A GIS disconnect switch risk assessment method, device, equipment and storage medium

Through the autoencoder and gradient enhancement decision tree model, the feature extraction and risk assessment of the GIS isolation switch motor power data is solved, and the problem of insufficient early warning capability of potential faults of GIS isolation switch in the prior art is solved, achieving high-precision early warning and Refining the identification of risk levels has significantly improved the safety and stability of the power system.

CN118520379BActive Publication Date: 2024-11-12POWER RES INST OF STATE GRID SHAANXI ELECTRIC POWER CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410985264.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-11-12
Estimated Expiration
2044-07-23

AI Technical Summary

Technical Problem

The existing technology is difficult to effectively warn and evaluate the potential mechanical failure of GIS isolation switches, and lacks the means to prevent and optimize maintenance of future risks, which affects the safe and stable operation of the power system.

Method used

Using the autoencoder and gradient-enhanced decision tree model, a risk assessment method is established through feature extraction and training of GIS isolation switch motor power data to achieve high-precision risk warning and refined risk level identification.

Benefits of technology

It realizes high-precision risk assessment and multi-level risk prediction, can accurately detect potential mechanical failures, improve the risk warning capabilities of GIS isolation switches, and ensure the safe and stable operation of the power system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118520379B_ABST
    Figure CN118520379B_ABST
Patent Text Reader

Abstract

The present invention relates to a GIS disconnector risk assessment method, device, equipment and storage medium, belonging to the field of electrical equipment detection technology, including: collecting switch motor power data for feature extraction to establish a first sample set training autoencoder; establishing a risk level for risk assessment based on the reconstruction error of the autoencoder; constructing a second sample set training gradient boosting decision tree model by inputting samples in the first sample set and the reconstruction error obtained after the samples in the first sample set are input into the autoencoder and the risk level annotated for the samples in the first sample set; enabling the gradient boosting decision tree model to classify the risk level of the samples in the first training sample set; for new samples, after feature extraction, the risk level classification and risk warning of the new samples are performed through the trained autoencoder and gradient boosting decision tree model. The present invention realizes GIS disconnector risk assessment, identifies potential mechanical failures, refines risk levels, and ensures the safety of the power system.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of electrical equipment detection, and in particular to a GIS isolating switch risk assessment method, device, equipment and storage medium. Background Art

[0002] Gas Insulated Switchgear (GIS) GIS disconnectors play a vital role in power systems. They can not only achieve reliable isolation and connection of power lines and equipment, but also effectively improve the stability and safety of power systems. Due to its unique design and superior performance, GIS disconnectors are widely used in high-voltage and ultra-high-voltage power transmission systems, especially in urban substations and underground substations with limited space. GIS disconnectors have advantages that traditional air-insulated switches cannot match.

[0003] In order to ensure the reliable operation of GIS disconnectors, the current mechanical status assessment of GIS disconnectors includes: assessment methods based on vibration signals, current and voltage characteristics, temperature detection, and mechanical stroke and speed; these methods have achieved certain results in practical applications, but they mainly focus on the assessment of the current mechanical status of GIS disconnectors and lack the ability to warn of future risks.

[0004] The safe and stable operation of the power system is crucial to modern society. As a key device in the power system, the failure of GIS disconnectors will not only lead to power supply interruption, but also may cause serious safety accidents. Therefore, in addition to real-time monitoring and evaluation of the mechanical status of GIS disconnectors, it is particularly important to establish effective risk assessment methods that can prevent failures, optimize maintenance and improve safety. Summary of the invention

[0005] In view of the above analysis, the present invention aims to provide a GIS disconnector risk assessment method, device, equipment and storage medium to achieve high-precision GIS disconnector risk assessment, identify potential mechanical failures, refine risk levels, and ensure the safety of the power system.

[0006] The purpose of the present invention is mainly achieved through the following technical solutions:

[0007] One aspect of the present invention discloses a GIS disconnector risk assessment method, comprising:

[0008] Step S1, collecting the motor power data of the GIS disconnector at various stages including starting, acceleration and constant speed operation of the motor, extracting features to establish a first sample set, and performing unsupervised training of the autoencoder; enabling the autoencoder to learn the variation law of the motor power data; and establishing the risk level of the GIS disconnector risk assessment according to the reconstruction error of the autoencoder;

[0009] Step S2, constructing a second sample set by taking the samples in the first sample set, the reconstruction errors obtained after the samples in the first sample set are input into the autoencoder, and the risk levels annotated on the samples in the first sample set, and performing supervised training on the gradient boosting decision tree model; enabling the gradient boosting decision tree model to classify the risk levels of the samples in the first sample set;

[0010] Step S3: extract features from the real-time collected GIS disconnector motor power data to obtain new samples, and perform risk level classification and risk warning on the new samples through the trained autoencoder and gradient boosting decision tree model.

