Operation and maintenance management method and device, electronic equipment and storage medium

Abstract

The application relates to an operation and maintenance management method and device, electronic equipment and a storage medium. The method comprises the following steps: obtaining operation and maintenance data; performing abnormality detection on the operation and maintenance data by using a multi-task prediction model to obtain a health evaluation result; wherein the multi-task prediction model comprises an expert network and a prediction network; the expert network comprises at least one shared expert network and a plurality of private expert networks; the shared expert network is used for feature extraction and fusion processing of the operation and maintenance data to obtain a global feature vector; the private expert network is used for feature processing of operation and maintenance data corresponding to a specific scene and / or a specific task to obtain a local feature vector; the prediction network is used for detecting a target feature vector obtained by fusing the global feature vector and the local feature vector to obtain the health evaluation result; the operation and maintenance quality is evaluated and analyzed based on the operation and maintenance data to obtain a target operation and maintenance evaluation result; and a target comprehensive result is determined according to the health evaluation result and the target operation and maintenance evaluation result.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of operation and maintenance management, in particular to an operation and maintenance management method and device, electronic equipment and storage medium. BACKGROUND

[0002] In modern operation and maintenance management, with the development of big data technology, the types and quantities of operation and maintenance data increase dramatically, and the business becomes complex and diversified. How to effectively collect, analyze and utilize these data has become a challenge.

[0003] The existing operation and maintenance management method mainly analyzes and evaluates through static indicators, expert experience or preset rules, which is difficult to capture new problems and changing trends in operation and maintenance work in a timely manner, and is difficult to realize state analysis and integrated evaluation of operation and maintenance data of different tasks and / or different dimensions. There is a problem of lag in problem discovery, which cannot meet the real-time and accuracy requirements.

[0004] Therefore, how to realize the real-time and accuracy of operation and maintenance monitoring and evaluation analysis is a problem to be solved. SUMMARY

[0005] Therefore, it is necessary to provide an operation and maintenance management method, device, distribution box and storage medium to solve the above technical problems.

[0006] In a first aspect, the present application provides an operation and maintenance management method, which comprises:

[0007] Obtaining operation and maintenance data in a power system; wherein the operation and maintenance data comprises at least one of network dimension data, work specification dimension data and power dimension data;

[0008] Using a multi-task prediction model to detect anomalies in the operation and maintenance data to obtain a health assessment result; wherein the multi-task prediction model comprises an expert network and a prediction network; the expert network comprises at least one shared expert network and a plurality of private expert networks; the shared expert network is used for feature extraction and fusion processing of the operation and maintenance data to obtain a global feature vector; the private expert network is used for feature processing of operation and maintenance data corresponding to a specific scene and / or a specific task to obtain a local feature vector; the prediction network is used for detecting a target feature vector obtained by fusing the global feature vector and the local feature vector to obtain the health assessment result;

[0009] Based on the operation and maintenance data, the operation and maintenance quality is evaluated and analyzed to obtain a target operation and maintenance evaluation result;

[0010] According to the health assessment result and the target operation and maintenance evaluation result, a target comprehensive result is determined.

[0011] In one of the embodiments, the operation and maintenance data in the power system is acquired, including:

[0012] The original operation and maintenance data in the power system is acquired;

[0013] The original operation and maintenance data is cleaned by using interpolation method, average value filling method and / or deletion method to obtain first data;

[0014] The first data is subjected to abnormal value identification to obtain second data;

[0015] The second data is subjected to data standardization processing to obtain the operation and maintenance data.

[0016] In one of the embodiments, the operation and maintenance quality is evaluated and analyzed based on the operation and maintenance data to obtain target operation and maintenance evaluation result, including:

[0017] The operation and maintenance data of each dimension is assigned with initial weight based on preset rule;

[0018] The importance degree score of the operation and maintenance data of each dimension is determined according to operation and maintenance target in current scene;

[0019] The initial weight of the operation and maintenance data is updated according to the importance degree score to obtain real-time weight;

[0020] The target operation and maintenance evaluation result is obtained according to evaluation score of the operation and maintenance data of each dimension and corresponding real-time weight.

[0021] In one of the embodiments, the method further includes:

[0022] The real-time operation and maintenance evaluation result in predetermined time window is acquired;

[0023] The weight optimization factor is calculated according to the real-time operation and maintenance evaluation result;

[0024] The evaluation deviation is determined by comparing the real-time operation and maintenance evaluation result with the weight optimization factor;

[0025] The real-time weight is updated according to the evaluation deviation and the weight optimization factor to obtain updated real-time weight;

[0026] The target operation and maintenance evaluation result is recalculated according to evaluation score of the operation and maintenance data of each dimension and corresponding updated real-time weight.

[0027] In one of the embodiments, the calculation expression of the weight optimization factor is as follows:

[0028] ;

[0029] wherein, indicates a start time of the predetermined time window; indicates an end time of the predetermined time window; indicates a real-time weight corresponding to the kth dimension; indicates a real-time operation and maintenance evaluation result of the operation and maintenance data of the kth dimension at time t.

