High-voltage switch cabinet fault early warning method and system, computer equipment or storage medium

Abstract

The invention belongs to the technical field of switch cabinet monitoring, and particularly relates to a high-voltage switch cabinet fault early warning method and system, computer equipment or a storage medium. The early warning method comprises the following steps: acquiring an ultrasonic signal, a transient ground voltage signal, an ultrahigh frequency detection signal, ozone concentration, oxygen concentration, an SF6 signal, a temperature signal and a humidity signal; performing data preprocessing on the acquired signal to obtain a processed signal; and inputting the ultrasonic signal and the preprocessed signal into the BP neural network model which is qualified in training, and outputting a prediction result. The operation state of the high-voltage switch cabinet can be monitored from multiple angles, possible limitation of single parameter detection is avoided, and the BP neural network has strong nonlinear mapping capability and self-learning capability, so that whether the high-voltage switch cabinet has a fault or not and possible types of the fault and the like can be accurately predicted; and potential fault hidden dangers can be found in advance.

Description

Technical Field

[0001] The present invention belongs to the technical field of switch cabinet monitoring, and in particular relates to a high-voltage switch cabinet fault early warning method, system, computer equipment or storage medium. Background Art

[0002] As an important equipment in the power system, the switchgear has an important impact on the operation of the entire system. Insulation degradation is inevitable during its operation. Insulation degradation will lead to partial discharge. If it is allowed to develop, it will cause serious damage to the insulation and even lose dielectric properties. Many switchgear burnouts and explosions are caused by this process. Because the number of switchgears is large and widely distributed, only a small number of them are equipped with online monitoring systems. For most switchgears, they are repaired in a power outage state, but the power outage maintenance time interval is long, and it is difficult to detect partial discharge faults in time. In the daily operation process, operation and maintenance personnel will also use partial discharge detection equipment to detect partial discharge signals outside the switchgear. However, due to the obstruction of the switchgear cabinet, the detection sensitivity and accuracy are not enough, and there is a detection cycle, it is difficult to detect partial discharge faults in time. If the partial discharge characteristics of electrical equipment in the accident latency period can be monitored in real time, whether partial discharge occurs and the type of discharge can be automatically determined, the discharge point can be located, and it can be processed in time, it can effectively prevent the failure of high-voltage switchgear electrical equipment and take precautions before it happens. There have been a lot of studies on the mechanism, phenomenon and detection methods of partial discharge. Based on the phenomena of sound, light, electricity, magnetism and chemical substances produced when partial discharge occurs, the methods currently used for partial discharge detection in switch cabinets mainly include ultrasonic method, transient voltage to ground method, pulse current method, ultra-high frequency method, etc. There are the following problems: Ultrasonic signals are easily interfered by the internal structure of the switch cabinet and the external environmental noise. There are various complex electrical equipment and mechanical structures inside the switch cabinet. Ultrasonic waves will be reflected, refracted and scattered during the propagation process, resulting in signal attenuation and distortion. For example, when there is vibration noise from other electrical equipment or external mechanical vibration around the switch cabinet, it will interfere with the ultrasonic detection signal and reduce the accuracy and reliability of the detection. The sensitivity of ultrasonic detection is limited, and it may not be able to effectively detect tiny partial discharge signals. Especially in the early stage of partial discharge, the discharge amount is small, and the ultrasonic signal generated is weak, which can easily be submerged by background noise, making it difficult to accurately determine whether partial discharge has occurred.

[0003] Transient ground voltage signals are susceptible to external electromagnetic interference. In complex electrical environments such as substations, there are a large number of electromagnetic devices and transmission lines. The electromagnetic radiation generated by these devices and lines will interfere with the detection of transient ground voltage signals. For example, when a large transformer is switched nearby or there is lightning activity, a large amount of electromagnetic interference will be generated, causing deviations in the transient ground voltage detection results. The positioning accuracy of this method for partial discharge positions is relatively low. When transient ground voltage signals propagate in the switch cabinet, they are affected by many factors, such as the characteristics of the insulating medium, the geometric structure of the switch cabinet, etc., making it difficult to accurately determine the location of partial discharge based on transient ground voltage signals.

