data according to the first system fault locating unit to generate first fault locating data, where the first system fault locating unit is a forward system fault locating unit of the two system fault locating units that depend on each other. The first system fault location network performs fault location on heuristic fault location data generated by performing heuristic search on the first fault location data according to a second system fault location unit to generate second fault location data, wherein the second system fault location unit is a backward system fault location unit in the two dependent system fault location units, and the heuristic fault location data is part of control data of the target abnormal working data.

Wherein each sensor control activity of the first fault location data and the second fault location data characterizes the abnormal working data as a confidence level of each candidate fault tag of the plurality of candidate fault tags. The target abnormal working data is the abnormal working data loaded to the first system fault locating unit, and the heuristic fault locating data is the abnormal working data generated by the first system fault locating unit performing heuristic search on the target abnormal working data according to the first fault locating data. The heuristic fault location data includes sensor control activity of a first control tag associated with the first fault location data, the first control tag being a control tag corresponding to the first system fault location unit.

In some alternative embodiments, the control tag corresponding to the second system fault location unit is a second control tag. The system operation node where the sensor control activity of the second control tag is located is in the system operation node where the sensor control activity of the first control tag is located. The sensor control activity of the first control tag may be focused on fault location first, and then the sensor control activity of the second control tag may be more refined for fault location within the system operational node where the sensor control activity of the first control tag is located. Each system fault locating unit can carry out system fault locating decision on the sensor control activity, and the confidence coefficient of the sensor control activity as each candidate fault label is determined.

In other words, the first system fault locating unit performs fault locating on the target abnormal working data, determines the confidence coefficient of each sensor control activity as each control label, and primarily determines the control label of each sensor control activity according to the confidence coefficient, and the first system fault locating unit focuses on the first control label for fault locating, so that heuristic fault locating data of a system working node where the sensor control activity including the first control label is located is loaded to the second system fault locating unit, and the second system fault locating unit can continue fault locating on the heuristic fault locating data and focuses on the second control label. If the first system fault locating unit is the first system fault locating unit in the plurality of system fault locating units of the first system fault locating network, the target abnormal working data is the loaded training abnormal working data.

In some alternative embodiments, in step 103, in the network knowledge learning process of the first system fault location network, the first system fault location network may further obtain sensor control linkage data between target system working nodes having node dependencies among the plurality of target system working nodes corresponding to the plurality of training abnormal working data, where the target system working nodes are system working nodes where sensor control activities of target control labels in the plurality of training abnormal working data are located. Thus, in the step 104, when each of the plurality of system fault locating units performs heuristic search on the abnormal working data, the system fault locating unit may perform heuristic search on the abnormal working data based on a relationship between sensor control linkage data between target system working nodes having node dependency relationships among the plurality of target system working nodes and a set heuristic search path.

In some alternative embodiments, in step 103, in the network knowledge learning process of the first system fault location network, the first system fault location network may further obtain time-space domain data of a plurality of target system working nodes corresponding to the plurality of training abnormal working data. In step 104, when each of the plurality of system fault locating units performs heuristic search on the abnormal working data, the system fault locating unit may perform heuristic search on the abnormal working data based on a relationship between sensor control linkage data and a set heuristic search path between target system working nodes having node dependency relationships among the plurality of target system working nodes and time-space domain data of the plurality of target system working nodes.

For example, the step 104 may be described by two system fault locating units depending on the above-mentioned system fault locating units: and for the first system fault positioning unit, the first system fault positioning unit performs heuristic search on the target abnormal working data based on the relation between sensor control linkage data of the first target system working node and sensor control linkage data of the second target system working node and a set heuristic search path and time-space domain data of the first target system working node or the second target system working node, wherein the first target system working node is a system working node where a sensor control activity of a first control tag corresponding to the first system fault positioning unit is located, and the second target system working node is a system working node where a sensor control activity of a second control tag corresponding to the second system fault positioning unit is located.

Step 105, the first system fault location network obtains a fault location loss value of each fault location data according to fault location marking data of a plurality of fault location data and the training abnormal working data.