[0030] In one of the embodiments, the multi-task detection model further comprises a gating network; wherein the gating network is configured to obtain a weight coefficient corresponding to each feature vector according to the global feature vector and / or the local feature vector; the expert network is connected to a plurality of the gating networks; and the gating network is connected to the prediction network in a one-to-one manner.

[0031] In one of the embodiments, the method further comprises:

[0032] In response to a viewing request of an operation and maintenance evaluation interface, the operation and maintenance evaluation interface is displayed; wherein the operation and maintenance evaluation interface is configured to display detailed information of operation and maintenance data in the power system; and the operation and maintenance evaluation interface comprises an operation and maintenance evaluation option.

[0033] When a triggering operation on the operation and maintenance evaluation option is received, the target comprehensive result is output in at least one of a chart form and a report form on the operation and maintenance evaluation interface, and / or the target comprehensive result is sent to a third-party terminal.

[0034] In a second aspect, the application further provides an operation and maintenance management device, which comprises:

[0035] An acquisition module is configured to acquire operation and maintenance data in a power system; wherein the operation and maintenance data comprises at least one of network dimension data, work specification dimension data and power dimension data.

[0036] A health evaluation module is configured to perform abnormality detection on the operation and maintenance data by using a multi-task prediction model to obtain a health evaluation result; wherein the multi-task prediction model comprises an expert network and a prediction network; the expert network comprises at least one shared expert network and a plurality of private expert networks; the shared expert network is configured to perform feature extraction and fusion processing on the operation and maintenance data to obtain a global feature vector; the private expert network is configured to perform feature processing on operation and maintenance data corresponding to a specific scene and / or a specific task to obtain a local feature vector; and the prediction network is configured to perform detection on a target feature vector obtained by fusing the global feature vector and the local feature vector to obtain the health evaluation result.

[0037] A quality evaluation module is configured to perform evaluation analysis on operation and maintenance quality based on the operation and maintenance data to obtain a target operation and maintenance evaluation result.

[0038] determining module configured to determine a target comprehensive result according to the health assessment result and the target operation and maintenance assessment result.

[0039] In a third aspect, the present application also provides an electronic device, comprising a processor and a memory for storing a computer program of the processor; wherein the processor is configured to implement the steps performed by the method according to any of the embodiments of the present application when executing the computer program.

[0040] In a fourth aspect, the present application also provides a computer readable storage medium. The computer readable storage medium has a computer program stored thereon, and the computer program, when executed by a processor, implements the steps performed by the method according to any of the embodiments of the present application.

[0041] In the operation and maintenance management method, on the one hand, through the fusion collection of multiple types of operation and maintenance data, the overall monitoring of the operation state of the power system can be realized, the misjudgment risk caused by the single dimension / static index judgment can be reduced, and the identification ability of potential faults or abnormal behaviors can be enhanced. On the other hand, through the shared expert network, the common features across tasks / scenes can be extracted, the generalization and stability of the model can be improved, through the multiple parallel processing private expert networks, fine modeling can be performed for specific scenes / tasks, the integration evaluation of operation and maintenance data of different tasks / dimensions can be realized, the abnormal detection precision can be improved, and the real-time requirement can be met. Moreover, according to the health assessment result and the target operation and maintenance assessment result, the target comprehensive result is determined, not only the state of the physical equipment of the power system, the state of the virtual network / power transmission, etc. are concerned, but also the personnel operation specification, the process execution situation, the response timeliness, etc. are combined, a comprehensive evaluation perspective is formed, and the precision and reliability of the operation and maintenance evaluation management are further improved. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 is a flowchart of an operation and maintenance management method according to an exemplary embodiment;

[0043] Figure 2 is a flowchart of an operation and maintenance management method according to an exemplary embodiment;

[0044] Figure 3 is a structural block diagram of an operation and maintenance management device according to an exemplary embodiment;

[0045] Figure 4 is an internal structure diagram of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION

[0046] In order to make the purposes, technical solutions and advantages of the present application clearer, further detailed description will be made to the present application in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and should not be used to limit the present application.

[0047] The terms "first", "second", "third" in the embodiments of the present application are only used for descriptive purposes and should not be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined with "first", "second", "third" can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "a plurality of" is at least two, such as two, three, etc., unless otherwise specifically limited. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, method, product or device including a series of steps or units is not limited to the listed steps or units, but can optionally include steps or units not listed, or can optionally include other steps or units inherent to the process, method, product or device.

[0048] In this paper, the term "embodiment" means that the specific features, structures or properties described in combination with the embodiment can be included in at least one embodiment of the present application. The phrase appears at various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment independent of or alternative to other embodiments. Those skilled in the art explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0049] In some embodiments, the operation and maintenance management method provided in the embodiments of the present application can be applied to electronic devices or cloud servers. Among them, the electronic device can be any mobile terminal or fixed terminal. The terminal can refer to a device that provides voice and / or data connectivity to a user. Exemplarily, the terminal can be an Internet of Things terminal, such as a sensor device, a mobile phone or a "cellular" phone and a computer with an Internet of Things terminal, for example, it can be a fixed, portable, pocket, handheld, computer built-in device. The cloud server can be any virtualized computing resource or physical server cluster. The server can refer to a platform that provides on-demand, scalable computing, storage, network and application services to users.