[0004] The pulse current method needs to be connected to the electrical circuit, which will interfere with the normal operation of the equipment. When installing the detection equipment, the detection circuit needs to be connected to the electrical circuit of the switch cabinet, which may change the parameters of the original electrical circuit and create potential risks to the normal operation of the equipment. At the same time, this connection method also increases the complexity and installation cost of the equipment. The detection bandwidth is limited, and the detection capability of high-frequency partial discharge signals is insufficient. With the continuous improvement of the voltage level of power equipment and the development of technology, the frequency range of partial discharge signals is also expanding, and the detection performance of the pulse current method in the high frequency band is relatively weak, which cannot meet the demand for high-frequency partial discharge signal detection.

[0005] UHF signals are severely attenuated when propagating in the switch cabinet. The cabinet structure and internal insulation materials of the switch cabinet have strong absorption and attenuation effects on UHF signals, resulting in large energy loss of UHF signals during propagation. This results in a limited detection range, and partial discharge signals far away from the sensor may not be effectively received. The cost of UHF detection equipment is high. Since UHF detection technology is relatively new, its detection equipment involves high-frequency electronic devices and complex signal processing systems, and the R&D and production costs are high, which to a certain extent limits its wide application.

[0006] In summary, existing methods are unable to accurately warn of faults in high-voltage switchgear. Summary of the invention

[0007] The purpose of the present invention is to provide a high-voltage switch cabinet fault early warning method, system, computer equipment or storage medium, which solves the problem that the prior art cannot accurately predict the high-voltage switch cabinet fault.

[0008] The present invention is achieved through the following technical solutions: The present invention discloses a high-voltage switch cabinet fault early warning method, comprising the following steps: S1. Obtain ultrasonic signals, transient ground voltage signals, UHF detection signals, ozone concentration, oxygen concentration, SF6 signals, temperature signals and humidity signals; S2, performing data preprocessing on transient ground voltage signals, ultra-high frequency detection signals, ozone concentration, SF6 signals, oxygen concentration, temperature signals and humidity signals to obtain processed signals; S3, input the ultrasonic signal, ozone concentration, oxygen concentration, temperature signal, SF6 signal, humidity signal, UHF detection signal and transient ground voltage signal pre-processed by S2 into the trained BP neural network model, and output the prediction result.

[0009] Furthermore, S1 is specifically: Ultra-high frequency sensors, ultrasonic sensors, and transient ground voltage sensors are arranged on the surface of the switch cabinet; Use multiple ultrasonic sensors to collect ultrasonic signals from the switch cabinet; A single transient ground voltage sensor is used to collect the transient ground voltage signal of the switch cabinet; A single UHF sensor is used to collect the UHF detection signal of the switch cabinet; Install ozone concentration sensors, oxygen concentration sensors, temperature sensors, humidity sensors, and SF6 sensors in the switch cabinet; Ozone concentration sensor: An ozone concentration sensor is used to collect the ozone concentration of the switch cabinet; Use oxygen concentration sensor to collect oxygen concentration of switch cabinet; Use temperature sensors to collect temperature signals from the switch cabinet; Use humidity sensors to collect humidity signals from switch cabinets; Ultrasonic signals, ozone concentration, oxygen concentration, SF6 signals, UHF detection signals, temperature signals, humidity signals and transient ground voltage signals are transmitted to the host computer via Lora wireless.

[0010] Further, in S2, the transient voltage signal includes a voltage amplitude signal and a pulse number, with the voltage amplitude signal being primary and the pulse number being secondary; The preprocessing of the transient voltage signal is specifically as follows: obtaining a low-frequency voltage signal through low-pass filtering, and obtaining a pulse signal through high-pass filtering; Data preprocessing is performed on the UHF detection signal, SF6 signal, ozone concentration, oxygen concentration, temperature signal and humidity signal. Specifically, the sliding average filtering method is used to remove the fluctuation and noise of the six signals.

