Method and device for determining fault state of cable joint, and electronic equipment

Abstract

The invention discloses a method and a device for determining a fault state of a cable joint, and electronic equipment. The method comprises the following steps: acquiring target data corresponding to a cable joint in a power transmission and transformation system; determining spatial-temporal characteristics corresponding to the power transmission and transformation operation data; according to the operation data of the cable joint, determining an operation age limit characteristic corresponding to the cable joint; determining a cable defect index corresponding to the cable joint according to the predetermined item data and a cable defect relationship corresponding to the predetermined item; determining weight values of the cable joint in a plurality of fault states; and determining a target fault state of the cable joint from the plurality of fault states according to the spatial-temporal characteristics, the operating age limit characteristics, the cable defect index and the weight values corresponding to the plurality of fault states. According to the method and the device, the technical problem that the judgment accuracy is low when the fault state of the cable joint is judged by using massive and discrete cable joint related data in the related technology is solved.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a method, a device and an electronic device for determining a fault state of a cable connector. Background Art

[0002] At present, when monitoring the fault status of cable joints, the method usually adopted is to monitor the current cable joint fault status based on the current massive and discrete cable joint related data. However, since the current massive and discrete cable joint related data is not only large in volume but also highly complex, it takes a lot of time and computing costs to monitor the fault status of cable joints based only on such data, and it is difficult to accurately judge the fault status of cable joints in a timely manner, which makes it difficult for dispatching and monitoring personnel to effectively analyze the hidden dangers of cable joint defects, and there is a technical problem of low accuracy in monitoring the fault status of cable joints.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present invention provide a method, device and electronic device for determining the fault status of a cable joint, so as to at least solve the technical problem in the related art that when using massive, discrete cable joint-related data to determine the fault status of the cable joint, the judgment accuracy is low.

[0005] According to one aspect of an embodiment of the present invention, a method for determining a fault state of a cable joint is provided, comprising: obtaining target data corresponding to a cable joint in a power transmission and transformation system, wherein the target data includes power transmission and transformation operation data, cable joint operation data, and predetermined item data, wherein the predetermined item data is data corresponding to a predetermined item determined in advance, and the predetermined item is an item whose characterization index of the degree of cable joint defect is greater than a predetermined threshold; determining the spatiotemporal characteristics corresponding to the power transmission and transformation operation data; determining an operating age characteristic corresponding to the cable joint based on the cable joint operation data; determining a cable defect index corresponding to the cable joint based on the predetermined item data and a cable defect relationship corresponding to the predetermined item; determining weight values ​​of the cable joint in a plurality of fault states respectively; determining a target fault state of the cable joint from the plurality of fault states based on the spatiotemporal characteristics, the operating age characteristic, the cable defect index, and the weight values ​​corresponding to the plurality of fault states respectively.

[0006] Optionally, the target fault state of the cable joint is determined from the multiple fault states based on the spatiotemporal characteristics, the operating years characteristics, the cable defect index, and the weight values ​​corresponding to the multiple fault states, including: when the target data also includes alarm data, determining the alarm frequency change characteristics, alarm frequency comparison characteristics and alarm duration comparison characteristics corresponding to the cable joint based on the alarm data, wherein the alarm frequency comparison characteristics represent the characteristics of the alarm frequencies of multiple alarm signals respectively occurring in the total alarm frequencies, and the alarm duration comparison characteristics represent the characteristics of the durations of the multiple alarm signals respectively occurring in the total durations; the target fault state of the cable joint is determined from the multiple fault states based on the spatiotemporal characteristics, the alarm frequency change characteristics, the alarm frequency comparison characteristics, the alarm duration comparison characteristics, the operating years characteristics, the cable defect index, and the weight values ​​corresponding to the multiple fault states.

[0007] Optionally, when the target data also includes alarm data, determining the alarm frequency change characteristics, alarm frequency comparison characteristics and alarm duration comparison characteristics corresponding to the cable joint based on the alarm data, including: determining multiple alarm signal categories of the cable joint, the weights corresponding to the multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, and the durations of the alarm signals corresponding to the multiple alarm signal categories based on the alarm data; determining the multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, and the durations of the alarm signals corresponding to the multiple alarm signal categories based on the alarm data; determining the multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, and the durations of the alarm signals corresponding to the multiple alarm signal categories based on the alarm data; determining the multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, and the durations of the alarm signals corresponding to the multiple alarm signal categories; determining the multiple alarm signal categories, the alarm frequency comparison characteristics and the alarm duration comparison characteristics corresponding to the cable joint based on the alarm data ... According to the weights corresponding to the alarm signal categories, the alarm frequency change characteristics corresponding to the cable joint are determined; according to the multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories and the weights corresponding to the multiple alarm signal categories, the alarm frequency comparison characteristics corresponding to the cable joint are determined; according to the multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, the alarm signal durations corresponding to the multiple alarm signal categories and the weights corresponding to the multiple alarm signal categories, the alarm duration comparison characteristics corresponding to the cable joint are determined.

[0008] Optionally, determining the spatiotemporal features corresponding to the power transmission and transformation operation data includes: inputting the power transmission and transformation operation data into a spatial feature extraction module in a feature detection model to extract the spatial features of the power transmission and transformation operation data, wherein the feature detection model is trained based on sample data; inputting the spatial features into a temporal feature extraction module in the feature detection model to extract the temporal features of the power transmission and transformation operation data; inputting the spatial features and the temporal features into a feature fusion module in the feature detection model to determine the spatiotemporal features corresponding to the power transmission and transformation operation data.

[0009] Optionally, determining the operating life characteristics corresponding to the cable joint based on the cable joint operating data includes: when the cable joint operating data includes cable joint maintenance data and the cable joint operating life, determining an aging regression factor based on the cable joint maintenance data; determining an operating failure probability value based on the cable joint operating life; determining the operating life characteristics corresponding to the cable joint based on the aging regression factor and the operating failure probability value.

[0010] Optionally, determining the cable defect index corresponding to the cable joint based on the predetermined item data and the cable defect relationship corresponding to the predetermined item includes: obtaining the threshold type of the predetermined item; determining a defect index determination function corresponding to the predetermined item based on the threshold type of the predetermined item, wherein the defect index determination function corresponding to the predetermined item represents the cable defect relationship corresponding to the predetermined item; inputting the predetermined item data into the defect index determination function to determine the cable defect index corresponding to the cable joint.

[0011] Optionally, determining the weight values ​​of the cable joints in multiple fault states respectively includes: determining a first number of evaluation indicators for evaluating the weight values ​​of the fault states; determining multiple groups of evaluation scores, wherein each group of scores in the multiple groups of evaluation scores includes a second number of scores, and the second number is the product of the first number and the number of the multiple fault states, the multiple groups of evaluation scores are obtained by evaluating multiple target objects, and the multiple groups of evaluation scores correspond one-to-one to the multiple target objects; constructing an evaluation matrix based on the multiple groups of evaluation scores; and determining the weight values ​​of the cable joints in multiple fault states respectively based on the evaluation matrix.

[0012] According to one aspect of an embodiment of the present invention, a device for determining a fault state of a cable joint is provided, comprising: an acquisition module, used to acquire target data corresponding to a cable joint in a power transmission and transformation system, wherein the target data includes power transmission and transformation operation data, cable joint operation data, and predetermined item data, wherein the predetermined item data is data corresponding to a predetermined item determined in advance, and the predetermined item is an item whose characterization index of the degree of cable joint defect is greater than a predetermined threshold; a first determination module, used to determine the spatiotemporal characteristics corresponding to the power transmission and transformation operation data; a second determination module, used to determine the operating years characteristics corresponding to the cable joint based on the cable joint operation data; a third determination module, used to determine the cable defect index corresponding to the cable joint based on the predetermined item data and the cable defect relationship corresponding to the predetermined item; a fourth determination module, used to determine the weight values ​​of the cable joint being in a plurality of fault states respectively; and a fifth determination module, used to determine the target fault state of the cable joint from the plurality of fault states based on the spatiotemporal characteristics, the operating years characteristics, the cable defect index, and the weight values ​​corresponding to the plurality of fault states respectively.

[0013] According to one aspect of an embodiment of the present invention, an electronic device is provided, comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement any of the above-mentioned methods for determining a fault state of a cable connector.

[0014] According to one aspect of an embodiment of the present invention, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device can execute any of the above-mentioned methods for determining a fault state of a cable connector.

[0015] In an embodiment of the present invention, target data corresponding to a cable joint in a power transmission and transformation system is obtained, wherein the target data includes power transmission and transformation operation data, cable joint operation data, and predetermined item data, wherein the predetermined item data is data corresponding to a predetermined item determined in advance, and the predetermined item is an item whose characterization index of a cable joint fault is greater than a predetermined threshold. The spatiotemporal features corresponding to the power transmission and transformation operation data are determined. Based on the cable joint operation data, the operating years feature corresponding to the cable joint is determined. Based on the predetermined item data and the cable defect relationship corresponding to the predetermined item, the cable defect index corresponding to the cable joint is determined. The weight values ​​of the cable joint in multiple fault states are determined. Based on the spatiotemporal features, the operating years feature, the cable defect index, and the weight values ​​corresponding to the multiple fault states, the target fault state of the cable joint is determined from the multiple fault states. It can be seen that the embodiment of the present invention adopts a method of multi-dimensional processing and analysis of massive and discrete data. By comprehensively considering the spatiotemporal characteristics, operating years characteristics, cable defect index, and weight values ​​corresponding to multiple fault states, it comprehensively considers multiple characteristics and state weights, avoids the limitations of a single indicator, and achieves the purpose of accurately determining the fault state of the cable joint, thereby achieving the technical effect of improving the accuracy of judging the fault state of the cable joint, and further solves the technical problem of low judgment accuracy when using massive and discrete cable joint-related data to judge the fault state of the cable joint in the related technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0017] Figure 1 is a flow chart of a method for determining a fault state of a cable connector according to an embodiment of the present invention;

[0018] Figure 2 is a flow chart of a method for determining a fault state provided by an optional embodiment of the present invention;

[0019] Figure 3 is a schematic diagram of the composition of a neural network provided in an optional embodiment of the present invention;

[0020] Figure 4 is a schematic diagram of the relationship between the number of iterations and the accuracy provided by an optional embodiment of the present invention;

[0021] Figure 5 is a schematic diagram of the relationship between the number of iterations and the loss value provided in an optional embodiment of the present invention;

[0022] Figure 6 4 is a structural block diagram of a device for determining a fault state of a cable connector according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0025] First, some nouns or terms that appear in the description of the embodiments of the present application are subject to the following explanations:

[0026] CNN model: Convolutional Neural Network is a deep learning model specially designed to process data with grid structure (such as images, videos, sounds and texts, etc.). The CNN model has a unique architecture and advantages in processing such data. It can automatically detect important features in the input data without manual feature engineering. This is one of the main reasons why it has achieved remarkable results in computer vision, natural language processing, audio recognition and other fields.