[0050] In some embodiments, as Figure 1 As shown in the figure, an operation and maintenance management method is provided, which comprises the following steps:

[0051] S101, obtaining operation and maintenance data in a power system; wherein the operation and maintenance data comprises at least one of network dimension data, work specification dimension data and power dimension data.

[0052] In the embodiments of the present application, the network dimension data can include, but is not limited to, at least one of the resource consumption amount, the failure rate, the load rate, the communication bandwidth utilization rate, the network delay, the packet loss rate, the router load, the network connectivity, the service classification data (such as service address, service port, service request mode, etc.), and the abnormality judgment data (request response time, request response quantity, request total number, etc.) of hardware / system / software / network connection resources.

[0053] In the embodiments of the present application, the work specification data can include, but is not limited to, at least one of the operation / patrol record data, the operation execution compliance data, the safety specification compliance data, and the work order processing efficiency data.

[0054] In the embodiments of the present application, the power dimension data can include, but is not limited to, at least one of the electrical dimension data (current / voltage), power, frequency, harmonic content, switch state, electrical equipment health state data, and relay protection data.

[0055] In one embodiment, the information operation and maintenance platform in the electronic device / cloud server can collect and store operation and maintenance data in the process of power system operation in real time through electronic transformers, sensors (temperature / humidity sensors), and / or network flow mirror devices, etc.

[0056] S102, performing abnormality detection on the operation and maintenance data by using a multi-task prediction model to obtain a health assessment result; wherein the multi-task prediction model includes an expert network and a prediction network; the expert network includes at least one shared expert network and a plurality of private expert networks; the shared expert network is used for feature extraction and fusion processing on the operation and maintenance data to obtain a global feature vector; the private expert network is used for feature processing on operation and maintenance data corresponding to a specific scene and / or a specific task to obtain a local feature vector; and the prediction network is used for detecting a target feature vector obtained by fusing the global feature vector and the local feature vector to obtain the health assessment result.

[0057] In the embodiments of the present application, the multi-task prediction model can include, but is not limited to, at least one of a multi-gate mixture-of-experts (MMoE) model, a progressive layered extraction (PLE) model, and a multitask deep learning (MDL) framework.

[0058] In the embodiments of the present application, the expert network indicates a professional sub-network for extracting high-order features related to the task through deep convolution, attention mechanism, etc., and usually includes multiple parallel structures. The expert network can avoid negative transfer (for example, the performance of the first task is improved, which leads to the performance of the second task to be reduced) by isolating the feature spaces of different tasks.

[0059] In the embodiments of the present application, the prediction network is used to integrate the outputs of the expert network to generate the final prediction result; the input of the prediction network is usually a fused representation of the multi-expert feature vectors.

[0060] In some embodiments, the power system has different power consumption equipment and power consumption peaks at different time periods, etc. Its network communication load, throughput is also different. For example, in summer, various regions open more refrigeration equipment, and are in the power consumption peak period, and the power system load is large, while in spring, the temperature is suitable, and compared with summer, it is in the power consumption trough, so the scenes corresponding to summer and spring are different; or, due to the various types of power distribution equipment and power consumption equipment in the power system, the use rules and health conditions are different, and the corresponding scenes are also different.

[0061] In some embodiments, the specific task can include but is not limited to at least one of a task of focusing on power load anomaly, a task of focusing on equipment state anomaly, a task of focusing on network communication anomaly, a task of focusing on network load anomaly, and a task of focusing on electrical safety anomaly.

[0062] Exemplarily, the expert network includes one shared expert network and three private expert networks; the shared expert network can include one or more parallelly arranged fully connected networks, which perform feature extraction and feature fusion on the operation and maintenance data to obtain a global feature vector; the first private expert network receives operation and maintenance data corresponding to the task of focusing on power distribution equipment anomaly, and outputs a first local feature vector; the second private expert network receives operation and maintenance data corresponding to the task of focusing on electrical safety anomaly, and outputs a second local feature vector; the third private expert network receives operation and maintenance data corresponding to the task of focusing on network communication anomaly, and outputs a third local feature vector; the prediction network performs fusion processing on the global feature vector and the local feature vectors, and performs multi-task prediction to obtain the health evaluation result of the power system.

[0063] In some embodiments, the multi-task detection model further includes a gating network; wherein the gating network is used to obtain a weight coefficient corresponding to each feature vector according to the global feature vector and / or the local feature vector; the expert network is connected to multiple gating networks; and the gating network is connected to the prediction network one by one.

[0064] In the embodiments of the present application, the gating network (Gating Network) is a neural network structure; the gating network is used to dynamically adjust the contribution degree of different expert networks or feature channels.