[0011] Furthermore, in S3, the training process of the BP neural network model is: S3.1, obtain sensor historical data, pre-process it, and input it into the BP neural network model; The sensor historical data includes ultrasonic signals, low-frequency voltage signals, pulse signals, UHF detection signals, ozone concentration, oxygen concentration, SF6 signals, temperature signals and humidity signals; The input layer of the BP neural network model inputs the above 9 sensor signals, which are , , , , , , , , ; The first-level PD warning weight coefficients corresponding to the 9 sensor signals are: , , , , , , , The first-level environmental warning weight coefficients are , , , , , , , , ; S3.2. Calculate the partial discharge warning value and environmental warning value based on the weight allocation coefficient; = +

[0012] = +

[0013] in, is the partial discharge warning value, is the environmental warning value, is the partial discharge warning threshold, is the environmental warning threshold; S3.3. Initial second-level weight coefficients , , based on the partial discharge warning value and the environmental warning value, the fault warning value z is obtained; the expression is:

[0014] represents the second-level partial discharge warning weight coefficient, Represents the second-level environmental warning weight coefficient; Represents the activation function of the warning; S3.4. Determine the status of the high-voltage switchgear according to the fault warning value z; When 0 < z < 0.35, the status of the high-voltage switchgear is normal; when 0.35 ≤ z ≤ 0.65, the status of the high-voltage switchgear is a warning; when 0.65 < z, the high-voltage switchgear is abnormal; According to and Judge the status of the high-voltage switchgear; when > 0.8 and > 0.8, the output of the high-voltage switchgear is abnormal; when > 0.8 or > 0.8, the high-voltage switchgear outputs a warning; otherwise it is normal; S3.5. If the results of the two judgments are inconsistent, calculate the fault warning value error , and the expression is:

[0015] Among them, is the actual expected value; Judge whether the fault warning value error is within the preset requirements. If it is, judge whether the first-level weight coefficient needs to be corrected; If not, adjust the second-level weight coefficient until the fault warning value error is within the preset requirements.

[0016] Furthermore, the adjustment method of the second-level weight coefficient is as follows: Calculate the corrected fault warning value , and the expression is:

[0017] The first correction error ; Among them, is the second-level weight error coefficient, j = 1, 2;

[0018] Among them, is the error backpropagation learning rate, and the parameter affects its iteration speed; Repeat S3.4 - S3.5, continue to judge whether the correction error is within the preset requirements. If not, continue to correct; Similarly, the calculation result of the nth correction is:

[0019] The nth correction error ; Compare with the preset requirements each time until the correction error Within the preset requirements, the optimal secondary weight coefficient is finally determined. , .

[0020] Further, it is determined whether the first-level weight coefficient needs to be corrected, specifically: First calculate the error of the partial discharge warning value , the expression is:

[0021] in, is the expected value of partial discharge warning, is the partial discharge warning value; Judgment Error Whether the error requirements are met. If so, it means that the first-level weight coefficient is appropriate; If it is not satisfied, the first-level weight coefficient is modified as follows: The nth corrected PD warning value ,for: = +

[0022] in is the nth first-order weighted partial discharge correction coefficient, and its value is:

[0023] Error after the nth correction , specifically:

[0024] when When the value of meets the error requirement, the optimal first-level partial discharge warning weight coefficient and the optimal first-level environmental warning weight coefficient are obtained.

[0025] Furthermore, in S3.3, the activation function of the early warning is the Sigmod function, which is: .