[0027] LSTM model: Long Short-Term Memory network is a special type of recurrent neural network (RNN), which is mainly designed to solve the long-term dependency problem in sequence data. LSTM network performs well in processing time series data (such as speech recognition, text generation, time series prediction, etc.) because it can effectively learn and memorize long-term patterns and dependencies in the sequence without suffering from the gradient vanishing or gradient exploding problems in traditional RNN.

[0028] MLP model: Multilayer Perceptron is a feedforward artificial neural network (ANN) consisting of multiple layers, including an input layer, one or more hidden layers, and an output layer. The MLP model learns and recognizes complex patterns in input data through the connections between neurons (or nodes) in each layer. Each neuron receives weighted inputs from neurons in the previous layer and processes these inputs through an activation function to produce outputs. The MLP model performs well in processing various types of data (such as image recognition, classification tasks, regression problems, etc.) because it can approximate complex nonlinear functional relationships by adjusting the connection weights and biases between neurons, thereby achieving accurate prediction and classification of data.

[0029] Example 1

[0030] According to an embodiment of the present invention, an embodiment of a method for determining a fault state of a cable connector is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0031] Figure 1 is a flow chart of a method for determining a fault state of a cable connector according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0032] Step S102, obtaining target data corresponding to the cable joint in the power transmission and transformation system, wherein the target data includes power transmission and transformation operation data, cable joint operation data, and predetermined item data, the predetermined item data is data corresponding to a predetermined item determined in advance, and the predetermined item is an item whose characterization index of the degree of cable joint defect is greater than a predetermined threshold;

[0033] In step S102 provided in the present application, target data corresponding to the cable connector in the power transmission and transformation system is obtained.

[0034] Among them, cable joints are involved. Cable joints are the connection points formed when the cable sections are connected as a whole to make it a continuous line after the cable is laid. Cable joints may fail due to various reasons such as aging, wear and moisture during long-term use.

[0035] Among them, target data is involved. Target data refers to data related to cable joints and used to monitor and analyze the fault status of cable joints, including power transmission and transformation operation data, cable joint operation data, and predetermined item data.

[0036] Among them, the power transmission and transformation operation data is involved. The power transmission and transformation operation data is data that can reflect the overall operation status of the cable joint and the power transmission and transformation system in which it is located, and can include continuous numerical time series data and discrete time series data;

[0037] Among them, the cable joint operation data is involved. The cable joint operation data is data that can reflect and evaluate the repair and operation status of the cable joint, which can be the repair data of the cable joint and the operation life data of the cable joint;

[0038] Among them, predetermined item data is involved, and the predetermined item data is data corresponding to a predetermined item that is predetermined and related to the degree of cable joint defects. The predetermined item is predetermined based on experience, standards or safety requirements, and its characterization index of the degree of cable joint defects is greater than a predetermined threshold. The predetermined threshold is the minimum value that can characterize the degree of cable joint defects. For example, if the characterization index corresponding to the winding insulation resistance change rate exceeds the predetermined threshold, it means that the winding insulation resistance change rate can represent the degree of cable joint defects. In this case, the winding insulation resistance change rate can be used as a predetermined item, and the corresponding winding insulation resistance change rate data is the predetermined item data under the predetermined item.

[0039] Step S104, determining the spatiotemporal characteristics corresponding to the power transmission and transformation operation data;

[0040] In step S104 provided in the present application, the spatiotemporal characteristics corresponding to the power transmission and transformation operation data are determined.

[0041] Among them, the time and space characteristics are involved, and the time and space characteristics are specifically manifested as the characteristics of the cable joint in time and space.

[0042] Since the power transmission and transformation operation data reflects the overall operating status of the cable joint and the power transmission and transformation system in which it is located, it can reflect the changes, fluctuations or regularities of the cable joint during operation. Through in-depth analysis of these power transmission and transformation operation data, the spatiotemporal characteristics corresponding to the power transmission and transformation operation data can be extracted.

[0043] It should be noted that the method for extracting the spatiotemporal features may be a deep learning algorithm model, which is not specifically limited here and can be customized according to actual applications and scenarios.

[0044] Since the spatiotemporal characteristics integrate the operating status information of the cable joint at different time points, as well as the changes of this status information in the spatial dimension (such as location, environmental impact, etc.), this multi-dimensional information fusion can more comprehensively reflect the real operation of the cable joint. Moreover, by analyzing the spatiotemporal characteristics, the specific mode of cable joint failure can be revealed, such as the association between the failure and the specific time (season, day and night changes), the specific spatial location (such as the environment where the cable joint is located, underground or above ground), which can help maintenance personnel understand and predict the possible causes and timing of the failure, and help improve the accuracy of fault status prediction.

[0045] Step S106, determining the operating age characteristics corresponding to the cable joint according to the cable joint operating data;

[0046] In step S106 provided in the present application, the operating age characteristics corresponding to the cable joint are determined based on the cable joint operating data.

[0047] Among them, the operating life characteristics are involved. The operating life characteristics are specifically manifested as the characteristics of the cable joint during use as the service life increases, which can reflect the aging condition of the cable joint.

[0048] Since cable joints are affected by multiple factors such as current, voltage, and temperature during operation, their performance will gradually change over time. Accordingly, these performance changes will be directly reflected in the cable joint operation data. Therefore, through in-depth analysis of these cable joint operation data, the operating age characteristics corresponding to the cable joint can be extracted.

[0049] It should be noted that there is no specific limitation on the method for extracting the operating life feature, and it can be customized according to actual applications and scenarios.

[0050] Since the service life characteristics directly reflect the degree of aging of the cable joint over time, it is an important parameter for evaluating its failure tendency. The material of the cable joint will gradually deteriorate over time, and its key properties such as insulation performance and mechanical strength may decrease, thereby increasing the risk of failure. Therefore, through the service life, this aging effect can be quantified, the degree of aging of the cable joint can be accurately evaluated, and then its impact on the fault state can be evaluated. Moreover, the operating state of the cable joint has a certain correlation with the age, and the service life characteristics can help predict the failure trend. For example, the cable joint may rarely fail in the early stage, but as the service time increases, the failure frequency may gradually increase, so the fault state can be better determined.

[0051] Step S108, determining a cable defect index corresponding to the cable joint according to the predetermined item data and the cable defect relationship corresponding to the predetermined item;

[0052] In step S108 provided in the present application, a cable defect index corresponding to the cable joint is determined based on the predetermined item data and the cable defect relationship corresponding to the predetermined item.

[0053] Among them, the cable defect relationship is involved. The cable defect relationship is a correspondence between predetermined items and the degree of cable defect, which can be expressed in the form of a function. Specifically, the cable defect relationship refers to the correlation between various monitoring data, operating parameters, external environmental factors, etc. during the operation of the cable and possible defects or failures of the cable. This relationship can be established through data mining, statistical analysis or machine learning models to identify which parameter changes or combinations are most indicative of the defective state of the cable. For example, if the temperature of the cable is monitored to be abnormally high, it is often related to the aging or overload of the insulating material inside the cable; similarly, if the high-frequency partial discharge signal increases, it may indicate the presence of partial discharge at the cable joint. These are all manifestations of the cable defect relationship.

[0054] Among them, the cable defect index is also involved. The cable defect index can quantify the degree of defect of the cable joint to better reflect the degree of cable defects.

[0055] In this step, first, multiple predetermined items that can characterize the degree of cable joint defects can be pre-determined based on actual experience and scenarios. Secondly, a cable defect relationship corresponding to the predetermined items is established, and then the data corresponding to the multiple predetermined items are obtained. Finally, based on the predetermined item data and the cable defect relationship corresponding to the predetermined items, the cable defect index corresponding to the cable joint is determined to quantify the degree of defect of the cable joint. This process achieves a quantitative assessment of the degree of defect of the cable joint by pre-determining the predetermined items and establishing a corresponding relationship between the predetermined items and the degree of cable defect, which helps to more accurately identify potential problems or risks in the cable joint in the future.

[0056] Step S110, determining weight values ​​of the cable connector being in a plurality of fault states respectively;

[0057] In step S110 provided in the present application, weight values ​​of the cable joints being in a plurality of fault states are determined.

[0058] Therein, multiple fault states are involved, and the multiple fault conditions are various fault states that may occur in the cable connector.

[0059] When evaluating the possible fault states of cable joints, the relative importance or frequency of occurrence of various possible abnormal or fault states (such as minor abnormality, general abnormality, serious abnormality, fault, etc.) is given to quantify the impact of different states in the comprehensive evaluation.