In some alternative embodiments, when fault location loss values of fault location data of each system fault location unit are obtained in a network knowledge learning process of the first system fault location network, an influence coefficient of a control tag corresponding to the system fault location unit is greater than an influence coefficient of other control tags.

And step 106, updating the function layer function weight information of the first system fault location network according to the fault location loss values until the function layer function weight information is not changed any more, and generating a second system fault location network, wherein the function layer function weight information of the first system fault location network at least comprises unit function weight information of each system fault location unit.

After the second system fault location network is generated, when target abnormal working data to be subjected to fault location is received, the second system fault location network can be used for carrying out fault location on the target abnormal working data according to the second system fault location network and a plurality of system fault location units to generate target fault location data, wherein the target abnormal working data is also abnormal working data of system scheduling software.

In some alternative embodiments, the first system fault location network and the second system fault location network each include a plurality of modality data location networks for respectively acquiring modality acquisition data of the abnormal operation data and performing fault location on the abnormal operation data based on different data modalities.

According to the mode acquisition data of the abnormal working data obtained according to different data modes, the fault positioning data of different data modes are combined, and the accuracy of the second system fault positioning network in positioning the abnormal working data faults can be improved. The first system fault location network may use the number of control activities of the active sensor control activities of the plurality of training abnormal operation data as the number of system fault location units in each of the modal data location networks when determining the number of system fault location units. In other words, the number of system fault locating units in each modal data location network is the number of control activities of the active sensor control activities. In the above network knowledge learning process, for each modal data positioning network, the training abnormal working data may be sequentially subjected to fault positioning according to the plurality of system fault positioning units, so that the plurality of fault positioning data of the plurality of modal data positioning networks are combined to generate final abnormal working data. The fault location data of each of the plurality of modal data location networks includes modal fault location data of a plurality of system fault location units.

In the process of calling the second system fault location network to perform fault location on the target abnormal working data, the plurality of mode data location networks can respectively acquire at least one mode acquisition data of the target abnormal working data according to the corresponding data modes, the plurality of system fault location units in each mode data location network perform fault location on each mode acquisition data, and the target fault location data is generated according to the fault location data of the plurality of mode data location networks. The fault location data of each of the plurality of modal data location networks includes modal fault location data of a plurality of system fault location units. According to the fault location data of the multi-mode data location network, the process of generating the target fault location data is the same as the learning of the network knowledge, and can be as follows: and carrying out feature fusion on the modal fault locating data of the system fault locating units of the modal data locating networks according to the influence coefficients corresponding to the modal data locating networks and the influence coefficients corresponding to the system fault locating units in each modal data locating network to generate target fault locating data.

Based on the steps, the first system fault location network is updated with the function layer function according to the training abnormal work data to obtain the second system fault location network, so that when the target abnormal work data is received, the second system fault location network which is learned according to the network knowledge can conduct fault location on the target abnormal work data, wherein the first system fault location network can determine the number of system fault location units according to the number of control activities of sensor control activities of the training abnormal work data, expansibility and pertinence in the network knowledge learning process can be improved, and accuracy in the subsequent fault location process is further improved.

Further application embodiments are described below in connection with fig. 2, which may include the following steps:

step 201, obtaining target abnormal working data to be subjected to fault location.

Step 202, a second system fault location network is invoked.

Wherein the second system fault location network includes a plurality of system fault location units. The number of the plurality of system fault location units is determined by the first system fault location network when the function layer function update is performed on the first system fault location network. The system fault locating units are used for locating faults of different system working nodes of abnormal working data, and the system fault locating units can be used for locating faults of target abnormal working data in sequence, so that a multi-dimensional fault locating process is realized.

The second system fault location network may also include a plurality of mode data location networks, which are respectively used for acquiring mode acquisition data of the abnormal working data and performing fault location on the abnormal working data based on different data modes, similar to the above description in step 106. Thus, a plurality of system fault locating units are included in each modal data location network. The system fault locating units respectively correspond to one control tag of the sensor control activity, namely, the system fault locating units are respectively used for focusing on one control tag of the fault locating sensor control activity.