[0026] The present invention also discloses a high-voltage switch cabinet fault early warning system, comprising: Data acquisition module, used to acquire ultrasonic signals, transient ground voltage signals, UHF detection signals, ozone concentration, oxygen concentration, SF6 signals, temperature signals and humidity signals; A data preprocessing module is used to perform data preprocessing on transient ground voltage signals, ultra-high frequency detection signals, ozone concentration, SF6 signals, oxygen concentration, temperature signals and humidity signals to obtain processed signals; The prediction module is used to input the ultrasonic signal, the pre-processed ozone concentration, the oxygen concentration, the temperature signal, the SF6 signal, the humidity signal, the UHF detection signal and the transient ground voltage signal into the trained BP neural network model and output the prediction result.

[0027] The present invention also discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the high-voltage switch cabinet fault warning method when executing the computer program.

[0028] The present invention also discloses a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the high-voltage switch cabinet fault warning method are implemented.

[0029] Compared with the prior art, the present invention has the following beneficial technical effects: The purpose of the present invention is to provide a high-voltage switchgear fault warning method, which can monitor the operating status of the high-voltage switchgear from multiple angles by acquiring various types of signals such as ultrasonic signals, transient ground voltage signals, ultra-high frequency detection signals, ozone concentration, oxygen concentration, SF6 signals, temperature signals and humidity signals, avoiding the limitations that may exist in single parameter detection, and more comprehensively and accurately grasping the actual situation of the equipment, thereby improving the reliability and accuracy of fault warning. The pre-processed multiple signals are input into the trained qualified BP neural network model to output the prediction results. The BP neural network has a strong nonlinear mapping ability and self-learning ability. It can establish a complex mapping relationship between the input signal and the fault state through learning and training of a large amount of known data, thereby making a more accurate prediction of whether the high-voltage switchgear has a fault and the possible type of the fault, etc., which is helpful to discover potential fault hazards in advance, take timely measures for maintenance, and ensure the safe and stable operation of the high-voltage switchgear.

[0030] Furthermore, for transient ground voltage signals, a preprocessing method of obtaining low-frequency voltage signals through low-pass filtering and pulse signals through high-pass filtering can effectively separate signal components with different frequency characteristics, which is helpful for subsequent more accurate analysis of fault-related characteristic information contained in transient ground voltage signals and improves the effectiveness of using the signal; The sliding average filtering method is used to remove fluctuations and noise from UHF detection signals, SF6 signals, ozone concentration, oxygen concentration, temperature signals and humidity signals, which can make these signals smoother and more stable, reduce the possibility of misjudgment due to signal fluctuations and noise interference, and make the data input into the model higher in quality, thereby improving the accuracy of fault prediction based on these data.

[0031] Furthermore, historical sensor data, including nine types of signals such as ultrasonic signals and low-frequency voltage signals, are used in the training process, and partial discharge warning weight coefficients and environmental warning weight coefficients are set for these signals respectively. This can fully consider the importance differences of different signals in fault warning and environmental monitoring, so that the model can reasonably allocate weights during the learning process and better mine the effective information in the data.

[0032] When the two judgment results are inconsistent, the fault warning value error is calculated to determine whether the secondary weight coefficient needs to be adjusted. This weight adjustment method based on error back propagation (such as calculating and correcting the fault warning value and correcting the error) allows the model to continuously optimize its own parameters to adapt to the actual data distribution and failure mode. Through multiple iterative corrections until the corrected error is within the preset requirements, the optimal secondary weight coefficient is finally determined, allowing the model to warn of faults more accurately.

[0033] Not only will the secondary weight coefficient be adjusted, but it will also determine whether the primary weight coefficient needs to be corrected. By calculating the error of the partial discharge warning value, the primary weight coefficient is similarly corrected multiple times until the error requirement is met, and the optimal first-level partial discharge warning weight coefficient and the optimal first-level environmental warning weight coefficient are obtained. In this way, the model can be fully optimized from the perspective of the signal weight of the input layer, improving the model's sensitivity and discrimination ability to different signals, thereby improving the accuracy and reliability of the entire fault warning system.