[0060] It should be noted that the method for determining the weight value corresponding to each fault state may be a fuzzy comprehensive evaluation method, which is not specifically limited here and can be customized according to actual applications and scenarios.

[0061] By setting weight values, parameters can be allocated according to the actual impact of each fault state, thereby improving the accuracy of fault prediction states.

[0062] Step S112, determining a target fault state of the cable joint from the multiple fault states based on the spatiotemporal characteristics, the operating years characteristics, the cable defect index, and the weight values ​​corresponding to the multiple fault states.

[0063] In step S112 provided in the present application, a target fault state of the cable joint is determined from multiple fault states based on spatiotemporal characteristics, operating years characteristics, cable defect index, and weight values ​​corresponding to multiple fault states.

[0064] After obtaining the spatiotemporal characteristics, operating years characteristics, cable defect index, and weight values ​​corresponding to multiple fault states, they are analyzed and the state that best matches the target data is selected from multiple fault states as the target fault state of the cable joint. It comprehensively considers multiple characteristics and state weights, avoids the limitations of a single indicator, achieves the purpose of accurately determining the fault state of the cable joint, and thus achieves the technical effect of improving the accuracy of cable joint fault state judgment.

[0065] It should be noted that the method for determining the target fault state of the cable connector from multiple fault states is not specifically limited here and can be adaptively set according to actual applications and scenarios.

[0066] Through the above steps S102-S112, the target data corresponding to the cable joint in the power transmission and transformation system is obtained, wherein the target data includes the power transmission and transformation operation data, the cable joint operation data, and the predetermined item data, the predetermined item data is the data corresponding to the predetermined item determined in advance, and the predetermined item is the item whose characterization index of the cable joint fault is greater than the predetermined threshold. Determine the spatiotemporal characteristics corresponding to the power transmission and transformation operation data. Based on the cable joint operation data, determine the operating years characteristics corresponding to the cable joint. Based on the predetermined item data and the cable defect relationship corresponding to the predetermined item, determine the cable defect index corresponding to the cable joint. Determine the weight values ​​of the cable joint in multiple fault states respectively. Based on the spatiotemporal characteristics, the operating years characteristics, the cable defect index, and the weight values ​​corresponding to the multiple fault states, determine the target fault state of the cable joint from the multiple fault states. It can be seen that the embodiment of the present invention adopts a method of multi-dimensional processing and analysis of massive and discrete data. By comprehensively considering the spatiotemporal characteristics, operating years characteristics, cable defect index, and weight values ​​corresponding to multiple fault states, it comprehensively considers multiple characteristics and state weights, avoids the limitations of a single indicator, and achieves the purpose of accurately determining the fault state of the cable joint, thereby achieving the technical effect of improving the accuracy of judging the fault state of the cable joint, and further solves the technical problem of low judgment accuracy when using massive and discrete cable joint-related data to judge the fault state of the cable joint in the related technology.

[0067] As an optional embodiment, based on spatiotemporal characteristics, operating years characteristics, cable defect index, and weight values ​​corresponding to multiple fault states, a target fault state of the cable joint is determined from multiple fault states, including: when the target data also includes alarm data, based on the alarm data, determining the alarm frequency change characteristics, alarm frequency comparison characteristics, and alarm duration comparison characteristics corresponding to the cable joint, wherein the alarm frequency comparison characteristics represent the characteristics of the alarm frequencies of multiple alarm signals accounting for the total alarm frequencies, and the alarm duration comparison characteristics represent the characteristics of the durations of multiple alarm signals accounting for the total durations; based on spatiotemporal characteristics, alarm frequency change characteristics, alarm frequency comparison characteristics, alarm duration comparison characteristics, operating years characteristics, cable defect index, and weight values ​​corresponding to multiple fault states, a target fault state of the cable joint is determined from multiple fault states.

[0068] In this embodiment, a further step is described for determining a target fault state of a cable joint from a plurality of fault states based on spatiotemporal characteristics, operating years characteristics, a cable defect index, and weight values ​​corresponding to the plurality of fault states, when the target data includes alarm data.

[0069] Among them, alarm data is involved. In the power system, the operating status of the cable joint is continuously monitored by various sensors and monitoring equipment. When the monitored parameters exceed the normal range or an abnormal situation occurs, the system will generate alarm data, which is the first-hand early warning information of possible problems with the cable joint.

[0070] Among them, the alarm frequency change feature is also involved. The alarm frequency change feature is used to characterize the degree of degradation trend of the cable joint operation, that is, the alarm frequency change feature refers to the change in the number of alarm signals generated by the cable joint within a specific time period. For example, whether the monitored alarm frequency shows an upward trend, this can reflect that the degree of deterioration of the cable joint operation status is relatively severe.

[0071] Among them, the alarm frequency comparison feature is also involved. The alarm frequency comparison feature indicates the proportion of the alarm frequency of various alarm signals to the total alarm frequency. This helps to identify which type of alarm is more frequent, which may mean which type of fault is more common or needs more attention.

[0072] Among them, the alarm duration comparison feature is also involved. The alarm duration comparison feature indicates the percentage of the duration of each alarm signal to the total alarm duration. A longer alarm duration may mean the persistence or severity of the fault, which requires more attention and processing.

[0073] In the steps involved in this embodiment, it not only takes into account the spatiotemporal characteristics, years of operation, cable defect index, etc., but also specifically incorporates alarm data, and further refines the evaluation of the cable joint status by analyzing the alarm frequency change characteristics, alarm frequency comparison characteristics, and alarm duration comparison characteristics. After these characteristics are quantified, together with other characteristics and the weight value of the fault state, the possible fault state of the cable joint is comprehensively predicted. That is, by comprehensively considering the spatiotemporal characteristics and multiple alarm signal characteristics, the operating status of the cable joint can be fully monitored, and the comprehensiveness and accuracy of the early warning can be improved.

[0074] Among them, by analyzing the changes in alarm frequency and duration, subtle changes in the cable connector status can be captured in time, achieving early warning and preventing serious faults from occurring.

[0075] It should be noted that in the above process, deep learning technology, such as LSTM (Long Short-Term Memory Network), can be used to process the time series characteristics of alarm data, capture the complex associations and long-term dependencies between alarm signals, and improve the accuracy of feature extraction. In addition, a dynamic weight adjustment mechanism can be developed to dynamically adjust the weight of each fault state according to the real-time changes in the power grid operating environment and the latest status of the cable connector, so that the prediction is closer to the actual operating conditions and improves the real-time and flexibility of the prediction.

[0076] For example, for a specific cable joint, if a high-frequency abnormal temperature alarm suddenly appears recently, the weights of the abnormal temperature alarm frequency change characteristics and frequency comparison characteristics can be increased to reflect the significant impact of this abnormal situation on the fault state of the cable joint, so as to better predict the fault.

[0077] As an optional embodiment, when the target data also includes alarm data, the alarm frequency change characteristics, alarm frequency comparison characteristics and alarm duration comparison characteristics corresponding to the cable joint are determined based on the alarm data, including: determining multiple alarm signal categories of the cable joint, the weights corresponding to the multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, and the durations of the alarm signals corresponding to the multiple alarm signal categories based on the alarm data; determining the alarm frequency change characteristics corresponding to the cable joint based on the multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, and the weights corresponding to the multiple alarm signal categories; determining the alarm frequency comparison characteristics corresponding to the cable joint based on the multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, and the weights corresponding to the multiple alarm signal categories; determining the alarm duration comparison characteristics corresponding to the cable joint based on the multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, the durations of the alarm signals corresponding to the multiple alarm signal categories, and the weights corresponding to the multiple alarm signal categories.

[0078] In this embodiment, specific steps are described for determining the alarm frequency change characteristics, alarm frequency comparison characteristics and alarm duration comparison characteristics corresponding to the cable joint based on the alarm data when the target data also includes alarm data.

[0079] Among them, the alarm signal category is involved. The alarm signal category is used to characterize alarm signals of different severity. For example, there are usually four categories of multiple alarm signal categories for cable joints in this field: accident, abnormality, over-limit, and notification.

[0080] Among them, the weights corresponding to multiple alarm signal categories are involved. The weights corresponding to the multiple alarm signal categories are used to characterize the degree of influence of each alarm signal on the operating status of the cable joint. There is no specific restriction on the method for determining the weight here. For example, in the above-mentioned four types of alarm signals, namely accidents, abnormalities, over-limit, and notifications, the hierarchical analysis method and entropy weight method can be used to quantify the degree of influence of the above four types of alarm signals on the operating status of the cable joint according to their severity, so as to determine the weights corresponding to the multiple alarm signal categories.

[0081] Among them, the alarm frequencies corresponding to multiple alarm signal categories are involved, and the alarm frequencies corresponding to multiple alarm signal categories are used to represent the number of times each alarm signal appears.

[0082] Among them, the duration of the alarm signal corresponding to multiple alarm signal categories is also involved. The duration of the alarm signal corresponding to multiple alarm signal categories is used to represent the duration of each alarm signal after it appears.

[0083] After obtaining multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, and the weights corresponding to the multiple alarm signal categories, the alarm frequency change characteristics corresponding to the cable joint can be determined based on the above data. By comparing the alarm frequencies of the same alarm signal category at different time points, the change trend of the alarm signal over time can be analyzed, and then the change trend of the cable joint status can be evaluated for better subsequent predictions.

[0084] After obtaining multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, and the weights corresponding to the multiple alarm signal categories, the alarm frequency comparison characteristics corresponding to the cable joint can be determined based on the above data. By calculating the proportion of different alarm signal categories in the total alarm frequency, it is possible to understand which type of alarm is more frequent, thereby identifying potential problem areas of the cable joint for better subsequent prediction.