[0011] Another aspect of the present invention discloses a device for implementing the above-mentioned GIS disconnector risk assessment method, comprising: a first module, a second module, a third module and a fourth module;

[0012] The first module is used to collect the motor power data of the GIS disconnector;

[0013] The second module is used to extract the features of the collected GIS disconnector motor power data;

[0014] The third module has a built-in trained autoencoder to calculate the reconstruction error of the output feature data of the second module;

[0015] The fourth module has a built-in trained gradient boosting decision tree model, which is used to perform risk prediction based on the feature data output by the second module and the reconstruction error output by the third module to obtain the risk level of the collected GIS disconnector motor power data.

[0016] Another aspect of the present invention also discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that when the processor executes the computer program, the GIS isolating switch risk assessment method as described above is implemented, and the GIS isolating switch risk prediction is performed to obtain the risk level of the GIS isolating switch.

[0017] Another aspect of the present invention also discloses a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the GIS isolating switch risk assessment method as described above, performs GIS isolating switch risk prediction, and obtains the risk level of the GIS isolating switch.

[0018] The beneficial effects of the present invention are as follows:

[0019] 1. The present invention achieves high-precision risk assessment: combining the advantages of autoencoders and gradient boosting decision trees, it can accurately detect and evaluate the mechanical state of GIS disconnectors and provide high-precision risk warnings.

[0020] 2. The present invention has anomaly detection capability: the autoencoder can identify potential mechanical failures by reconstructing anomalies in error detection data.

[0021] 3. The present invention realizes multi-level risk prediction: by marking and predicting different risk levels, risk warning is made more detailed and specific, making it easier to take corresponding preventive measures.

[0022] The solution of the present invention ensures the safety of the power system and effectively improves the risk warning capability of the GIS disconnector, thereby ensuring the safe and stable operation of the power system and avoiding power outages or equipment damage caused by mechanical failures. The present invention can significantly improve the accuracy and real-time performance of the mechanical status assessment of the GIS disconnector, and has important application value and broad promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings are only used for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. In the entire drawings, the same reference symbols represent the same components;

[0024] Figure 1 is a flow chart of a GIS disconnector risk assessment method in an embodiment of the present invention;

[0025] Figure 2 is a ROC curve diagram of an example in an embodiment of the present invention;

[0026] Figure 3 is an example performance evaluation index diagram in an embodiment of the present invention;

[0027] Figure 4 The present invention is a device composition connection diagram of the GIS disconnect switch risk assessment method in an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used to illustrate the principles of the present invention together with the embodiments of the present invention.

[0029] Embodiment 1

[0030] This embodiment discloses a GIS disconnector risk assessment method; a method for performing GIS disconnector risk assessment using motor power data;

[0031] Motor power data is an important physical quantity that reflects the mechanical state of GIS disconnectors. The operation process of GIS disconnectors is mainly completed by the drive motor, and the running state of the motor directly affects the mechanical action of the GIS disconnector. Therefore, by monitoring and analyzing motor power data, mechanical faults of GIS disconnectors can be effectively identified and diagnosed.

[0032] When the GIS disconnector is opened and closed, the drive motor undergoes a series of dynamic changes, which are reflected in the fluctuation of the motor power. For example, during the closing process, the motor needs to overcome the resistance of the moving and static contacts, which will cause an instantaneous increase in the motor power. If the moving and static contacts are stuck or worn, the change characteristics of the motor power will be different from the normal state. Similarly, during the opening process, the rate and amplitude of the motor power drop can also reflect the mechanical state of the GIS disconnector. By analyzing these power change characteristics, potential mechanical faults such as sticking, wear, and inadequate opening and closing can be identified.