Step 203, loading the target abnormal working data into the second system fault location network, and performing fault location on the target abnormal working data by the second system fault location network according to the plurality of system fault location units to generate a plurality of fault location data.

If the second system fault location network includes a plurality of modal data location networks, the step 203 may be: loading the target abnormal working data into a second system fault location network, acquiring at least one mode acquisition data of the target abnormal working data by a plurality of mode data location networks in the second system fault location network according to corresponding data modes respectively, performing fault location on each mode acquisition data by a plurality of system fault location units in each mode data location network, and generating target fault location data according to the fault location data of the plurality of mode data location networks.

In the same manner as in step 104, each system fault location unit may perform fault location and heuristic search on the abnormal working data to generate fault location data, for example, for two system fault location units that are dependent (such as an association relationship with a fault triggering reason) in the plurality of system fault location units, the first fault location unit may perform fault location on the target abnormal working data according to the first system fault location unit, to generate first fault location data, where the first system fault location unit is a forward system fault location unit in the two system fault location units that are dependent. The computer equipment performs fault location on heuristic fault location data generated by performing heuristic search on the first fault location data according to a second system fault location unit to generate second fault location data, wherein the second system fault location unit is a backward system fault location unit in the two dependent system fault location units, and the heuristic fault location data is part of control data of the target abnormal working data. The heuristic fault location data comprises sensor control activities of a first control tag associated with the first fault location data, wherein the first control tag is a control tag corresponding to the first system fault location unit.

In step 204, the second system fault location network generates target fault location data according to the plurality of fault location data.

When a plurality of modal data localization networks are included, the fault localization data of each of the plurality of modal data localization networks includes modal fault localization data of a plurality of system fault localization units. Therefore, in the process of generating the target fault location data by the second system fault location network according to the fault location data of the plurality of modal data location networks, the feature fusion can be performed on the modal fault location data of the plurality of system fault location units of the plurality of modal data location networks according to the influence coefficients corresponding to the plurality of modal data location networks and the influence coefficients corresponding to each system fault location unit in each modal data location network, so as to generate the target fault location data. The influence coefficients corresponding to the plurality of modal data positioning networks and the influence coefficients corresponding to each system fault positioning unit in each modal data positioning network are determined based on a regression modeling algorithm.

Based on the steps, the first system fault location network is subjected to function layer function updating according to the training abnormal work data to obtain the second system fault location network, so that when the target abnormal work data is obtained, the second system fault location network after network knowledge learning can be used for performing fault location on the target abnormal work data, the first system fault location network can be used for determining the number of system fault location units according to the control activity number of the sensor control activities of the training abnormal work data, and therefore, the function layer function updating can be directly performed on the abnormal work data of different sensor control activities in the framework of the first system fault location network, the expansibility and pertinence in the network knowledge learning process are improved, and the accuracy in the fault location process is further improved.

On the basis of the above description, the embodiment of the application may further include the following steps.

Step 301, determining a target fault label corresponding to each sensor control activity based on the target fault positioning data. For example, candidate fault tags with a confidence greater than the set confidence may be used as target fault tags for the corresponding sensor control activities.

Step 302, obtaining a fault repair knowledge graph related to the target fault label.

For example, fault repair knowledge maps related to different fault labels may be preconfigured, and the fault repair knowledge maps may include a plurality of fault repair knowledge points to be matched, and a tree node sequence and/or a tree node execution instance sequence corresponding to a knowledge tree corresponding to each fault repair knowledge point to be matched.

And step 303, performing software repair processing on the sensor signal acquisition processing system based on the fault repair knowledge graph and the control behavior of the corresponding sensor control activity.

Step 303 may include:

Based on the steps, the tree node sequence and the tree node execution example sequence required by matching the to-be-matched fault repair knowledge points are determined according to the repair scheduling relationship between the tree node sequence of the sensor control service and the tree node execution example sequence of the scheduling software service, so that the matching characteristics of the to-be-matched fault repair knowledge points are generated. And then determining the matching characteristics of the tree node sequence according to the fault repair characteristic vectors of the fault repair knowledge points to be matched, and guiding the repair scheduling software service according to the matching characteristics.