[0085] After obtaining multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, the duration of the alarm signals corresponding to the multiple alarm signal categories, and the weights corresponding to the multiple alarm signal categories, it is necessary to determine the alarm duration comparison characteristics corresponding to the cable joint based on the above data. By analyzing the distribution of durations of different alarm signal categories, alarms with longer durations can be identified. These alarms may correspond to more serious fault conditions, so as to make better subsequent predictions.

[0086] As an optional embodiment, determining the spatiotemporal features corresponding to power transmission and transformation operation data includes: inputting the power transmission and transformation operation data into a spatial feature extraction module in a feature detection model to extract the spatial features of the power transmission and transformation operation data, wherein the feature detection model is trained based on sample data; inputting the spatial features into a temporal feature extraction module in the feature detection model to extract the temporal features of the power transmission and transformation operation data; inputting the spatial features and the temporal features into a feature fusion module in the feature detection model to determine the spatiotemporal features corresponding to the power transmission and transformation operation data.

[0087] In this embodiment, specific steps of determining the spatiotemporal features corresponding to the power transmission and transformation operation data are described.

[0088] Among them, a feature detection model is involved, and the feature detection model is obtained by training the sample data. The model is usually based on machine learning or deep learning technology. In this embodiment, it can be a model obtained by combining CNN (convolutional neural network) and LSTM (long short-term memory network) to automatically identify and extract features that are helpful for analysis and prediction from the input features.

[0089] Among them, a spatial feature extraction module is involved, which can be a module constructed based on the CNN network model. In the feature detection model, it is used to extract the characteristics of the spatial distribution of data, such as the distribution pattern of partial discharge data at different positions of the cable joint, or the gradient change of temperature data on the surface of the equipment.

[0090] Among them, a time feature extraction module is involved, which can be a module constructed based on the LSTM network model. In the feature detection model, it is used to analyze the changing trend of data over time, such as the temperature fluctuation over time, the changing law of local discharge intensity, etc.

[0091] Among them, the feature fusion module is also involved, which comprehensively processes the spatial features and time features to generate a more comprehensive and deeper data feature representation, which is particularly important for the state monitoring and fault prediction of complex systems. That is, it can form spatiotemporal features that can fully characterize the state of the cable joint. Through the fused spatiotemporal features, the comprehensive operating state of the cable joint under different time and space environments can be captured, thereby providing accurate data support for fault judgment.

[0092] In the steps involved in this embodiment, the feature detection model is used to extract and fuse the features of the power transmission and transformation operation data. First, the original power transmission and transformation operation data is input into the spatial feature extraction module in the model, which automatically identifies and extracts the spatial characteristics in the data through an algorithm (such as CNN), and then the extracted spatial features are sent to the temporal feature extraction module (such as LSTM) to analyze the change patterns of these spatial features over time, such as the fluctuation trend of the intensity of partial discharge signals. Finally, the spatial features and temporal features are integrated into the feature fusion module, which comprehensively processes these two types of features to generate a spatiotemporal feature that contains spatial and temporal dimension information.

[0093] By extracting spatiotemporal features, we can capture the dynamic changes and spatial distribution characteristics of the operating status, provide more comprehensive data support for fault prediction, and improve the accuracy of prediction. The fusion of spatial and temporal features enhances the adaptability of the model to different types of data and improves the generalization ability of the model in different scenarios.

[0094] For example:

[0095] The CNN-LSTM model can be used to determine the spatiotemporal features corresponding to the power transmission and transformation operation data. First, the power transmission and transformation operation data of the cable joint is input into the CNN module, and the convolution layer and pooling layer of the CNN module are used to capture the spatial features in the power transmission and transformation operation data. The convolution operation can identify the local features of the data distribution on the cable joint, and the pooling layer compresses the features by downsampling, thereby removing redundant information and retaining important features, such as temperature distribution. Secondly, the spatial features output by the CNN are passed to the LSTM module as an input sequence. The LSTM module models the changes of these spatial features at different time steps through its memory gate and forget gate structure to capture the dynamic changes in time. For example, LSTM can identify the temperature fluctuations of the cable joint in different time periods, thereby establishing features in the time dimension. The spatial features extracted by CNN and the temporal features extracted by LSTM are integrated to form a complete spatiotemporal feature. A comprehensive analysis of the acquired spatiotemporal features can capture the comprehensive operating status of the cable joint under different time and space environments, thereby providing accurate data support for the subsequent step of fault judgment.

[0096] As an optional embodiment, the operating life characteristics corresponding to the cable joint are determined based on the cable joint operating data, including: when the cable joint operating data includes the cable joint maintenance data and the cable joint operating life, determining the aging regression factor based on the cable joint maintenance data; determining the operation failure probability value based on the cable joint operating life; determining the operating life characteristics corresponding to the cable joint based on the aging regression factor and the operation failure probability value.

[0097] In this embodiment, specific steps of determining the operating age characteristics corresponding to the cable joint based on the cable joint operating data are described.

[0098] Among them, cable joint maintenance data is involved. Cable joint maintenance data refers to the records of regular or irregular maintenance, inspection, and repair activities carried out during the service life of the cable joint, including the time, type (such as major repairs, minor repairs), reasons, results and other detailed information of the maintenance.

[0099] Among them, the service life of the cable joint is involved. The service life of the cable joint refers to the cumulative time from the first time the cable joint was put into use to the present time. The service life is an important parameter for evaluating the aging of power equipment and predicting its failure probability.

[0100] Among them, the aging regression factor is involved. The aging regression factor is a characteristic value of the operating years extracted based on the cable joint maintenance data, which is used to characterize the health status of the cable joint after different degrees of maintenance. In the overhaul and maintenance of power equipment, the performance and life of the equipment will be restored after repair, but usually it will not return to a brand new state. The aging regression factor is a numerical expression used to quantify the degree of performance degradation of the equipment after maintenance relative to its most original state or the state of the new equipment. It is used to characterize the health status of the cable joint after different degrees of maintenance.

[0101] Among them, the operational failure probability value is also involved. In statistics, it refers to the probability of a specific device failing within a specific operating period, which can be predicted based on historical failure data, the operating conditions of the equipment, and environmental conditions. For example, as the operating time increases, the aging process of the cable joint will accelerate, and the failure probability will increase accordingly. Therefore, the higher the operational failure probability value, the greater the possibility that the cable joint will fail within a specific time in the future. The operational failure probability value can be determined in combination with a statistical model, such as the Weibull distribution model. The Weibull distribution model is a statistical model used to describe the law of equipment failure rate changes over time, and can better fit the failure behavior of power equipment. The specific model is not limited here, and can be set adaptively according to the actual application and scenario.

[0102] In the steps involved in this embodiment, the aging regression factor is determined based on the cable joint maintenance data. Since this factor reflects the degree of performance recovery of the cable joint after maintenance, if the cable joint is overhauled, the aging regression factor may be 0.7, while it may be 0.4 after a minor repair. This shows that even after maintenance, the performance recovery of the cable joint may not reach a brand new state, and the degree of recovery after overhaul is higher. Next, according to the service life of the cable joint, the Weibull distribution model or other statistical models can be used to determine the operation failure probability value, that is, the probability of failure of the cable joint under the current service life. Finally, the aging regression factor and the operation failure probability value are combined to determine the service life characteristics of the cable joint. This characteristic not only reflects the natural aging process of the cable joint over time, but also takes into account the impact of maintenance on its performance recovery, so that the evaluation result is closer to the actual operating conditions.

[0103] By quantifying the aging regression factor and operational failure probability value, the operating status and potential failures of cable connectors can be predicted more accurately, providing data support for preventive maintenance.

[0104] As an optional embodiment, based on predetermined item data and the cable defect relationship corresponding to the predetermined item, a cable defect index corresponding to the cable joint is determined, including: obtaining the threshold type of the predetermined item; determining a defect index determination function corresponding to the predetermined item based on the threshold type of the predetermined item, wherein the defect index determination function corresponding to the predetermined item represents the cable defect relationship corresponding to the predetermined item; inputting the predetermined item data into the defect index determination function to determine the cable defect index corresponding to the cable joint.

[0105] In this embodiment, specific steps of determining a cable defect index corresponding to a cable joint based on predetermined item data and a cable defect relationship corresponding to the predetermined item are described.

[0106] Among them, the threshold type of the predetermined item is involved. The predetermined item usually refers to the key parameters or indicators used to characterize the health status of the equipment, such as temperature, partial discharge intensity, insulation resistance, etc. The threshold type refers to the normal range and abnormal warning boundary set for these parameters, which are divided into absolute thresholds and relative thresholds. The absolute threshold is set according to the limit value of the safe operation of the equipment, such as the upper limit and lower limit, while the relative threshold is set based on the statistical analysis of the historical data of the equipment, such as the mean and standard deviation of the equipment performance.

[0107] Among them, a defect index determination function is involved, which is used to quantify and evaluate the degree to which the predetermined item parameters deviate from the normal range, and the impact of this deviation on the health status of the cable joint. The function takes the predetermined item data of the cable joint as input and outputs a value indicating the degree of cable defect. The larger the value, the more serious the defect of the cable joint.

[0108] In the steps involved in this embodiment, first, the predetermined item data related to the cable joint status monitoring and the threshold types of these predetermined items are obtained. Then, according to the threshold type of each predetermined item, the corresponding defect index determination function is determined, which describes the defect relationship between the predetermined item data of the cable joint and its health state, that is, when the specific parameters of the cable joint deviate from the normal range, how this deviation will affect the health state of the cable joint, which is specifically manifested as the calculation of the defect index. Finally, the predetermined item data of the cable joint is input into the corresponding defect index determination function, and the defect index of the cable joint is calculated to quantify the health state and failure risk of the cable joint.

[0109] Through the defect index determination function, the predetermined item data of the cable joint can be accurately quantified into the cable defect index, making the condition evaluation of the cable joint more objective, quantitative, and easy to compare and analyze.