[0034] Calculate the partial discharge warning value and environmental warning value, and comprehensively consider these two factors in the subsequent calculation of the fault warning value and the fault judgment process. This method can simultaneously consider the impact of partial discharge and environmental factors on the high-voltage switchgear, avoiding the problem of focusing on a single factor and ignoring other potential fault factors. It is more in line with the complex situation of equipment failure in the actual operating environment, thereby providing a more comprehensive and effective fault warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a schematic diagram of the fault warning principle of the high-voltage switch cabinet of the present invention; Figure 2 It is a flow chart of a high-voltage switch cabinet fault early warning method of the present invention; Figure 3 The present invention is a block diagram of a high-voltage switch cabinet fault warning system. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solution and advantages of the present invention more clear, the following is further described in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the embodiments described are only part of the embodiments of the present invention, not all embodiments.

[0037] The components described and shown in the drawings and embodiments of the present invention may be arranged and designed in various configurations. Therefore, the detailed description of the embodiments of the present invention provided in the following drawings is not intended to limit the scope of the claimed invention, but merely represents a selected embodiment of the present invention. Based on the drawings and embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0038] It should be noted that the terms "comprises", "includes" or any other variations are intended to cover non-exclusive inclusion, so that a process, element, method, article or equipment that includes a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to the process, element, method, article or equipment.

[0039] The features and performances of the present invention are further described in detail below in conjunction with the embodiments.

[0040] Example 1 like Figure 1 and Figure 2 As shown, the present invention discloses a high-voltage switch cabinet fault early warning method, comprising the following steps: S1. Obtain ultrasonic signals, transient ground voltage signals, UHF detection signals, ozone concentration, oxygen concentration, SF6 signals, temperature signals and humidity signals; S2, performing data preprocessing on transient ground voltage signals, ultra-high frequency detection signals, ozone concentration, SF6 signals, oxygen concentration, temperature signals and humidity signals to obtain processed signals; S3, input the ultrasonic signal, ozone concentration, oxygen concentration, temperature signal, SF6 signal, humidity signal, UHF detection signal and transient ground voltage signal processed by S2 into the trained BP neural network model, and output the prediction result.

[0041] The preprocessing of the transient voltage signal is specifically as follows: obtaining a low-frequency voltage signal through low-pass filtering, and obtaining a pulse signal through high-pass filtering; Data preprocessing is performed on the UHF detection signal, SF6 signal, ozone concentration, oxygen concentration, temperature signal and humidity signal. Specifically, the sliding average filtering method is used to remove the fluctuation and noise of the six signals.

[0042] Example 2 S1 is specifically: Ultra-high frequency sensors, ultrasonic sensors, and transient ground voltage sensors are arranged on the surface of the switch cabinet; Use multiple ultrasonic sensors to collect ultrasonic signals from the switch cabinet; A single transient ground voltage sensor is used to collect the transient ground voltage signal of the switch cabinet; A single UHF sensor is used to collect the UHF detection signal of the switch cabinet; Install ozone concentration sensors, oxygen concentration sensors, temperature sensors, humidity sensors, and SF6 sensors in the switch cabinet; Ozone concentration sensor: An ozone concentration sensor is used to collect the ozone concentration of the switch cabinet; Use oxygen concentration sensor to collect oxygen concentration of switch cabinet; Use temperature sensors to collect temperature signals from the switch cabinet; Use humidity sensors to collect humidity signals from switch cabinets; Ultrasonic signals, ozone concentration, oxygen concentration, SF6 signals, UHF detection signals, temperature signals, humidity signals and transient ground voltage signals are transmitted to the host computer via Lora wireless.

[0043] Example 3 This paper mainly introduces the training process of BP neural network model.