For some exemplary design considerations, the sensor in the embodiments of the present application may be a non-contact self-calibration current-voltage integrated sensor, and may include, for example, an insulation box, a current sensor installed in the insulation box, a voltage sensor, a signal processing circuit, and a self-calibration combination metering unit. The current sensor is formed by uniformly winding a metering winding on an annular iron core, two ends of a wire are connected with a bidirectional TVS tube and a precise power resistor in parallel, the precise power resistor is used as a sampling resistor to convert a current signal into a voltage signal, and then the detected voltage signal US1 is sent to a self-calibration combination metering unit. The voltage sensor is characterized in that an annular capacitive screen welding lead is connected with an anode of an input end, a GDT tube, a bidirectional TVS tube and a capacitor are connected in parallel, the capacitor and a sampling resistor are connected in series to form a resistor-capacitor voltage divider, so that a voltage sampling signal is output, the sampling signal is subjected to impedance transformation through an operational amplifier follower, and the sampling signal is isolated and then is sent to a self-calibration combination metering unit. The self-calibration combination metering unit transmits voltage and current sampling signals to a metering chip, and the metering chip calculates the voltage, the current, the phase and the frequency and converts analog signals into digital signals. After the MCU reads the digital signal of the metering chip, the sampling signal is compensated, so that a standard value is output, and when the parameter is read by the intelligent electronic equipment, a message is output according to IEC62056-21 protocol.

In the current sensor, in order to prevent the interference of the high-current conductor, the current sensor coil is arranged in the shielding shell, and is poured by epoxy resin for insulation treatment, and the bidirectional TVS tube plays a role in overcurrent protection.

In the voltage sensor, the voltage output interface uses piezoresistor, self-recovery insurance, TVS tube, diode and protection circuit for overvoltage, overcurrent and reverse connection prevention.

In the self-calibration combination metering unit, the current sampling signal US1 and the voltage sampling signal US2 are transmitted to the metering chip, and the metering chip calculates the voltage, the current, the phase and the frequency, and the analog signal is converted into a digital signal. After the MCU reads the digital signal of the metering chip, the sampling signal is compensated, so that a standard value is output, and when the parameter is read by the display, a message can be output according to IEC62056-21 protocol.

Fig. 3 illustrates a hardware architecture diagram of a data processing system 100 of a sensor signal acquisition processing system for implementing the data processing method of the sensor signal acquisition processing system according to an embodiment of the present application, where, as shown in fig. 3, the data processing system 100 of the sensor signal acquisition processing system may include a processor 110, a machine-readable storage medium 120, a bus 130, and a communication unit 140.

In an alternative embodiment, the data processing system 100 of the sensor signal acquisition processing system may be a single server or a group of servers. The server farm may be centralized or distributed (e.g., the data processing system 100 of the sensor signal acquisition processing system may be a distributed system). In an alternative embodiment, the data processing system 100 of the sensor signal acquisition processing system may be local or remote. For example, the data processing system 100 of the sensor signal acquisition processing system may access information and/or data stored in the machine-readable storage medium 120 via a network. As another example, the data processing system 100 of the sensor signal acquisition processing system may be directly connected to the machine readable storage medium 120 to access stored information and/or data. In an alternative embodiment, the data processing system 100 of the sensor signal acquisition processing system may be implemented on a cloud platform. For example only, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-layer cloud, or the like, or any combination thereof.

The machine-readable storage medium 120 may store data and/or instructions. In an alternative embodiment, the machine-readable storage medium 120 may store data acquired from an external terminal. In an alternative embodiment, machine-readable storage medium 120 may store data and/or instructions that are used by data processing system 100 of the sensor signal acquisition processing system to perform or use the exemplary methods described herein. In alternative embodiments, machine-readable storage medium 120 may include mass storage, removable storage, volatile read-write memory, read-only memory, and the like, or any combination thereof. Exemplary mass storage devices may include magnetic disks, optical disks, solid state disks, and the like. Exemplary removable memory may include flash drives, floppy disks, optical disks, memory cards, compact disks, tape, and the like.