[0110] As an optional embodiment, determining the weight values ​​of a cable joint in multiple fault states respectively includes: determining a first number of evaluation indicators for evaluating the weight values ​​of the fault states; determining multiple groups of evaluation scores, wherein each group of scores in the multiple groups of evaluation scores includes a second number of scores, and the second number is the product of the first number and the number of the multiple fault states, the multiple groups of evaluation scores are obtained by evaluating multiple target objects, and the multiple groups of evaluation scores correspond one-to-one to the multiple target objects; constructing an evaluation matrix based on the multiple groups of evaluation scores; and determining the weight values ​​of the cable joint in multiple fault states respectively based on the evaluation matrix.

[0111] In this embodiment, specific steps of determining weight values ​​of a cable joint being in a plurality of fault states are described.

[0112] Among them, a first number of evaluation indicators are involved, and the evaluation indicators are parameters or characteristics used as evaluation criteria in fault state assessment, which are used to quantify various aspects of the cable joint fault state.

[0113] Among them, a second number of scores is involved, which is a set of scores given by multiple target objects (such as experts, field engineers or data analysis models) for each evaluation indicator under different fault conditions. Each set of evaluation scores contains a score for each fault condition, reflecting the evaluation results of different target objects for the same evaluation indicator under different fault conditions.

[0114] For the first and second quantities mentioned above, the first quantity refers to the number of evaluation indicators, and the second quantity refers to the total number of scores in each group of evaluation scores, which is equal to the first quantity (the number of evaluation indicators) multiplied by the number of multiple fault states. For example, if there are 5 evaluation indicators and 4 fault states, then each group of evaluation scores will contain 20 scores.

[0115] Among them, the evaluation matrix is ​​involved. The evaluation matrix is ​​a matrix in which multiple groups of evaluation scores are arranged according to the structure of evaluation indicators and fault states. Each row represents an evaluation indicator, each column represents a fault state, and each element in the matrix is ​​the score of the target object for the indicator under the fault state.

[0116] In the steps involved in this embodiment, it is necessary to determine a set of evaluation indicators that cover the key aspects of the operating status of the cable joint. Then, multiple target objects (experts or intelligent models) give evaluation scores for each indicator under each fault state to form multiple groups of evaluation scores, each of which includes the scores under all fault states. Next, the multiple groups of evaluation scores are organized into an evaluation matrix, the rows of which correspond to the evaluation indicators, the columns correspond to the fault states, and the elements in the matrix are the scores given by the target objects. Finally, by mathematically processing the evaluation matrix (for example, fuzzy comprehensive evaluation method or statistical methods such as average and weighted average), the weight values ​​of the cable joint under each fault state are determined, and these weight values ​​will reflect the possibility of the cable joint being in different fault states.

[0117] Specifically, the cable joint will show different fault states under different operating conditions. In order to evaluate the impact of these fault states on the overall operation of the cable joint, the weight values ​​of the cable joint in multiple fault states are introduced to indicate the relative impact of each fault state on the overall operation of the cable joint. First, it is necessary to determine the first number of evaluation indicators used to evaluate the weight values ​​of the fault states. These indicators should be able to effectively reflect the characteristics of various fault states of the cable joint. For example, the first number of evaluation indicators is defined as 4 evaluation indicators, and these 4 evaluation indicators will be used to quantify the impact of each fault state on the overall operation of the cable joint. Then, multiple groups of evaluation scores are determined, wherein each group of scores in the multiple groups of evaluation scores includes a second number of scores, and the second number is the product of the first number and the number of multiple fault states. For example, 4 evaluation indicators are defined and there are 4 fault states, then the "second number" should be 16, indicating that each group of evaluation scores contains 16 scores. Multiple groups of evaluation scores are obtained by evaluating multiple target objects, and multiple groups of evaluation scores correspond to multiple target objects one by one. For example, multiple experienced experts in this field are used as target objects to score each fault state of the cable joint. The multiple groups of evaluation scores of all target objects are summarized to construct an evaluation matrix. The rows of the evaluation matrix represent different evaluation indicators for the current fault state, and the columns represent the scores of the current fault state under different evaluation indicators. The data of each column in the evaluation matrix can be normalized so that different evaluation indicators have the same dimension. According to the evaluation matrix, the weight values ​​of the cable joints in multiple fault states are determined. The weight value corresponding to each evaluation indicator can be determined according to the relative importance of each evaluation indicator to the fault state, and then the final weighted total score is obtained by weighted average, that is, the weight value for each fault state. The higher the weight value of the fault state, the greater its impact on the cable joint. By assigning weight values ​​to different fault states of the cable joint, the impact of each fault state on the cable joint can be more objectively evaluated, providing data support for the subsequent judgment of the fault state of the cable joint.

[0118] By determining the weight value of the fault state, a refined evaluation of the health status of the cable joint can be achieved, avoiding the limitations of a single fault signal evaluation and improving the accuracy and comprehensiveness of fault diagnosis. Moreover, the determination of multiple groups of evaluation scores not only takes into account expert experience, but also may combine the evaluation results of the intelligent analysis model, making the evaluation process more objective and scientific.

[0119] Based on the above embodiments and optional embodiments, an optional implementation is provided, which is described in detail below.

[0120] In the related art, when using massive, discrete cable joint related data to judge the cable joint fault status, there is a technical problem of low judgment accuracy. For example: the judgment of the cable joint fault status relies on massive, discrete power data, but the current massive, discrete power data has not been classified or judged in any way. Directly using these data to judge the cable joint fault status requires a lot of time and computing costs, and the judgment accuracy of the cable joint fault status is very low.

[0121] In view of this, an optional embodiment of the present invention provides a method and system for determining the fault status of a cable joint, which can also be called a multi-parameter status assessment method and system for a distribution cable joint, which can effectively solve the technical problem of low judgment accuracy when using massive and discrete cable joint-related data to judge the fault status of the cable joint in the related technology.

[0122] (I) Method for determining the fault status of cable joints:

[0123] Figure 2 is a flow chart of a method for determining a fault state provided by an optional embodiment of the present invention, such as Figure 2 As shown, the optional implementation methods of the present application are described in detail below.

[0124] S1, obtaining target data corresponding to the cable joints in the power transmission and transformation system, wherein the target data includes power transmission and transformation operation data, cable joint operation data, and predetermined item data, the predetermined item data is data corresponding to a predetermined item determined in advance, and the predetermined item is an item whose characterization index of the cable joint fault is greater than a predetermined threshold.

[0125] For example, it is possible to obtain the alarm information of the cable joints to be associated and excavated (the same as the above-mentioned alarm data), defect management records and maintenance records (the same as the above-mentioned cable joint operation data), offline test reports (the same as the above-mentioned predetermined item data) and power transmission and transformation online monitoring data (the same as the above-mentioned power transmission and transformation operation data).

[0126] It should be noted that the above data can come from test reports (LMS system), defect records (defect management system), equipment ledger information (control cloud) and local data stored in online detection software, and alarm information (G600 system). In the future, based on multi-source data fusion technology, the alarm information, defect management records and maintenance records, offline test reports and online monitoring data of cable joints will be mined through correlation, and these data will be divided into multiple dimensions to comprehensively evaluate the fault status of cable joints.

[0127] S2, determine the spatiotemporal characteristics corresponding to the power transmission and transformation operation data;

[0128] For example, a CNN-LSTM model can be constructed and trained to extract the spatiotemporal features corresponding to the cable joint power transmission and transformation operation data.

[0129] The power transmission and transformation operation data can often effectively characterize its state changes, such as operating load, overload conditions, current anomalies and temperature rise conditions. An optional implementation method of the present invention adopts a data-driven form in this dimension to establish an association between the power transmission and transformation operation data and the cable joint fault state. By converting the power transmission and transformation operation data into a supervised learning form, as the input layer of the CNN-LSTM model training, the cable state feature values ​​predicted by the model are output through the S-shaped growth curve (sigmoid function). The attention mechanism is introduced into the algorithm model, and the intermediate output results of the LSTM encoder for the input sequence are retained to ensure that the model selectively learns these inputs and the correlation between the output sequence and the input.

[0130] Specifically, S2 also includes:

[0131] S21, inputting the power transmission and transformation operation data into the spatial feature extraction module in the feature detection model to extract the spatial features of the power transmission and transformation operation data, wherein the feature detection model is obtained by training based on the sample data;

[0132] S22, inputting the spatial features into a time feature extraction module in the feature detection model to extract the time features of the power transmission and transformation operation data;

[0133] S23, inputting the spatial features and the temporal features into the feature fusion module in the feature detection model to determine the temporal and spatial features corresponding to the power transmission and transformation operation data.

[0134] The above S21-S23 specifically include:

[0135] For example, the acquired power transmission and transformation operation data can be input into a constructed and trained CNN-LSTM model.

[0136] Among them, the input data of the model include ultrasonic partial discharge, high-frequency partial discharge, sound, cable temperature, oil temperature monitoring data, external meteorological data, cable abnormality inducement characteristics and the mark of the adjacent circuit breaker action or the connected bus voltage over-limit alarm data. The former is a continuous numerical time series data, which is stored in the measured data in the power grid, and the latter mark is a discrete time series data. In order to determine the inducement mark type of discrete alarm data, after counting the cable failure cases, the distribution cable was found to be abnormal by inspection, online monitoring, live detection, power outage test, background alarm, light gas alarm, etc., and effective measures were taken in time to avoid the expansion of the problem, or the defect continued to develop until the failure occurred. After sorting out the inducements of cable abnormality, the inducement mark types are mainly divided into short circuit, aging or wear, moisture or water ingress, overvoltage, and partial discharge. If there is a cable abnormality inducement before the cable failure, it can be marked according to the corresponding serial number of the above type, if not, it is marked as none.