[0065] In an embodiment, the gating network can assign different weights to different expert networks according to the characteristics of the operation and maintenance data or task requirements; and by adjusting the weights of shared experts and private experts, the competition between tasks can be alleviated.

[0066] In some embodiments, the gating network generates a set of normalized weights through a nonlinear transformation, multiplies the weights with the outputs of the corresponding expert networks, and sums them to generate the fused features.

[0067] S103, based on the operation and maintenance data, evaluating and analyzing the operation and maintenance quality to obtain a target operation and maintenance evaluation result.

[0068] In an embodiment, the operation and maintenance data can be divided into multiple different dimension data; the electronic device can establish a linear weighting model, use the operation and maintenance data to evaluate and analyze the operation and maintenance quality, and obtain a target operation and maintenance evaluation result.

[0069] For example, the expression of the linear weighting model is as follows:

[0070] ;

[0071] wherein, indicates the weight corresponding to the ith dimension; indicates the standardized score corresponding to the ith dimension.

[0072] For example, the calculation method of the standardized score is as follows:

[0073] For positive dimensions (such as the inspection completion rate): ;

[0074] For negative dimensions (such as the power outage duration proportion): ;

[0075] wherein, indicates the minimum value of the ith type of dimension in the operation and maintenance data, representing the "worst performance" boundary; indicates the maximum value of the ith type of dimension in the operation and maintenance data, representing the "best performance" boundary.

[0076] S104, determining a target comprehensive result according to the health evaluation result and the target operation and maintenance evaluation result.

[0077] In an embodiment, the electronic device can utilize a weight superposition method to fuse the health assessment result and the target operation and maintenance assessment result to obtain a target comprehensive result of the power system; or can determine the target comprehensive result according to a comparison between the health assessment result and the target operation and maintenance assessment result and a key threshold value. For example, if the health assessment result is less than a health critical value (60), the target comprehensive result can be determined as the health assessment result, highlighting the risks in the device / operation process and the need for attention.

[0078] In an embodiment, the electronic device can evaluate and score a service provider providing operation and maintenance services for the power system according to the target comprehensive result, facilitating subsequent selection of the service provider.

[0079] In the above operation and maintenance management method, on the one hand, through the fusion collection of multiple types of operation and maintenance data, the power system operation state can be monitored comprehensively, the misjudgment risk caused by single dimension / static index judgment can be reduced, and the identification ability of potential faults or abnormal behaviors can be enhanced. On the other hand, through the shared expert network, the common features across tasks / scenes can be extracted, the generalization and stability of the model can be improved, through the multiple parallel processing private expert networks, fine modeling can be performed for specific scenes / tasks, the integration evaluation of operation and maintenance data of different tasks / dimensions can be realized, the abnormal detection accuracy can be improved, and the real-time requirement can be met. Moreover, according to the health assessment result and the target operation and maintenance assessment result, the target comprehensive result is determined, not only the state of the physical equipment of the power system, the state of the virtual network / power transmission, etc. are concerned, but also the personnel operation specification, the process execution situation, the response timeliness, etc. are combined to form a comprehensive evaluation perspective, and the accuracy and reliability of the operation and maintenance evaluation management are further improved.

[0080] In an embodiment, the operation and maintenance data in the power system are obtained, including:

[0081] The original operation and maintenance data in the power system are obtained;

[0082] The original operation and maintenance data are cleaned by using an interpolation method, an average value filling method and / or a deletion method to obtain first data;

[0083] The first data are subjected to abnormal value identification to obtain second data;

[0084] The second data are subjected to data standardization processing to obtain the operation and maintenance data.

[0085] In some embodiments, the electronic device can obtain operation data such as voltage, current, power, device temperature, and switch state from intelligent electronic devices such as transformers, circuit breakers, mutual inductors, etc. The data is extracted from power automation systems such as SCADA (Supervisory Control and Data Acquisition), EMS (Energy Management System), and DMS (Distribution Management System). SCADA focuses on real-time data acquisition and monitoring, and can provide real-time operation status of devices; EMS is used for power grid scheduling and power generation control, and contains data such as power flow and load of overall operation of the power grid; DMS focuses on distribution networks, and has data such as distribution transformer monitoring and feeder automation. The operation and maintenance business data such as inspection records (inspection time, inspection personnel, and discovered problems), fault repair and processing records (fault occurrence time, fault type, and repair time), and device account (device model, commissioning date, and manufacturer information) are obtained from production management systems (PMS) and operation and maintenance work order systems. The above data is used as raw operation and maintenance data.

[0086] In one embodiment, the electronic device can use interpolation method to process missing values for raw operation and maintenance data with strong time series, use average filling method to process missing values for raw operation and maintenance data with relatively uniform data distribution, and directly delete missing values for raw operation and maintenance data with weak context association, to obtain first data.