[0033] like Figure 1 As shown, the GIS disconnector risk assessment method provided in this embodiment includes:

[0034] Step S1, collecting the motor power data of the GIS disconnector at various stages including starting, acceleration and constant speed operation of the motor, extracting features to establish a first sample set, and performing unsupervised training of the autoencoder; enabling the autoencoder to learn the variation law of the motor power data; and establishing the risk level of the GIS disconnector risk assessment according to the reconstruction error of the autoencoder;

[0035] Step S2, constructing a second sample set by taking the samples in the first sample set, the reconstruction errors obtained after the samples in the first sample set are input into the autoencoder, and the risk levels annotated on the samples in the first sample set, and performing supervised training on the gradient boosting decision tree model; enabling the gradient boosting decision tree model to classify the risk levels of the samples in the first sample set;

[0036] Step S3: extract features from the real-time collected GIS disconnector motor power data to obtain new samples, and perform risk level classification and risk warning on the new samples through the trained autoencoder and gradient boosting decision tree model.

[0037] Specifically, step S1 includes:

[0038] Step S101, establishing a first sample set by standardizing the collected GIS disconnector motor power data as samples for training the autoencoder;

[0039] The GIS disconnector motor power data comes from the sensor or monitoring device installed on the GIS disconnector, which records the motor power changes under different mechanical conditions; the motor power data feature matrix obtained as a sample after feature extraction ,in, is the number of groups of power data, is the number of features of a set of power data;

[0040] The extracted power data features include the mean, standard deviation, variance, skewness, kurtosis, maximum, minimum, range, root mean square, absolute deviation, energy, and entropy of power.

[0041] Specifically, a set of power data collected is ,in is the length of a set of power data, and the extracted power data features are:

[0042] Mean : ;

[0043] Standard Deviation : ;

[0044] variance : ;

[0045] Skewness : ;

[0046] Kurtosis : ;

[0047] Maximum value: ;

[0048] Minimum: ;

[0049] scope : ;

[0050] RMS : ;

[0051] Absolute Deviation : ;

[0052] energy : ;

[0053] entropy : ;

[0054] Then, the features of a set of power data constitute a column of a sample in the first sample set:

[0055] ;

[0056] By The power data feature matrix of a sample composed of group power data for:

[0057] .

[0058] Step S102: using the first sample set to train the autoencoder to minimize the reconstruction error;

[0059] The autoencoder model includes an encoder and a decoder;

[0060] The encoder is used to compress the input data into a low-dimensional representation z; the high-dimensional motor power data Compressing to a low-dimensional representation z extracts the main features of the data and reduces data redundancy;

[0061] The decoder is used to reconstruct the original input from the low-dimensional representation z Close data , realize data recovery;

[0062] Encoder The output z is:

[0063] ;

[0064] in, is the encoder weight matrix, is the encoder bias vector, is the activation function;

[0065] Optional, activation function Use functions including ReLU functions.

[0066] Decoder Reconstructed data for:

[0067] ;

[0068] in, is the weight matrix of the decoder, is the decoder bias vector, is the activation function;

[0069] Activation Function Functions including the Sigmoid function are used.

[0070] During the training process of the autoencoder model, the objective function to minimize the reconstruction error is:

[0071] ;

[0072] in, , The first sample set samples, is the number of samples in the first sample set.

[0073] By minimizing the error between the original data and the reconstructed data, the autoencoder can learn the main characteristics and changing patterns of the motor power data.

[0074] Step S103: Setting reconstruction error threshold and , the reconstruction error output by the autoencoder is divided into different risk levels;

[0075] For each sample the reconstruction error is calculated:

[0076] ;

[0077] Reconstruction error It represents the difference between the original power data and the reconstructed data, reflecting the autoencoder's ability to compress and recover data. If the reconstruction error is large, it means that the sample data is abnormal, which may be caused by a mechanical failure.

[0078] Setting reconstruction error threshold and , the samples are divided into different risk levels according to the reconstruction error. and Can be selected based on manual experience.

[0079] Assume the risk level is :

[0080] .

[0081] Risk Level Indicates the risk level of the GIS disconnector in the current operating state. Low risk means the mechanical state is normal, medium risk means there may be a minor fault, and high risk means a serious fault requires immediate inspection.