[0137] The CNN-LSTM model is composed of a CNN module and an LSTM module. The spatiotemporal feature values ​​predicted by the model are output through the sigmoid function, and the Attention mechanism is introduced into the model. By retaining the intermediate output results of the LSTM encoder for the input sequence, the CNN-LSTM model is guaranteed to selectively learn these inputs and the correlation between the output sequence and the input.

[0138] Firstly, the power transmission and transformation operation data is input into the CNN module to extract spatial features.

[0139] For the CNN module, four consecutive one-dimensional convolution Conv1D layers are used to extract the spatial features of each sample data. Each Conv1D layer consists of a convolution layer, a batch normalization layer, and a maximum pooling layer. The default input format of the Conv1D layer is three-dimensional data of [batch_size, length, channel], where batch_size is the batch size, batch_size is set to 256, length is the number of sample features, and channel is a feature map in a certain sense. For batch_size samples, the dimension of the data signal input to the model is [160, 1]. The convolution layer of the first Conv1D layer has a convolution kernel of size 9, the convolution layer of the second Conv1D layer has 512 filters and a convolution kernel of size 9, the convolution layer of the third Conv1D layer has 256 filters and a convolution kernel of size 3, and the convolution layer of the fourth Conv1D layer has 128 filters and a convolution kernel of size 3.

[0140] Then the output spatial features are used as the input of the LSTM module to extract the temporal features.

[0141] For the LSTM module, LSTM controls the transmission state through three gate units: input gate, forget gate, and output gate. The input of the next LSTM layer is the time series of the output of the previous LSTM. The output of the last layer of CNN is used as the input of the first layer of LSTM, and the input tensor dimension is [256, 7, 128].

[0142] That is, the input of LSTM must be a three-dimensional array of shape batch size.time steps.unitsl, where batch size is the batch size, time steps is the time step, and units is the number of units in the input sequence. LSTM controls the transmission state through three gated units: input gate, forget gate, and output gate. The input of the next LSTM layer is the time series of the output of the previous LSTM. These gated units are composed of 4 neural networks, and the input of each neural network is composed of the hidden layer input ht-1 at time t-1 and the input at time t. The output of the last layer of CNN is used as the input of the first layer of LSTM. The input tensor dimension is [256, 7, 128]. The first LSTM has 256 hidden units, and the dimension of ht-1 is

[256] . Then the dimension of the LSTM unit parameter matrix is ​​[256+128, 4*256]. At this time, the input at the moment is [256, 128]. Since it will be concatenated with ht-1 to obtain the input matrix [256, 128+256], the input matrix will be multiplied by the parameter matrix {256+128, 4*256], and the output is [256, 4*256]. That is, the output of each neural network in the LSTM unit is [256, 256], so the output dimension of the first LSTM layer is [256, 7, 256]. Similarly, the second LSTM layer has 32 hidden units, so the output dimension is [256, 7, 32].

[0143] Finally, the temporal and spatial features are integrated and analyzed.

[0144] The data flattening layer Flatten in the CNN-LSTM model is used to flatten the output of the LSTM layer, converting the data from multi-dimensional to one-dimensional, and then connected to the fully connected layer to output the maximum eigenvalue, which is the spatiotemporal eigenvalue.

[0145] It should also be noted that the model training method uses categorical cross entropy as the loss function. Use GPU to speed up the model calculation speed and reduce the training time. Select the stochastic gradient descent method with small batches and momentum parameters as the model parameter optimizer. The learning rate update adopts a fixed cycle reduction strategy, and the initial learning rate is set to 0.01. In order to expand the data set and enhance the generalization ability of the model, an optional implementation method of the present invention will use the ten-fold cross method to divide the total number of samples in the data set into 10 parts roughly equally. In each experiment, 7 of them, i.e. 70% of the total number of samples, are selected as training samples, and the rest are used for testing.

[0146] S3: Based on the alarm data, determine the alarm frequency change characteristics, alarm frequency comparison characteristics and alarm duration comparison characteristics corresponding to the cable joint, wherein the alarm frequency comparison characteristics represent the characteristics of the alarm frequencies of various alarm signals respectively occurring in the total alarm frequency, and the alarm duration comparison characteristics represent the characteristics of the duration of the occurrence of various alarm signals respectively occurring in the total duration.

[0147] For example, according to the impact and correlation degree of different alarm information (the same as the above-mentioned alarm signal) on the operating status of the cable joint, the alarm situation can be quantified and the alarm frequency change characteristics, alarm frequency comparison characteristics, and alarm duration comparison characteristics can be extracted.

[0148] In this dimension, the alarm frequency and overload conditions are quantified based on the impact and correlation of different alarm information on the operating status of the cable joint.

[0149] For example, based on the experience of equipment monitoring and operation management, alarm information has obvious seasonal patterns, so the alarm situation is quantified and the alarm frequency change characteristics, alarm frequency comparison characteristics and alarm duration comparison characteristics are extracted. Usually, there are four types of equipment alarm information: accidents, abnormalities, over-limit, and notifications. According to the severity of the four types of alarm information, the degree of its impact on the operation status of the cable joint is measured. For example, if more than 10 alarms are frequently issued in one day, it is considered a false alarm caused by a system problem.

[0150] According to the severity of the four types of alarm information, the degree of influence on the operating status of the cable joint is measured, and the hierarchical analysis method and entropy weight method are used to obtain the weights of the four types of alarm information W = (w1, w2, w3, w4) = (0.43, 0.28, 0.19, 0.03) of the optional implementation mode of the present invention.

[0151] After obtaining multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, and the weights corresponding to the multiple alarm signal categories, it is necessary to determine the alarm frequency change characteristics corresponding to the cable joint, wherein the above determination method can be shown by the following formula:

[0152]

[0153] Among them, M is the alarm frequency change characteristic corresponding to the cable joint, fi is the frequency of occurrence of the i-th alarm information; m is the total number of alarm states, and wi is the weight of the i-th alarm signal.

[0154] After obtaining multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, and the weights corresponding to the multiple alarm signal categories, it is necessary to determine the alarm frequency comparison feature corresponding to the cable joint; wherein the above determination method can be shown by the following formula:

[0155]

[0156] Among them, A1 is the alarm frequency change characteristic corresponding to the cable joint, ni is the number of times the i-th alarm signal occurs in the recent period, and N is the total number of times the i-th alarm signal occurs in the recent period.

[0157] After obtaining multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, the alarm signal durations corresponding to the multiple alarm signal categories, and the weights corresponding to the multiple alarm signal categories, it is necessary to determine the alarm duration comparison feature corresponding to the cable joint; wherein the above determination method can be shown by the following formula:

[0158]

[0159] Among them, A2i is the alarm duration comparison feature corresponding to the cable joint, bi and ci are the normalized indicators of frequency and duration, respectively, and Ti is the duration of the i-th type of frequent alarm signal.

[0160] S4: Determine the operating life characteristics corresponding to the cable joint based on the cable joint operating data.

[0161] For example, in the intermediate step, an equipment aging model based on the operating years can be constructed to extract the equipment degradation factor and the Weibull distribution model fitting results of the operating failure.

[0162] Distribution cables will inevitably age during operation. Therefore, even if cables with different operating years are in the same state, their failure probabilities are different. During the operation of cable joints, the failure rate function caused by operation aging can be fitted by the Weibull distribution model, where the determination method of the operation failure probability value can be shown as follows:

[0163]

[0164] Among them, r(t) is the operational failure probability value, α and β are the shape parameter and scale parameter of the Weibull distribution model respectively, t is the operating life of the equipment, and α=1.24 and β=122.55 are determined based on statistical analysis.

[0165] The aging regression factor a is calculated as follows:

[0166] a=0.4[1-0.005(u1-1)]+0.7[1-0.005(u2-1)]

[0167] Among them, u1 and u2 are degradation indexes respectively. According to the cable joint maintenance data, for major repair records, u1=0.7 and u2=0.5, for minor repair records, u1=0.4 and u2=0.3, and the major repair record is the maintenance that exceeds the repair complexity threshold, and the minor repair is the maintenance that does not exceed the repair complexity threshold.

[0168] S5: Determine the cable defect index corresponding to the cable joint based on the predetermined item data and the cable defect relationship corresponding to the predetermined item.

[0169] For example, the cable degradation degree characteristic corresponding to each test characteristic quantity may be determined based on an offline cable test report.

[0170] According to the analysis of the cable fault mechanism, the test characteristic quantity (same as the above-mentioned predetermined item) is a reliable data to characterize the cable fault, and the qualified standard (attention value) of the cable test characteristic quantity is sorted out. The present invention adopts the membership function based on the upper limit and the lower limit to measure the degree of degradation reflected by the cable test characteristic quantity. If the test characteristic quantity has a specified lower limit, the upper limit attention value specified by the test characteristic quantity is set to a, indicating that the larger the test characteristic quantity detection value V, the greater the tendency of defects. On the contrary, if the test characteristic quantity has a specified lower limit, assuming that the lower limit attention value specified by the test characteristic quantity is b, indicating that the smaller the test characteristic quantity detection value y, the greater the tendency of defects. Therefore, a membership function as shown in the formula is constructed. After the membership function is determined, the degree of cable degradation corresponding to each test characteristic quantity can be determined. The test characteristic quantities selected by the present invention are: winding DC resistance unbalance (%), absorption ratio, winding insulation resistance change rate (%), core and clamp resistance change rate (%), winding dielectric loss (%).