[0044] The training process of the BP neural network model is: S3.1, obtain sensor historical data, pre-process it, and input it into the BP neural network model; The sensor historical data includes ultrasonic signals, low-frequency voltage signals, pulse signals, UHF detection signals, ozone concentration, oxygen concentration, SF6 signals, temperature signals and humidity signals; The input layer of the BP neural network model inputs the above 9 sensor signals, which are , , , , , , , , ; The first-level PD warning weight coefficients corresponding to the 9 sensor signals are: , , , , , , , The first-level environmental warning weight coefficients are , , , , , , , , ; S3.2. Calculate the partial discharge warning value and the environmental warning value based on the weight distribution coefficient; = +

[0045] = +

[0046] where, is the partial discharge warning value, is the environmental warning value, is the partial discharge warning threshold, is the environmental warning threshold; S3.3. The initial second-level weight coefficient , , obtain the fault warning value z based on the partial discharge warning value and the environmental warning value; the expression is:

[0047] represents the second-level partial discharge warning weight coefficient, represents the second-level environmental warning weight coefficient; represents the activation function of the warning; The activation function selects the Sigmod function, specifically: .

[0048] S3.4. Judge the state of the high-voltage switchgear according to the fault warning value z; When 0 < z < 0.35, the state of the high-voltage switchgear is normal. When 0.35 ≤ z ≤ 0.65, the state of the high-voltage switchgear is a warning. When 0.65 < z, the high-voltage switchgear is abnormal; According to and judge the state of the high-voltage switchgear; when > 0.8 and > 0.8, the output of the high-voltage switchgear is abnormal; when > 0.8 or > 0.8, the high-voltage switchgear outputs a warning; otherwise it is normal; S3.5. If the results of the two judgments are inconsistent, calculate the fault warning value error , the expression is:

[0049] where, is the actual expected value; Determine whether the fault warning value error is within the preset requirement range. If so, start S3.7; If not, start S3.6; S3.6. Adjust the secondary weight coefficient. The adjustment method is as follows: Calculate and correct fault warning value , the expression is:

[0050] First error correction ; in, is the secondary weight error coefficient, j=1,2;

[0051] in, is the error reverse learning rate, the parameter affects its iteration speed; Repeat S3.4-S3.5 to continue to determine whether the correction error is within the preset requirement range. If not, continue correction; Similarly, the calculation result of the nth repair is:

[0052] The nth correction error ; Compare with the preset requirement range each time until the error is corrected Within the preset requirements, the optimal secondary weight coefficient is finally determined. , ; S3.7. Determine whether the primary weight coefficient needs to be corrected, specifically: First calculate the error of the partial discharge warning value , the expression is:

[0053] in, is the expected value of partial discharge warning, is the partial discharge warning value; Judgment Error Whether the error requirements are met. If so, it means that the first-level weight coefficient is appropriate; If it is not satisfied, the first-level weight coefficient is modified as follows: The nth corrected PD warning value ,for: = +

[0054] in is the nth first-order weighted partial discharge correction coefficient, and its value is:

[0055] Error after the nth correction , specifically:

[0056] when When the value of meets the error requirement, the optimal first-level partial discharge warning weight coefficient and the optimal first-level environmental warning weight coefficient are obtained.

[0057] Example 4 like Figure 3 As shown, the present invention also discloses a high-voltage switch cabinet fault warning system, comprising: Data acquisition module, used to acquire ultrasonic signals, transient ground voltage signals, UHF detection signals, ozone concentration, oxygen concentration, SF6 signals, temperature signals and humidity signals; A data preprocessing module is used to perform data preprocessing on transient ground voltage signals, ultra-high frequency detection signals, ozone concentration, SF6 signals, oxygen concentration, temperature signals and humidity signals to obtain processed signals; The prediction module is used to input the ultrasonic signal, the pre-processed ozone concentration, the oxygen concentration, the temperature signal, the SF6 signal, the humidity signal, the UHF detection signal and the transient ground voltage signal into the trained BP neural network model and output the prediction result.