[0087] In one embodiment, the electronic device can identify abnormal values based on power business rules, for example, compare each piece of operation and maintenance data with electrical quantity threshold value / device state threshold value, and mark data exceeding the threshold value range as abnormal values; or, based on statistical methods, identify abnormal values in operation and maintenance data conforming to normal distribution. For example, through mean-standard deviation method (Grubbs' method), calculate the mean and standard deviation of operation and maintenance data of the same dimension, and identify and mark data outside the predetermined range as abnormal values; remove, smooth replace, or reduce the weight of the abnormal values, to obtain second data.

[0088] In some embodiments, the electronic device can normalize the dimension direction of operation and maintenance data of multiple dimensions, divide it into positive and negative dimensions, and use Min-Max scaling or Z-Score standardization to normalize the second data, to obtain operation and maintenance data.

[0089] ​In the embodiments of the present application, by performing data cleaning and outlier identification on the original operation and maintenance data, invalid interference data can be eliminated, real fault signals can be retained, data quality and reliability can be improved, the input requirements of subsequent evaluation models can be adapted, and the performance and stability of the models can be enhanced.

[0090] In some embodiments, as shown in Figure 2 The operation and maintenance quality is evaluated and analyzed based on the operation and maintenance data to obtain a target operation and maintenance evaluation result, including:

[0091] S1031, an initial weight is assigned to the operation and maintenance data of each dimension based on a preset rule;

[0092] S1032, an importance score of the operation and maintenance data of each dimension is determined according to an operation and maintenance target in a current scenario;

[0093] S1033, the initial weight of the operation and maintenance data is updated according to the importance score to obtain a real-time weight;

[0094] S1034, the target operation and maintenance evaluation result is obtained according to the evaluation score of the operation and maintenance data of each dimension and the corresponding real-time weight.

[0095] In one embodiment, the electronic device can calculate the evaluation score of the operation and maintenance data of each dimension according to a machine learning model (such as a decision tree, a neural network, a gradient model, etc.).

[0096] In some embodiments, the electronic device can set the weight corresponding to the operation and maintenance data of each dimension to be the same value. Alternatively, the weight can be non-uniformly distributed according to the importance of each dimension in the operation of the power system, for example, a lower weight is assigned to the operation and maintenance data of the network delay dimension, and a higher weight is assigned to the operation and maintenance data of the work order processing efficiency dimension. Alternatively, the contribution of each dimension feature in the machine learning model to the prediction result (evaluation score) of the model can be determined according to the SHAP value (SHapley Additive exPlanations), and the initial weight can be assigned according to the contribution.

[0097] In one embodiment, the electronic device can determine the importance degree score of the operation and maintenance data in the dimensions of fault response time, device health status, etc. according to the operation and maintenance target in the current scene, for example, the failure rate of the power system in the current quarter is significantly increased, and the operation and maintenance target is to reduce the failure rate, then the importance degree score of the operation and maintenance data in the dimensions of fault response time, device health status, etc. is higher; or the current scene is the power consumption peak period, and the operation and maintenance target is to ensure the continuous safety of power consumption, then the importance degree score of the operation and maintenance data in the electrical dimensions of load balancing and voltage stability is higher; according to the importance degree score, the initial weight (or current weight) of the operation and maintenance data is updated, the weight corresponding to the dimension with higher importance degree score is increased, and the weight corresponding to the dimension with lower importance degree score is decreased, to obtain the updated real-time weight; according to the evaluation score (such as the standardized score) of the operation and maintenance data in each dimension and the real-time weight, the target operation and maintenance evaluation result is determined.

[0098] In the embodiments of the present application, the real-time weight of each dimension is dynamically adjusted according to the operation and maintenance target in the current scene, which ensures that the target operation and maintenance evaluation result meets the actual demand, improves the accuracy and strategy guiding significance of the target operation and maintenance evaluation result, and meets the evaluation needs of different scenes; also enhances the explainability and transparency of the evaluation result, which is convenient for positioning the problem root.

[0099] In some embodiments, the method further comprises:

[0100] obtaining real-time operation and maintenance evaluation results in a predetermined time window;

[0101] calculating a weight optimization factor according to the real-time operation and maintenance evaluation results;

[0102] comparing the real-time operation and maintenance evaluation results with the weight optimization factor to determine an evaluation deviation;

[0103] updating the real-time weight according to the evaluation deviation and the weight optimization factor to obtain an updated real-time weight;

[0104] recomputing the target operation and maintenance evaluation result according to the evaluation score of the operation and maintenance data in each dimension and the corresponding updated real-time weight.

[0105] In one embodiment, the electronic device determines a first real-time operation and maintenance evaluation result according to the operation and maintenance data at a first time in a predetermined time window; determines a second real-time operation and maintenance evaluation result according to the operation and maintenance data at a second time in the predetermined time window; and so on, until the real-time operation and maintenance evaluation result at the end time of the predetermined time window is determined; calculates a weight optimization factor according to the real-time operation and maintenance evaluation results in the predetermined time window; compares the real-time operation and maintenance evaluation result at the current time with the weight optimization factor, calculates the difference value between the two, and obtains an evaluation deviation; and updates the real-time weight according to the evaluation deviation and the weight optimization factor.