[0171] According to the offline cable test report, determine the cable defect degree characteristics corresponding to each predetermined item. Obtain the threshold type of the predetermined item in this field. Use the upper limit-based and lower limit-based membership functions to measure the defect degree reflected by the cable predetermined items. If the predetermined item has a specified upper limit, let the upper limit attention value specified for the predetermined item be a, indicating that the larger the predetermined item detection value y is, the greater the tendency of defects. Conversely, if the predetermined item has a specified lower limit, assuming that the lower limit attention value specified for the predetermined item is b, indicating that the smaller the predetermined item detection value y is, the greater the tendency of defects. Therefore, construct the membership function shown in the following formula:

[0172]

[0173] Wherein, y is the predetermined item, and U(y) is the degree of defect of the cable caused by the predetermined item.

[0174] After the membership function is determined, the degree of cable degradation corresponding to each predetermined item can be determined. The predetermined items selected in the optional implementation mode of the present invention are: winding DC resistance unbalance (%), absorption ratio, winding insulation resistance change rate (%), core and clamp resistance change rate (%), winding dielectric loss (%).

[0175] S6: Determine weight values ​​of the cable joint in multiple fault states respectively.

[0176] When determining the weight value, a fuzzy comprehensive evaluation method can be used. The fuzzy comprehensive evaluation method is a decision-making method for processing fuzzy information. The following are the steps of the fuzzy comprehensive evaluation method in an optional embodiment of the present invention:

[0177] Determine the evaluation indicators: First, determine the indicators that need to be evaluated. The evaluation indicators of this scheme are the influencing factors of the four fault states.

[0178] Set evaluation level: Set a fuzzy evaluation level for each evaluation indicator.

[0179] Construct a fuzzy evaluation matrix: The performance of the evaluated object on each evaluation index is represented by a fuzzy set, and a fuzzy evaluation matrix is ​​constructed. The rows of the matrix represent the evaluation index, and the columns represent the evaluated object.

[0180] Determine the evaluation weight: Based on the subjective opinions of field engineers or expert knowledge, determine the weight of each evaluation indicator to reflect its importance in the overall evaluation.

[0181] Perform fuzzy calculation: Use fuzzy logic operations to multiply the fuzzy evaluation matrix and the evaluation weight to obtain the fuzzy evaluation value of each evaluated object on each evaluation indicator.

[0182] Fuzzy comprehensive evaluation: For each evaluated object, its fuzzy evaluation value on each evaluation index is integrated to obtain the final fuzzy evaluation result.

[0183] Defuzzification: Convert fuzzy evaluation results into specific values ​​or levels for comparison and ranking.

[0184] Make a decision: According to the defuzzified evaluation results, make corresponding weight values ​​for multiple fault states.

[0185] S7: Build a multi-layer perceptron model and use it to obtain the probability results of the four states.

[0186] The main feature of a multilayer perceptron (MLP) is that it has multiple neuron layers. Its basic structure includes an input layer, a hidden layer, and an output layer. The number of hidden layers can be large or small. The input layer to the hidden layer can be regarded as a fully connected layer, and the hidden layer to the output layer can be regarded as a classification recognizer.

[0187] The perceptron model constructed in the optional implementation mode of the present invention has 17 input layers, 22 intermediate layers, and 2 output layers. The final output is the confidence interval of the cable joint evaluation state. The fuzzy comprehensive evaluation result is a probability model for the expert to preliminarily judge the state result based on the maintenance records and operation data characteristics.

[0188] In an optional implementation manner of the present invention, the status warning results are divided into four categories: normal, general, abnormal, and dangerous, which are output results of a normalized exponential function (softmax function).

[0189] In an optional implementation manner of the present invention, a distribution cable of a distribution network system is selected as an application object for the problem of cable joint fault status assessment.

[0190] Experts preliminarily judged the probability model of the state results based on the maintenance records and operation data characteristics, and obtained the probability results of the four states, which were input together with the extracted feature vectors as the input layer of the MLP model. The MLP model conducted a four-classification experiment and used classification cross entropy as the loss function.

[0191] Figure 3 is a schematic diagram of the composition of a neural network provided in an optional embodiment of the present invention, Figure 4 is a schematic diagram of the relationship between the number of iterations and the accuracy provided by an optional embodiment of the present invention, Figure 5 is a schematic diagram of the relationship between the number of iterations and the loss value provided in an optional embodiment of the present invention. Table 1 is an example table of the model prediction accuracy provided in an optional embodiment of the present invention. Figure 3 , 4, 5 and Table 1, the model training stops when it is 168 times. At this time, the accuracy of the MLP model on the training set tends to 100%, the loss value is almost 0, and the model training effect is optimal. By observing the loss curve, the loss value of the validation set decreases with the decrease of the training set, and finally tends to be stable, and there is no overfitting phenomenon, which proves that the network optimization degree is high. Finally, the final state is determined according to the output confidence interval, and the inspection results are reported to the inspection personnel to complete the early warning task.

[0192] Table 1

[0193]

[0194] S8: Determine a target fault state of the cable joint from the multiple fault states based on the spatiotemporal characteristics, the operating years characteristics, the cable defect index, and the weight values ​​corresponding to the multiple fault states.

[0195] Taking the system's distribution cable fault 5 hours before the occurrence as the base point, a time series data group with a duration of 3 days was extracted and constructed. No abnormal inducement features occurred before the cable fault, and voltage over-limit appeared in the alarm term. The time series data group was resampled to 30 minutes and input into the previously trained feature extraction model. The maximum value of the predicted defect label, 0.75, was taken as the extracted spatiotemporal feature value.

[0196] By traversing the power transmission and transformation defect logs, it was found that there were more than 2,000 cable-related alarm data in this year, including 781 accident-related data.

[0197] The alarm frequency change feature, alarm frequency comparison feature and alarm duration comparison feature vector obtained from the formula are: [0.92, 1.46, 1.42, 1.35, 1.12].

[0198] The cable was put into operation on October 8, 2019. There was no major repair record before this. All minor repairs were carried out during routine inspections. The most recent routine inspection was on December 16, 2019, with a total of 5 minor repair records. The service life characteristics, i.e., the aging return factor and the service failure probability values, were 0.23 and 10.18×10 -4 .

[0199] The electrical test report was collected on June 18, 2020. The test results were qualified, and the cable defect index results were as follows: 0.972, 2, 0, 0, 0.331.

[0200] Finally, based on the opinions of the expert group or relevant on-site engineers and the fuzzy comprehensive evaluation method, the corresponding weight values ​​of multiple fault states are obtained as (0.2, 0.3, 0.3, 0.2).

[0201] The 17 feature vectors are the input to the MLP.

[0202] The output result is 0.67, which is an abnormal state.

[0203] (ii) For the fault status determination system of the cable joint:

[0204] The system mainly consists of five modules, namely online monitoring data module, cable offline ledger module, operation data feature extraction module, offline ledger feature extraction module and multi-layer perceptron module.

[0205] Among them, the online monitoring data module includes time series data such as ambient temperature, ambient humidity, transient partial discharge, ultrasonic partial discharge, high-frequency partial discharge, sound, cable temperature, etc. The data is written into the database after configuring the database through the cloud server and connecting the time series data to the database on the cloud server.

[0206] The cable offline ledger module mainly includes test reports (LMS system), defect records (defect management system), equipment ledger information (control cloud) and alarm information lists.

[0207] By configuring the IP, the running data feature extraction module is connected and communicated with the server. After issuing the instruction, the offline time series data is transmitted to the feature extraction module, and then the spatiotemporal features of the input data are obtained.

[0208] The ledger feature extraction module gradually extracts feature vectors such as alarm information, operating years and experimental reports through the above implementation steps.

[0209] The MLP module takes a multi-dimensional feature vector as input and outputs status warning information.

[0210] Through the above optional implementation, at least the following beneficial effects can be achieved:

[0211] (1) Compared with the related art, the optional implementation mode of the present invention can significantly reduce the computing cost and improve the data processing efficiency. By introducing a deep learning algorithm, such as the CNN-LSTM model, effective dimensionality reduction of data is achieved, which significantly reduces the computing cost. The MLP model is used to efficiently analyze the spatiotemporal characteristics, alarm frequency change characteristics, alarm frequency comparison characteristics, alarm duration comparison characteristics, operating years characteristics, cable defect index, and weight values ​​corresponding to multiple fault states. This process not only greatly reduces the computing component, but also speeds up the data analysis, providing strong support for rapid response to cable joint fault conditions.

[0212] (2) Compared with the related art, the optional implementation method of the present invention optimizes the data processing process and improves the accuracy and reliability of data analysis. The optional implementation method of the present invention divides the massive and discrete cable joint data into multiple dimensions, including spatiotemporal characteristics, alarm frequency change characteristics, alarm frequency comparison characteristics, alarm duration comparison characteristics, operating years characteristics, cable defect index, and weight values ​​corresponding to multiple fault states. This division reduces the complexity of processing massive and discrete cable joint data and improves processing efficiency. By performing multi-dimensional analysis on the data, the fault state of the cable joint can be grasped more comprehensively and accurately. This processing method not only improves the accuracy of the analysis, but also enhances the reliability of the results, providing a strong guarantee for the accurate judgment of the cable fault state.

[0213] (3) Compared with the related art, the present invention avoids the limitation of a single indicator. By systematically combing the monitoring information of multiple types of cable equipment, including but not limited to alarm records, defect records, maintenance logs, test results and online monitoring data of power transmission and transformation, the present invention realizes comprehensive consideration of data in multiple dimensions. On this basis, the present invention extracts spatiotemporal characteristics, alarm frequency change characteristics, alarm frequency comparison characteristics, alarm duration comparison characteristics, operating years characteristics, cable defect index, and weight values ​​corresponding to multiple fault states, which are used to evaluate the fault state of the cable joint. The present invention is combined with the MLP model for analysis, so that the optional implementation method of the present invention can accurately determine the fault state of the cable joint in terms of cable fault state judgment, thereby achieving the technical effect of improving the accuracy of cable joint fault state judgment, and further solving the technical problem of low judgment accuracy when using massive and discrete cable joint related data to judge the cable joint fault state in the related art.