[0058] Example 5 The present invention also discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the high-voltage switch cabinet fault warning method are implemented. The memory may include a memory, such as a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk memory, etc. The processor, the network interface, and the memory are interconnected through an internal bus. The internal bus may be an industrial standard architecture bus, a peripheral component interconnection standard bus, an extended industrial standard architecture bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory is used to store programs. Specifically, the program may include a program code, and the program code includes computer operation instructions. The memory may include a memory and a non-volatile memory, and provide instructions and data to the processor.

[0059] Example 6 The present invention also discloses a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the high-voltage switch cabinet fault warning method are implemented. Specifically, the computer-readable storage medium includes, but is not limited to, for example, a volatile memory and / or a non-volatile memory. The volatile memory may include a random access memory and / or a cache memory, etc. The non-volatile memory may include a read-only memory, a hard disk, a flash memory, an optical disk, a magnetic disk, etc.

[0060] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer-usable program code.

[0061] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0062] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0063] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A high-voltage switch cabinet fault early warning method, characterized in that: It includes the following steps: S1. Obtain ultrasonic signals, transient earth voltage signals, UHF detection signals, ozone concentration, oxygen concentration, SF6 signals, temperature signals, and humidity signals; S2. Perform data preprocessing on the transient earth voltage signals, UHF detection signals, ozone concentration, SF6 signals, oxygen concentration, temperature signals, and humidity signals to obtain processed signals; S3. Input the ultrasonic signals, the ozone concentration, oxygen concentration, temperature signals, SF6 signals, humidity signals, UHF detection signals, and transient earth voltage signals after preprocessing in S2 into a trained BP neural network model and output a prediction result.

2. A high-voltage switch cabinet fault early warning method according to claim 1, characterized in that: Specifically, S1 is as follows: Arrange UHF sensors, ultrasonic sensors, and transient earth voltage sensors on the surface of the switch cabinet body; Use multiple ultrasonic sensors to collect ultrasonic signals of the switch cabinet; Use a single transient earth voltage sensor to collect transient earth voltage signals of the switch cabinet; Use a single UHF sensor to collect UHF detection signals of the switch cabinet; Arrange ozone concentration sensors, oxygen concentration sensors, temperature sensors, humidity sensors, and SF6 sensors inside the switch cabinet; The ozone concentration sensor is used to collect the ozone concentration of the switch cabinet; Use an oxygen concentration sensor to collect the oxygen concentration of the switch cabinet; Use a temperature sensor to collect the temperature signal of the switch cabinet; Use a humidity sensor to collect the humidity signal of the switch cabinet; The ultrasonic signals, ozone concentration, oxygen concentration, SF6 signals, UHF detection signals, temperature signals, humidity signals, and transient earth voltage signals are wirelessly transmitted to the upper computer via Lora.

3. A high-voltage switch cabinet fault early warning method according to claim 1, characterized in that: In S2, the transient earth voltage signal includes a voltage amplitude signal and a pulse number, with the voltage amplitude signal being the main one and the pulse number being the auxiliary one; The specific preprocessing of the transient earth voltage signal is: obtain a low-frequency voltage signal through low-pass filtering and obtain a pulse signal through high-pass filtering; The data preprocessing of the UHF detection signals, SF6 signals, ozone concentration, oxygen concentration, temperature signals, and humidity signals is specifically: use a moving average filtering method to remove the fluctuations and noises of the six signals.

4. A high-voltage switch cabinet fault early warning method according to claim 3, characterized in that: In S3, the training process of the BP neural network model is: S3.

1. Obtain historical sensor data, preprocess it, and input it into the BP neural network model; The historical sensor data includes ultrasonic signals, low-frequency voltage signals, pulse signals, UHF detection signals, ozone concentration, oxygen concentration, SF6 signals, temperature signals, and humidity signals; The input layer of the BP neural network model inputs the above 9 sensor signals, which are , , , , , , , , ; The first-level PD warning weight coefficients corresponding to the 9 sensor signals are: , , , , , , , The first-level environmental warning weight coefficients are , , , , , , , , ; S3.