[0082] Step S2 includes:

[0083] Step S201, constructing a second sample set for training a gradient boosting decision tree model;

[0084] In the second sample set, the reconstruction error obtained by inputting the samples in the first sample set and the samples in the first sample set into the autoencoder is used as training data ; The risk level of the samples in the first sample set is used as the labeled data for training ;

[0085] Step S202: construct a gradient boosting decision tree model;

[0086] The gradient boosting decision tree model is:

[0087] ;

[0088] in, Indicates the predicted risk level; For the A decision tree, is the number of decision trees; the gradient boosting decision tree realizes the input data by combining multiple decision tree models. risk level prediction.

[0089] Step S203: training a gradient boosting decision tree model

[0090] During the training process, the gradient boosting decision tree model is trained by minimizing the loss function: the minimized loss function is:

[0091] ;

[0092] in, is the loss function; The second sample set The labeled data of samples; The gradient boosting decision tree model is The risk level predicted by each sample;

[0093] By minimizing the loss function, the model can accurately predict the risk level of the input data and improve the prediction performance of the model.

[0094] In each iteration, the error of the previous iteration is corrected by adding new decision trees:

[0095] ;

[0096] in, For the The output of a decision tree, is the learning rate.

[0097] By gradually adding new decision trees, the prediction error of the model is gradually corrected, thereby improving the accuracy and robustness of the model.

[0098] Specifically, step S3 includes:

[0099] Step S301: Collect GIS disconnector motor power data to extract features and obtain a first new sample ;

[0100] Step S302: The first new sample Input the trained autoencoder to get the first new sample The reconstruction error ;

[0101] Step S303: The first new sample and the corresponding reconstruction error Input constitutes the second new sample ;

[0102] Step S304: The second new sample Input the trained gradient boosting decision tree model for the first new sample Risk prediction, get risk level :

[0103] ;

[0104] It represents the risk level prediction result for the new sample, reflecting the mechanical status and failure risk of the current GIS disconnector.

[0105] Issue risk warnings based on prediction results and promptly notify operation and maintenance personnel to conduct inspections and maintenance.

[0106] This embodiment also provides a specific risk assessment example and performance analysis:

[0107] In this example, a dataset of 1,000 samples is set up, each sample contains 10 features, representing motor power data. We also assume that there are three risk levels: low risk, medium risk, and high risk. The following is the basic structure of the fictitious data:

[0108] Sample size: 1000

[0109] Number of features: 10

[0110] Risk level: Low risk (0), Medium risk (1), High risk (2)

[0111] Risk assessment is performed using the method mentioned in the present invention:

[0112] The results and graphs are generated using fictitious data, where the ROC curve is as follows Figure 2 As shown: Performance evaluation indicators are as follows: Figure 3 As shown;

[0113] Performance evaluation indicators:

[0114] Accuracy: 0.78

[0115] ROC curve and AUC value:

[0116] (ROC curve (Receiver Operating Characteristic Curve) and AUC value (Area Under Curve, the ROC curve represents the performance of the model at different thresholds, and the AUC value is the area under the ROC curve, which indicates the ability of the model to distinguish between positive and negative samples. The closer the AUC value is to 1, the better the model performance.)

[0117] Category 0 (low risk): AUC = 0.85

[0118] Category 1 (intermediate risk): AUC = 0.80

[0119] Category 2 (high risk): AUC = 0.75

[0120] Bar chart of energy evaluation indicators:

[0121] Accuracy: Indicates the accuracy of the model's predictions, which is about 0.78 in the figure.

[0122] Precision: It indicates the proportion of samples that are actually positive among all samples predicted to be positive, which is about 0.79 in the figure.

[0123] Recall: It indicates the proportion of samples that are correctly predicted to be positive among all samples that are actually positive. The figure shows that it is about 0.78.

[0124] F1 Score: It is the harmonic mean of precision and recall, which is about 0.78 as shown in the figure.

[0125] Receiver Operating Characteristic (ROC) Curve:

[0126] Figure 2 Curves 1, 2, and 3 represent the model’s ability to distinguish between three risk levels (low risk, medium risk, and high risk).

[0127] The area under the curve (AUC value) represents the performance of the model. The closer the AUC value is to 1, the better the model performance.

[0128] For example, the AUC value for category 0 (low risk) is approximately 0.85, the AUC value for category 1 (intermediate risk) is approximately 0.80, and the AUC value for category 2 (high risk) is approximately 0.75.

[0129] These graphs and indicators help illustrate the effectiveness and accuracy of the GIS disconnector risk assessment approach based on motor power data.