[0106] In some embodiments, the calculation expression of the weight optimization factor is as follows:

[0107] ;

[0108] wherein, indicates the starting time of the predetermined time window; indicates the ending time of the predetermined time window; indicates the real-time weight corresponding to the kth dimension; indicates the real-time operation and maintenance evaluation result of the operation and maintenance data of the kth dimension at time t.

[0109] In some embodiments, the update expression of the real-time weight is as follows:

[0110] ;

[0111] wherein, indicates the learning rate, representing the step size of weight update, and the value range is 0-1; indicates the real-time weight of the kth dimension at time t; indicates the real-time weight of the kth dimension at time t+1; s indicates the weight optimization factor.

[0112] In the embodiments of the present application, in the operation and maintenance of the power system, the device state, the environmental condition (such as the load difference caused by seasonal change) and the like change with time; by obtaining the real-time evaluation result in the predetermined time window, the recent operation and maintenance data features (such as the change of the influence degree of the transformer oil temperature, the load rate and the like indicators on the operation and maintenance quality in the summer high temperature) can be captured. The calculation of the weight optimization factor and the update of the weight can make the evaluation model dynamically adapt to the current scene; and by comparing the real-time evaluation result with the weight optimization factor to determine the deviation, and then updating the weight based on the deviation, the model can be gradually corrected, for example, the newly put into operation equipment is different from the old equipment in the operation law, the initial evaluation is easy to misjudge, and the dynamic update of the weight can make the model learn the operation and maintenance data features of the new equipment, and improve the evaluation accuracy for the new scene and the new equipment.

[0113] In some embodiments, the method further comprises:

[0114] in response to a viewing request of an operation and maintenance evaluation interface, displaying the operation and maintenance evaluation interface; wherein the operation and maintenance evaluation interface is used to display the detail information of the operation and maintenance data in the power system; the operation and maintenance evaluation interface comprises an operation and maintenance evaluation option;

[0115] when receiving a triggering operation on the operation and maintenance evaluation option, outputting the target comprehensive result in at least one of a chart form and a report form on the operation and maintenance evaluation interface, and / or sending the target comprehensive result to a third-party terminal.

[0116] In some embodiments, the target comprehensive result and corresponding detailed information (such as statistical charts, classification tables, etc.) can be output in text or voice form on the display component of the electronic device; or, can be sent to a third-party terminal for output on the third-party terminal.

[0117] In the embodiments of the present application, by outputting the target comprehensive result, the operation and maintenance situation of the power system can be intuitively determined, and the operator can select a service provider or locate the problem of the power system according to the operation and maintenance situation, so as to make corresponding adjustment or processing on the power system.

[0118] It should be understood that, although each step in the flowchart involved in each of the above embodiments is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise stated herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other sequences. Moreover, as described above, at least part of the steps in the flowchart involved in each of the above embodiments can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be alternately executed with at least part of other steps or steps or stages in other steps.

[0119] Based on the same inventive concept, the embodiments of the present application also provide an operation and maintenance management device for implementing the above-mentioned operation and maintenance management method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more operation and maintenance management device embodiments provided below can refer to the limitations of the operation and maintenance management method described above, and will not be repeated here.

[0120] In one embodiment, as shown in Figure 3 An operation and maintenance management device is provided, and the device comprises:

[0121] The acquisition module 10 is configured to acquire operation and maintenance data in the power system; wherein the operation and maintenance data comprises at least one of network dimension data, work specification dimension data and power dimension data;

[0122] The health assessment module 20 is configured to perform anomaly detection on the operation and maintenance data by using a multi-task prediction model to obtain a health assessment result; wherein the multi-task prediction model comprises an expert network and a prediction network; the expert network comprises at least one shared expert network and a plurality of private expert networks; the shared expert network is configured to perform feature extraction and fusion processing on the operation and maintenance data to obtain a global feature vector; the private expert network is configured to perform feature processing on operation and maintenance data corresponding to a specific scene and / or a specific task to obtain a local feature vector; and the prediction network is configured to perform detection on a target feature vector obtained by fusing the global feature vector and the local feature vector to obtain the health assessment result.

[0123] The quality assessment module 30 is configured to perform evaluation analysis on operation and maintenance quality based on the operation and maintenance data to obtain a target operation and maintenance evaluation result.

[0124] The determination module 40 is configured to determine a target comprehensive result according to the health assessment result and the target operation and maintenance evaluation result.

[0125] In one embodiment, the acquisition module 10 is configured to perform the following steps:

[0126] Acquire original operation and maintenance data in the power system;

[0127] Perform data cleaning on the original operation and maintenance data by using an interpolation method, an average value filling method and / or a deletion method to obtain first data;

[0128] Perform outlier identification on the first data to obtain second data;

[0129] Perform data standardization processing on the second data to obtain the operation and maintenance data.