In a specific implementation, the plurality of processors 110 execute computer executable instructions stored by the machine readable storage medium 120, so that the processors 110 may execute a data processing method of the sensor signal acquisition processing system according to the above method embodiment, where the processors 110, the machine readable storage medium 120, and the communication unit 140 are connected through the bus 130, and the processors 110 may be used to control transceiving actions of the communication unit 140.

The specific implementation process of the processor 110 may refer to the above-mentioned method embodiments executed by the data processing system 100 of the sensor signal acquisition processing system, and the implementation principle and technical effects are similar, which are not described herein again.

In addition, the embodiment of the application also provides a readable storage medium, wherein computer executable instructions are preset in the readable storage medium, and when a processor executes the computer executable instructions, the data processing method of the sensor signal acquisition processing system is realized.

Likewise, it should be noted that in order to simplify the presentation disclosed herein and thereby aid in understanding one or more inventive embodiments, various features are sometimes grouped together in a single embodiment, figure, or description thereof. Likewise, it should be noted that in order to simplify the presentation disclosed herein and thereby aid in understanding one or more inventive embodiments, various features are sometimes grouped together in a single embodiment, figure, or description thereof.

Claims

1. A data processing method of a sensor signal acquisition processing system, the method comprising:

2. The data processing method of the sensor signal acquisition processing system according to claim 1, wherein the performing, for each of the plurality of system fault locating units, fault locating of the loaded abnormal operation data by a preceding system fault locating unit, fault locating of the abnormal operation data generated by performing heuristic search based on the fault locating data of the preceding system fault locating unit by a following system fault locating unit, includes:

aiming at each phase-dependent system fault locating unit in the plurality of system fault locating units, carrying out fault locating on target abnormal working data according to a first system fault locating unit to generate first fault locating data, wherein the first system fault locating unit is a forward system fault locating unit in each phase-dependent system fault locating unit, and the target abnormal working data refers to the abnormal working data loaded to the first system fault locating unit;

3. The method for processing data of a sensor signal acquisition processing system according to claim 2, wherein the plurality of system fault location units respectively correspond to one control tag of a sensor control activity;

4. The data processing method of the sensor signal acquisition processing system according to claim 1, wherein the network knowledge learning process of the second system fault location network includes:

5. The method for processing data of a sensor signal acquisition processing system according to claim 4, wherein each training abnormal working data a priori defines fault location annotation data, the fault location annotation data characterizing target fault location data of the training abnormal working data;

6. The method for data processing in a sensor signal acquisition processing system of claim 4, further comprising:

7. The method for data processing in a sensor signal acquisition processing system of claim 6, further comprising:

8. The data processing method of the sensor signal acquisition processing system according to any one of claims 1 to 7, wherein the first system fault location network and the second system fault location network each include a plurality of modality data location networks for acquiring modality acquisition data of abnormal operation data and performing fault location on the abnormal operation data based on different data modalities, respectively;

9. A data processing method of a sensor signal acquisition processing system according to any one of claims 1-7, characterized in that the method further comprises:

determining matching features for the tree node sequence according to the respective matching features for at least one to-be-matched fault restoration knowledge point and the respective fault restoration feature vectors for the at least one to-be-matched fault restoration knowledge point, so as to determine a tree node execution instance sequence to be restored according to the matching features for the tree node sequence and the restoration scheduling relationship, and restore the scheduling software service;

repairing the scheduling software service according to the tree node to be repaired and/or the update tree node;

10. A data processing system of a sensor signal acquisition processing system, characterized in that the data processing system of the sensor signal acquisition processing system comprises a processor and a machine-readable storage medium having stored therein machine-executable instructions loaded and executed by the processor to implement the data processing method of the sensor signal acquisition processing system of any of claims 1-9.