[0130] In a more preferred solution of this embodiment, when collecting motor power data of the GIS disconnector, the sensor or monitoring device installed on the GIS disconnector adopts a motor power data collection method that eliminates the influence of parasitic circuits; the specific collection method includes:

[0131] 1) For GIS disconnectors that require power collection, connect the current and voltage collection circuit of the GIS disconnector motor; when someone is on duty, collect the current and voltage data of the GIS disconnector motor at all stages of the entire operation process, including starting, acceleration and constant speed operation;

[0132] The purpose of adopting a manned mode is to promptly deal with potential safety hazards caused by parasitic loops of the voltage acquisition circuit during the acquisition process to avoid safety accidents.

[0133] 2) Perform current and voltage phase difference detection during the entire operation process based on the collected current and voltage data of the GIS disconnector motor;

[0134] The detection method may adopt an existing phase difference detection method, such as a phase detection method through an XOR gate.

[0135] 3) According to the detected phase difference data and the collected current data, function fitting is performed to obtain the fitting function ;in, is the effective value of the motor current, is the phase difference between current and voltage, is the fitting function;

[0136] The fitting can be performed by existing fitting methods including second-order fitting, or by other fitting methods. Here, the emphasis is on the use of fitting means rather than a specific fitting technology. Different fitting technologies can achieve the goal of the present invention, and the only difference is the level of accuracy.

[0137] After the second-order fitting, the relationship between the effective value of the current and the phase difference is:

[0138] ;

[0139] In the formula, is the fitting coefficient.

[0140] 4) When collecting power data for the GIS disconnector risk assessment of this embodiment, only the current collection circuit of the GIS disconnector motor is connected. In the case of no one on duty, the current of the GIS disconnector motor is collected at each stage of the whole operation process including starting, acceleration and constant speed operation. The effective value of the current in the cycle is calculated based on the instantaneous value of the sampled current, and the fitting function is used to calculate the current value. Calculate the phase difference between the current and voltage corresponding to the effective value of the current, and then infer the instantaneous voltage value corresponding to the sampled instantaneous current value;

[0141] 5) Calculate the instantaneous sampling power of the GIS disconnector drive motor using the sampled instantaneous current value and the reversed instantaneous voltage value.

[0142] In the formal collection, the parasitic loop problem caused by direct voltage sampling, which may cause the GIS isolation switch to fail to operate normally and bring risks to the site, is avoided; and only the current sampling circuit needs to be connected, and no personnel are required to be on duty, which simplifies the detection process and reduces the detection cost.

[0143] In summary, the method provided in this embodiment has the following beneficial effects:

[0144] 1. Achieved high-precision risk assessment: Combining the advantages of autoencoders and gradient boosting decision trees, it can accurately detect and evaluate the mechanical status of GIS isolation switches and provide high-precision risk warnings.

[0145] 2. The present invention has anomaly detection capability: the autoencoder can identify potential mechanical failures by reconstructing anomalies in error detection data.

[0146] 3. The present invention realizes multi-level risk prediction: by marking and predicting different risk levels, risk warning is made more detailed and specific, making it easier to take corresponding preventive measures.

[0147] The solution of the embodiment of the present invention ensures the safety of the power system and effectively improves the risk warning capability of the GIS disconnector, thereby ensuring the safe and stable operation of the power system and avoiding power outages or equipment damage due to mechanical failures; the present invention can significantly improve the accuracy and real-time performance of the mechanical status assessment of the GIS disconnector, and has important application value and broad promotion prospects.

[0148] Embodiment 2

[0149] Another embodiment of the present invention discloses a device for a GIS disconnector risk assessment method, such as Figure 4 As shown, it includes: a first module, a second module, a third module and a fourth module;

[0150] The first module is used to collect the motor power data of the GIS disconnector;

[0151] The second module is used to extract the features of the collected GIS disconnector motor power data;

[0152] The third module has a built-in trained autoencoder to calculate the reconstruction error of the output feature data of the second module;

[0153] The fourth module has a built-in trained gradient boosting decision tree model, which is used to perform risk prediction based on the feature data output by the second module and the reconstruction error output by the third module to obtain the risk level of the collected GIS disconnector motor power data.

[0154] The specific technical details and beneficial effects of this embodiment are the same as those in Embodiment 1. Please refer to them for details and will not be described in detail here.

[0155] Embodiment 3

[0156] An embodiment of the present invention discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor.