[0214] (4) Compared with the previous state employment method, the present invention abandons the drawback of monitoring a single parameter without considering the correlation between the trend changes of multiple state parameters. While comprehensively considering the nonlinear relationship between multi-dimensional historical and online monitoring data, it extracts the dimensional features of multiple cable joint operation status evaluation by combing the monitoring information of multiple types of cable equipment, covering alarms, defect records, maintenance, testing and online monitoring of power transmission and transformation, etc. On this basis, an MLP perception model is established: by constructing a multi-classification cable state probability output model, it can predict its possible defects 5 hours in advance and achieve good results, determine the cable state warning indicators, and provide sufficient time for monitoring personnel to deploy cable operation and maintenance work in advance. Through experimental analysis, the method and system of the present invention can effectively assist equipment monitoring personnel to grasp the cable operation status in real time and make timely decisions and disposal.

[0215] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0216] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods of various embodiments of the present invention.

[0217] Example 2

[0218] According to an embodiment of the present invention, there is also provided a device for implementing the above-mentioned method for determining the fault state of a cable joint. Figure 6 is a structural block diagram of a device for determining a fault state of a cable connector according to an embodiment of the present invention. Figure 6 As shown, the device includes: an acquisition module 602, a first determination module 604, a second determination module 606, a third determination module 608, a fourth determination module 610 and a fifth determination module 612. The device is described in detail below.

[0219] The acquisition module 602 is used to acquire target data corresponding to the cable joint in the power transmission and transformation system, wherein the target data includes power transmission and transformation operation data, cable joint operation data, and predetermined item data, wherein the predetermined item data is data corresponding to a predetermined item determined in advance, and the predetermined item is an item whose characterization index of the degree of cable joint defect is greater than a predetermined threshold; the first determination module 604 is connected to the acquisition module 602 and is used to determine the spatiotemporal characteristics corresponding to the power transmission and transformation operation data; the second determination module 606 is connected to the first determination module 604 and is used to determine the operating age characteristics corresponding to the cable joint based on the cable joint operation data. ; A third determination module 608 is connected to the second determination module 606, and is used to determine the cable defect index corresponding to the cable joint based on the predetermined item data and the cable defect relationship corresponding to the predetermined item; a fourth determination module 610 is connected to the third determination module 608, and is used to determine the weight values ​​of the cable joint in multiple fault states respectively; a fifth determination module 612 is connected to the fourth determination module 610, and is used to determine the target fault state of the cable joint from multiple fault states based on the spatiotemporal characteristics, the operating years characteristics, the cable defect index, and the weight values ​​corresponding to the multiple fault states respectively.

[0220] It should be noted here that the above-mentioned acquisition module 602, the first determination module 604, the second determination module 606, the third determination module 608, the fourth determination module 610 and the fifth determination module 612 correspond to steps S102 to S112 in the method for determining the fault state of the cable connector, and the instances and application scenarios implemented by the multiple modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above-mentioned embodiment 1.

[0221] Example 3

[0222] According to another aspect of an embodiment of the present invention, there is further provided an electronic device, comprising: a processor; and a memory for storing instructions executable by the processor, wherein the processor is configured to execute instructions to implement any of the above methods for determining a fault state of a cable connector.

[0223] Example 4

[0224] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device can execute any of the above-mentioned methods for determining a fault state of a cable connector.

[0225] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0226] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0227] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0228] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed over multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0229] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0230] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0231] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for determining a fault state of a cable joint, characterized in that: include: Acquire target data corresponding to a cable joint in a power transmission and transformation system, wherein the target data includes power transmission and transformation operation data, cable joint operation data, and predetermined item data, wherein the predetermined item data is data corresponding to a predetermined item determined in advance, and the predetermined item is an item whose characterization index of a cable joint fault is greater than a predetermined threshold; Determining the spatiotemporal characteristics corresponding to the power transmission and transformation operation data; Determining the operating age characteristics corresponding to the cable joint according to the operating data of the cable joint; Determining a cable defect index corresponding to the cable joint according to the predetermined item data and the cable defect relationship corresponding to the predetermined item; Determining weight values ​​of the cable joint being in a plurality of fault states respectively; According to the spatiotemporal characteristics, the operating years characteristics, the cable defect index, and weight values ​​corresponding to a plurality of fault states, a target fault state of the cable joint is determined from the plurality of fault states.

2. The method according to claim 1, characterized in that The method of determining a target fault state of the cable joint from the multiple fault states based on the spatiotemporal characteristics, the operating life characteristics, the cable defect index, and weight values ​​corresponding to the multiple fault states respectively includes: In the case where the target data also includes alarm data, determining the alarm frequency change feature, the alarm frequency comparison feature and the alarm duration comparison feature corresponding to the cable joint according to the alarm data, wherein the alarm frequency comparison feature represents the feature of the alarm frequencies of multiple alarm signals respectively occurring in the total alarm frequencies, and the alarm duration comparison feature represents the feature of the duration of the multiple alarm signals respectively occurring in the total duration; Based on the spatiotemporal characteristics, the alarm frequency change characteristics, the alarm frequency comparison characteristics, the alarm duration comparison characteristics, the operating years characteristics, the cable defect index, and the weight values ​​corresponding to multiple fault states, the target fault state of the cable joint is determined from the multiple fault states.

3. The method according to claim 2, characterized in that In the case where the target data also includes alarm data, determining the alarm frequency change feature, the alarm frequency comparison feature and the alarm duration comparison feature corresponding to the cable joint according to the alarm data, including: Determine, based on the alarm data, multiple alarm signal categories of the cable joint, weights corresponding to the multiple alarm signal categories, alarm frequencies corresponding to the multiple alarm signal categories, and durations of the alarm signals corresponding to the multiple alarm signal categories; Determine the alarm frequency change characteristics corresponding to the cable joint according to the multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, and the weights corresponding to the multiple alarm signal categories; Determine an alarm frequency comparison feature corresponding to the cable joint according to the multiple alarm signal categories, the alarm frequencies respectively corresponding to the multiple alarm signal categories, and the weights respectively corresponding to the multiple alarm signal categories; Based on the multiple alarm signal categories, the alarm frequencies corresponding to the multiple alarm signal categories, the alarm signal durations corresponding to the multiple alarm signal categories, and the weights corresponding to the multiple alarm signal categories, the alarm duration comparison characteristics corresponding to the cable joint are determined.

4. The method according to claim 1, characterized in that: The determining of the spatiotemporal characteristics corresponding to the power transmission and transformation operation data includes: Inputting the power transmission and transformation operation data into a spatial feature extraction module in a feature detection model to extract spatial features of the power transmission and transformation operation data, wherein the feature detection model is trained based on sample data; Inputting the spatial features into a time feature extraction module in a feature detection model to extract the time features of the power transmission and transformation operation data; The spatial features and the temporal features are input into a feature fusion module in a feature detection model to determine the spatiotemporal features corresponding to the power transmission and transformation operation data.

5. The method according to claim 1, characterized in that The determining, based on the cable joint operation data, the operation life characteristics corresponding to the cable joint includes: In the case where the cable joint operation data includes cable joint maintenance data and cable joint operation years, determining an aging regression factor based on the cable joint maintenance data; Determining an operational failure probability value based on the operational life of the cable joint; According to the aging regression factor and the operational failure probability value, an operational life characteristic corresponding to the cable joint is determined.

6. The method according to claim 1, characterized in that Determining the cable defect index corresponding to the cable joint according to the predetermined item data and the cable defect relationship corresponding to the predetermined item includes: Obtaining a threshold type of the predetermined item; Determining a defect index determination function corresponding to the predetermined item according to the threshold type of the predetermined item, wherein the defect index determination function corresponding to the predetermined item represents a cable defect relationship corresponding to the predetermined item; The predetermined item data is input into the defect index determination function to determine the cable defect index corresponding to the cable joint.

7. The method according to any one of claims 1 to 6, characterized in that The determining of the weight values ​​of the cable joints being in a plurality of fault states respectively comprises: Determining a first number of evaluation indicators for evaluating fault state weight values; Determine a plurality of groups of evaluation scores, wherein each group of scores in the plurality of groups of evaluation scores includes a second number of scores, the second number being the product of the first number and the number of the plurality of fault states, the plurality of groups of evaluation scores being obtained by evaluating a plurality of target objects, and the plurality of groups of evaluation scores corresponding to the plurality of target objects one-to-one; constructing an evaluation matrix according to the multiple groups of evaluation scores; According to the evaluation matrix, weight values ​​of the cable joints in a plurality of fault states are determined.

8. A device for determining a fault state of a cable joint, characterized in that: include: An acquisition module is used to acquire target data corresponding to a cable joint in a power transmission and transformation system, wherein the target data includes power transmission and transformation operation data, cable joint operation data, and predetermined item data, wherein the predetermined item data is data corresponding to a predetermined item determined in advance, and the predetermined item is an item whose characterization index of a cable joint fault is greater than a predetermined threshold; A first determination module is used to determine the spatiotemporal characteristics corresponding to the power transmission and transformation operation data; A second determination module is used to determine the operating age characteristics corresponding to the cable joint according to the cable joint operating data; A third determination module is used to determine a cable defect index corresponding to the cable joint according to the predetermined item data and the cable defect relationship corresponding to the predetermined item; A fourth determination module, used to determine weight values ​​of the cable joint being in a plurality of fault states respectively; The fifth determination module is used to determine the target fault state of the cable joint from the multiple fault states based on the spatiotemporal characteristics, the operating years characteristics, the cable defect index, and the weight values ​​corresponding to the multiple fault states.

9. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the method for determining a fault state of a cable joint according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the method for determining a fault state of a cable joint according to any one of claims 1 to 7.

Images