[0130] In one embodiment, the quality assessment module 30 is configured to perform the following steps:

[0131] Assign an initial weight to the operation and maintenance data of each dimension based on a preset rule;

[0132] Determine an importance score of the operation and maintenance data of each dimension according to an operation and maintenance target in a current scene;

[0133] Update the initial weight of the operation and maintenance data according to the importance score to obtain a real-time weight;

[0134] Obtain the target operation and maintenance evaluation result according to the evaluation score of the operation and maintenance data of each dimension and the corresponding real-time weight.

[0135] In one embodiment, the device further comprises:

[0136] The acquisition module 10 is configured to acquire real-time operation and maintenance evaluation results in a predetermined time window.

[0137] The calculation module is configured to calculate a weight optimization factor according to the real-time operation and maintenance evaluation results.

[0138] The comparison module is configured to compare the real-time operation and maintenance evaluation results with the weight optimization factor to determine an evaluation deviation.

[0139] The update module is configured to update the real-time weight according to the evaluation deviation and the weight optimization factor to obtain an updated real-time weight.

[0140] The determination module 40 is configured to re-calculate the target operation and maintenance evaluation result according to the evaluation scores of the operation and maintenance data of each dimension and the corresponding updated real-time weight.

[0141] In one embodiment, the calculation expression of the weight optimization factor is as follows:

[0142] ;

[0143] wherein, indicates a starting time of the predetermined time window; indicates an ending time of the predetermined time window; indicates a real-time weight corresponding to the kth dimension; indicates a real-time operation and maintenance evaluation result of the operation and maintenance data of the kth dimension at time t.

[0144] In one embodiment, the multi-task detection model further comprises a gating network; wherein the gating network is configured to acquire a weight coefficient corresponding to each feature vector according to the global feature vector and / or the local feature vector; the expert network is connected to a plurality of the gating networks; and the gating network is connected to the prediction network in a one-to-one manner.

[0145] In one embodiment, the apparatus further comprises:

[0146] The display module is configured to display the operation and maintenance evaluation interface in response to a viewing request of the operation and maintenance evaluation interface; wherein the operation and maintenance evaluation interface is configured to display detailed information of operation and maintenance data in the power system; and the operation and maintenance evaluation interface comprises an operation and maintenance evaluation option.

[0147] The display module is configured to output the target comprehensive result in at least one of a chart form and a report form on the operation and maintenance evaluation interface and / or send the target comprehensive result to a third-party terminal when receiving a triggering operation on the operation and maintenance evaluation option.

[0148] The various modules in the operation and maintenance management apparatus can be implemented in whole or in part by software, hardware, and combinations thereof. The various modules can be embedded in or independent of a processor of an electronic device in hardware form, or stored in a memory of the electronic device in software form, so as to be called and executed by a processor to perform operations corresponding to the various modules.

[0149] In one embodiment, an electronic device is provided, an internal structure diagram of which can be as shown in Figure 4 The electronic device includes a processor, a memory, a communication interface, a display unit, and an input device connected through a method bus. The processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operation method and a computer program. The internal memory provides an environment for running the operation method and the computer program in the non-volatile storage medium. The communication interface of the electronic device is configured to perform wired or wireless communication with an external terminal. The wireless communication can be achieved through WIFI, mobile cellular network, NFC (Near Field Communication), or other technologies. The computer program is executed by the processor to implement an image processing method. The display screen of the electronic device can be a liquid crystal display screen or an electronic ink display screen. The input device of the electronic device can be a touch layer overlaid on the display screen, or a key, trackball, or touchpad arranged on the shell of the electronic device, or an external keyboard, touchpad, or mouse, etc.

[0150] Those skilled in the art can understand that Figure 4 The structure shown in the above embodiment is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the electronic device to which the scheme of the present application is applied. Specifically, the electronic device can include more or fewer components than those shown in the diagram, or combine certain components, or have a different arrangement of components.

[0151] In one embodiment, a computer readable storage medium is provided, which stores a computer program. The computer program is executed by a processor to implement the steps in the above method embodiments.

[0152] In one embodiment, a computer program product is provided, which includes a computer program. The computer program is executed by a processor to implement the steps performed by the processor of the electronic device of any one of the above embodiments.

[0153] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties.

[0154] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, etc., without being limited thereto.