[0157] In one example, the processor may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0158] The memory may include a read-only memory (ROM), a random access memory (RAM), a disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical or other physical / tangible memory storage device. Therefore, typically, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software including computer executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the GIS disconnector risk assessment method according to the first embodiment.

[0159] The processor runs a computer program corresponding to the executable program code by reading the executable program code stored in the memory, so as to implement the GIS disconnector risk assessment method in the first embodiment.

[0160] In one example, a communication interface and a bus may also be included.

[0161] Among them, the memory, processor, and communication interface are connected through a bus and complete communication with each other.

[0162] The communication interface is mainly used to realize the communication between the modules, devices, units and / or equipment in the embodiments of the present application. The input device and / or output device can also be accessed through the communication interface.

[0163] The bus includes hardware, software or both, and the components of the electronic device are coupled to each other. For example, but not limitation, the bus may include an accelerated graphics port (Accelerated Graphics Port, AGP) or other graphics bus, an enhanced industry standard architecture (Enhanced Industry Standard Architecture, EISA) bus, a front side bus (Front Side Bus, FSB), a hypertransport (HyperTransport, HT) interconnection, an industry standard architecture (Industry Standard Architecture, ISA) bus, an infinite bandwidth interconnection, a low pin count (Lowpincount, LPC) bus, a memory bus, a micro channel architecture (MicroChannel Architecture, MCA) bus, a peripheral component interconnect (Peripheral Component Interconnect, PCI) bus, a PCI-Express (PCI-E) bus, a serial advanced technology attachment (Serial Advanced Technology Attachment, SATA) bus, a video electronics standards association local (Video Electronics Standards Association Local Bus, VLB) bus or other suitable buses or a combination of two or more of these. Where appropriate, the bus may include one or more buses. Although embodiments of the present application describe and illustrate a particular bus, the present application contemplates any suitable bus or interconnect.

[0164] Embodiment 4

[0165] One embodiment of the present invention discloses a computer-readable storage medium, on which a computer program instruction is stored. When the computer program instruction is executed by a processor, the GIS disconnector risk assessment method in the first embodiment can be implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. The above-mentioned computer-readable storage medium may include a non-transitory computer-readable storage medium, such as a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc., which is not limited here.

[0166] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A GIS disconnector risk assessment method, characterized in that: include: Step S1, collecting the motor power data of the GIS disconnector at various stages including starting, acceleration and constant speed operation of the motor, extracting features to establish a first sample set, and performing unsupervised training of the autoencoder; enabling the autoencoder to learn the variation law of the motor power data; and establishing the risk level of the GIS disconnector risk assessment according to the reconstruction error of the autoencoder; The GIS disconnector motor power data comes from the sensor or monitoring device installed on the GIS disconnector, which records the motor power changes under different mechanical conditions; the motor power data feature matrix obtained as a sample after feature extraction ,in, is the number of groups of power data, is the number of features of a set of power data; The extracted power data features include the mean, standard deviation, variance, skewness, kurtosis, maximum, minimum, range, root mean square, absolute deviation, energy, and entropy of power; When collecting motor power data of GIS disconnector, the sensor or monitoring equipment installed on the GIS disconnector adopts a motor power data collection method that eliminates the influence of parasitic circuits; The motor power data collection method comprises: 1) For GIS disconnectors that require power collection, connect the current and voltage collection circuit of the GIS disconnector motor; when someone is on duty, collect the current and voltage data of the GIS disconnector motor at all stages of the entire operation process, including starting, acceleration and constant speed operation; 2) Perform current and voltage phase difference detection during the entire operation process based on the collected current and voltage data of the GIS disconnector motor; 3) Based on the detected phase difference data and the collected current data, a second-order fitting is performed to obtain the fitting function ;in, is the effective value of the motor current, is the phase difference between current and voltage, is the fitting function; 4) When collecting power data for GIS disconnector risk assessment, only the current collection circuit of the GIS disconnector motor is connected. Under unattended conditions, the current of the GIS disconnector motor is collected at all stages of the entire operation process, including starting, acceleration and constant speed operation. The effective current value of the cycle is calculated based on the instantaneous value of the sampled current, and the fitting function is used to calculate the current value of the cycle. Calculate the phase difference between the current and voltage corresponding to the effective value of the current, and then infer the instantaneous voltage value corresponding to the sampled instantaneous current value; 5) Calculate the instantaneous sampling power of the GIS isolation switch drive motor using the sampled instantaneous current value and the reversed instantaneous voltage value; Step S2, constructing a second sample set by taking the samples in the first sample set, the reconstruction errors obtained after the samples in the first sample set are input into the autoencoder, and the risk levels annotated on the samples in the first sample set, and performing supervised training on the gradient boosting decision tree model; enabling the gradient boosting decision tree model to classify the risk levels of the samples in the first sample set; Step S3: extract features from the real-time collected GIS disconnector motor power data to obtain new samples, and perform risk level classification and risk warning on the new samples through the trained autoencoder and gradient boosting decision tree model.