2. Calculate the partial discharge warning value and the environmental warning value based on the weight distribution coefficient; = + = + in, is the partial discharge warning value, is the environmental warning value, is the partial discharge warning threshold, is the environmental warning threshold; S3.

3. Initial second-level weight coefficients , , based on the partial discharge warning value and the environmental warning value, the fault warning value z is obtained; the expression is: represents the second-level partial discharge warning weight coefficient, Represents the second-level environmental warning weight coefficient; represents the activation function of the warning; S3.

4. Judge the state of the high-voltage switch cabinet according to the fault warning value z; When 0 < z < 0.35, the state of the high-voltage switch cabinet is normal; when 0.35 ≤ z ≤ 0.65, the state of the high-voltage switch cabinet is a warning; when 0.65 < z, the high-voltage switch cabinet is abnormal; according to and Determine the status of the high-voltage switchgear; when >0.8 and >0.8, the output of the high-voltage switch cabinet is abnormal; when >0.8 or >0.8, the high voltage switch cabinet outputs an early warning; otherwise it is normal; S3.

5. If the results of the two judgments are inconsistent, calculate the fault warning value error , the expression is: in, is the actual expected value; Judge whether the error of the fault warning value is within the preset requirement range. If it is, judge whether the primary weight coefficient needs to be corrected; If not, adjust the secondary weight coefficient until the error of the fault warning value is within the preset requirement range.

5. A high-voltage switch cabinet fault early warning method according to claim 4, characterized in that: The adjustment method of the secondary weight coefficient is as follows: Calculate and correct fault warning value , the expression is: First error correction ; in, is the secondary weight error coefficient, j=1,2; in, is the error reverse learning rate, the parameter affects its iteration speed; Repeat S3.4-S3.5 to continue to determine whether the correction error is within the preset requirement range. If not, continue correction; Similarly, the calculation result of the nth repair is: The nth correction error ; Compare with the preset requirement range each time until the error is corrected Within the preset requirements, the optimal secondary weight coefficient is finally determined. , .

6. A high-voltage switch cabinet fault early warning method according to claim 4, characterized in that: Determine whether the first-level weight coefficient needs to be corrected, specifically: First calculate the error of the partial discharge warning value , the expression is: in, is the expected value of partial discharge warning, is the partial discharge warning value; Judgment Error Whether the error requirements are met. If so, it means that the first-level weight coefficient is appropriate; If it is not satisfied, the first-level weight coefficient is modified as follows: The nth corrected PD warning value ,for: = + in is the nth first-order weighted partial discharge correction coefficient, and its value is: Error after the nth correction , specifically: when When the value of meets the error requirement, the optimal first-level partial discharge warning weight coefficient and the optimal first-level environmental warning weight coefficient are obtained.

7. A high-voltage switch cabinet fault early warning method according to claim 4, characterized in that: In S3.3, the activation function of the early warning is the Sigmod function, which is: 。 8. A high-voltage switch cabinet fault warning system, characterized in that: include: Data acquisition module, used to acquire ultrasonic signals, transient ground voltage signals, UHF detection signals, ozone concentration, oxygen concentration, SF6 signals, temperature signals and humidity signals; A data preprocessing module is used to perform data preprocessing on transient ground voltage signals, ultra-high frequency detection signals, ozone concentration, SF6 signals, oxygen concentration, temperature signals and humidity signals to obtain processed signals; The prediction module is used to input the ultrasonic signal, the pre-processed ozone concentration, the oxygen concentration, the temperature signal, the SF6 signal, the humidity signal, the UHF detection signal and the transient ground voltage signal into the trained BP neural network model and output the prediction result.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the high-voltage switch cabinet fault warning method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, the steps of the high-voltage switch cabinet fault warning method as claimed in any one of claims 1 to 7 are implemented.

Images