[0155] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0156] The above embodiments only express several embodiments of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. An operation and maintenance management method, characterized by, The method comprises: obtaining operation and maintenance data in a power system; wherein the operation and maintenance data comprises at least one of network dimension data and power dimension data; performing anomaly detection on the operation and maintenance data using a multi-task prediction model to obtain a health assessment result; wherein the multi-task prediction model comprises an expert network and a prediction network; the expert network comprises at least one shared expert network and a plurality of private expert networks; the shared expert network is used for feature extraction and fusion processing on the operation and maintenance data to obtain a global feature vector; the private expert network is used for feature processing on operation and maintenance data corresponding to a specific scene and / or a specific task to obtain a local feature vector; the prediction network is used for detecting a target feature vector obtained by fusing the global feature vector and the local feature vector to obtain the health assessment result; based on the operation and maintenance data, performing evaluation and analysis on operation and maintenance quality to obtain a target operation and maintenance evaluation result, comprising: assigning an initial weight to the operation and maintenance data of each dimension based on a preset rule; determining an importance score of the operation and maintenance data of each dimension according to an operation and maintenance target in the current scene; updating the initial weight of the operation and maintenance data according to the importance score to obtain a real-time weight; obtaining the target operation and maintenance evaluation result according to the evaluation score of the operation and maintenance data of each dimension and the corresponding real-time weight; obtaining real-time operation and maintenance evaluation results within a predetermined time window; calculating a weight optimization factor according to the real-time operation and maintenance evaluation results; comparing the real-time operation and maintenance evaluation results with the weight optimization factor to determine an evaluation deviation; updating the real-time weight according to the evaluation deviation and the weight optimization factor to obtain an updated real-time weight; recomputing the target operation and maintenance evaluation result according to the evaluation score of the operation and maintenance data of each dimension and the corresponding updated real-time weight; determining a target comprehensive result according to the health assessment result and the target operation and maintenance evaluation result.

2. The method of claim 1, wherein, The operation and maintenance data in the power system is obtained, comprising: obtaining original operation and maintenance data in the power system; performing data cleaning on the original operation and maintenance data using an interpolation method, an average value filling method and / or a deletion method to obtain first data; performing outlier identification on the first data to obtain second data; performing data standardization processing on the second data to obtain the operation and maintenance data.

3. The method of claim 1, wherein, The calculation expression of the weight optimization factor is as follows: wherein t0 indicates a starting time of the predetermined time window; t1 indicates an ending time of the predetermined time window; w k indicates a real-time weight corresponding to the kth dimension; f k k (x k (t)) indicates a real-time operation and maintenance evaluation result of the operation and maintenance data of the kth dimension at time t; x k (t) indicates the operation and maintenance data of the kth dimension at time t.

4. The method of claim 1, wherein, The multi-task detection model further comprises a gating network; wherein the gating network is used to obtain a weight coefficient corresponding to each feature vector according to the global feature vector and / or the local feature vector; the expert network is connected to a plurality of the gating networks; the gating network is connected to the prediction network one by one.

5. The method of claim 1, wherein, The method further comprises: in response to a viewing request of an operation and maintenance evaluation interface, displaying the operation and maintenance evaluation interface; wherein the operation and maintenance evaluation interface is used to display detailed information of operation and maintenance data in the power system; the operation and maintenance evaluation interface comprises an operation and maintenance evaluation option; When receiving a triggering operation on the operation and maintenance evaluation option, output the target comprehensive result on the operation and maintenance evaluation interface in at least one of a chart form and a report form, and / or send the target comprehensive result to a third-party terminal.

6. An operation and management device characterized by comprising: The device comprises: An acquisition module is configured to acquire operation and maintenance data in a power system, wherein the operation and maintenance data comprises at least one of network dimension data and power dimension data; A health evaluation module is configured to perform anomaly detection on the operation and maintenance data by using a multi-task prediction model to obtain a health evaluation result, wherein the multi-task prediction model comprises an expert network and a prediction network; the expert network comprises at least one shared expert network and a plurality of private expert networks; the shared expert network is configured to perform feature extraction and fusion processing on the operation and maintenance data to obtain a global feature vector; the private expert network is configured to perform feature processing on operation and maintenance data corresponding to a specific scene and / or a specific task to obtain a local feature vector; and the prediction network is configured to perform detection on a target feature vector obtained by fusing the global feature vector and the local feature vector to obtain the health evaluation result; A quality evaluation module is configured to perform evaluation analysis on operation and maintenance quality based on the operation and maintenance data to obtain a target operation and maintenance evaluation result, comprising: assigning an initial weight to the operation and maintenance data of each dimension based on a preset rule; determining an importance degree score of the operation and maintenance data of each dimension according to an operation and maintenance target in a current scene; updating the initial weight of the operation and maintenance data according to the importance degree score to obtain a real-time weight; and obtaining the target operation and maintenance evaluation result according to an evaluation score of the operation and maintenance data of each dimension and the corresponding real-time weight; The acquisition module is configured to acquire a real-time operation and maintenance evaluation result within a predetermined time window; A calculation module is configured to calculate a weight optimization factor according to the real-time operation and maintenance evaluation result; A comparison module is configured to compare the real-time operation and maintenance evaluation result with the weight optimization factor to determine an evaluation deviation; An update module is configured to update the real-time weight according to the evaluation deviation and the weight optimization factor to obtain an updated real-time weight; A determination module is configured to re-calculate the target operation and maintenance evaluation result according to an evaluation score of the operation and maintenance data of each dimension and the corresponding updated real-time weight; The determination module is configured to determine a target comprehensive result according to the health evaluation result and the target operation and maintenance evaluation result.

7. An electronic device, comprising: The device comprises a processor and a memory; wherein the memory is configured to store a computer program; and the processor is configured to implement the method of any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 5.

Images