2. The GIS disconnector risk assessment method according to claim 1 is characterized in that: The step S1 comprises: Step S101, establishing a first sample set by extracting features from the collected GIS disconnector motor power data and using the data as samples for training the autoencoder; Step S102: using the first sample set to train the autoencoder to minimize the reconstruction error; Step S103: Setting reconstruction error threshold and , the reconstruction error output by the autoencoder is divided into different risk levels.

3. The GIS disconnector risk assessment method according to claim 2 is characterized in that: The autoencoder model includes an encoder and a decoder; The encoder is used to compress the input data into a low-dimensional representation z to reduce data redundancy; Decoder, used to reconstruct the motor power data feature matrix from the low-dimensional representation z Close data , realize data recovery; Encoder The low-dimensional representation z is: ;in, is the encoder weight matrix, is the encoder bias vector, is the activation function; Decoder Reconstructed data for: ;in, is the weight matrix of the decoder, is the decoder bias vector, is the activation function; During the training process of the autoencoder model, the objective function to minimize the reconstruction error is: ; in, , No. Motor power data feature matrix The first sample set samples, is the number of samples in the first sample set.

4. The GIS disconnector risk assessment method according to claim 3 is characterized in that: The risk level of GIS isolation switch risk assessment is established based on the reconstruction error of the autoencoder. Motor power data feature matrix Risk level for: ; in, and is the reconstruction error threshold; reconstruction error ; .

5. The GIS disconnector risk assessment method according to claim 4 is characterized in that: The step S2 comprises: Step S201, constructing a second sample set for training a gradient boosting decision tree model; In the second sample set, the reconstruction error obtained by inputting the samples in the first sample set and the samples in the first sample set into the autoencoder is used as training data ; The risk level of the samples in the first sample set is used as the labeled data for training ; Step S202: construct a gradient boosting decision tree model; The gradient boosting decision tree model is: ; in, Indicates the predicted risk level; For the A decision tree, is the number of decision trees; Step S203, training a gradient boosting decision tree model; During the training process, the gradient boosting decision tree model is trained by minimizing the loss function: the minimized loss function is: ; in, is the loss function; The second sample set The labeled data of samples; The gradient boosting decision tree model is The risk level predicted by the sample.

6. The GIS disconnector risk assessment method according to claim 5 is characterized in that: Step S3 includes: Step S301: Collect GIS disconnector motor power data to extract features and obtain a first new sample ; Step S302: The first new sample Input the trained autoencoder to get the first new sample The reconstruction error ; Step S303: The first new sample and the corresponding reconstruction error Input constitutes the second new sample ; Step S304: The second new sample Input the trained gradient boosting decision tree model for the first new sample Risk prediction, get risk level ; It reflects the mechanical status and failure risk of the current GIS disconnector.

7. A device for implementing the GIS disconnector risk assessment method according to any one of claims 1 to 6, characterized in that: include: Module 1, Module 2, Module 3 and Module 4; The first module is used to collect the motor power data of the GIS disconnector; The second module is used to extract the features of the collected GIS disconnector motor power data; The third module has a built-in trained autoencoder to calculate the reconstruction error of the output feature data of the second module; The fourth module has a built-in trained gradient boosting decision tree model, which is used to perform risk prediction based on the feature data output by the second module and the reconstruction error output by the third module to obtain the risk level of the collected GIS disconnector motor power data.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the GIS isolating switch risk assessment method according to any one of claims 1 to 6 is implemented, and the GIS isolating switch risk prediction is performed to obtain the risk level of the GIS isolating switch.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the GIS isolating switch risk assessment method according to any one of claims 1 to 6 is implemented, and the GIS isolating switch risk prediction is performed to obtain the risk level of the GIS isolating switch.