Medium-voltage grounding fault intelligent diagnosis system and control method thereof

Abstract

The invention relates to the technical field of medium-voltage grounding fault intelligent diagnosis, in particular to a medium-voltage grounding fault intelligent diagnosis system and a control method thereof, and the system comprises a data collection module, a signal processing module, a feature extraction module, a fault diagnosis module, and a display and alarm module. The data acquisition module is used for acquiring transient current signals in a medium-voltage distribution line, adopts a Rogowski coil and has a temperature compensation function; the signal processing module is used for decomposing the collected transient current signals through multi-resolution wavelet transform, and the signal processing module is used for extracting time-frequency characteristics of the signals; the feature extraction module is used for calculating an energy moment Qj under each decomposition scale so as to represent energy distribution of a fault signal in different frequency bands; according to the method, the time-frequency characteristics of the transient signals are extracted through wavelet transform, and the least square multi-stage support vector machine is combined, so that the high-resistance grounding fault in a complex environment can be effectively identified, and the diagnosis precision and reliability are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medium voltage ground fault intelligent diagnosis, and in particular to a medium voltage ground fault intelligent diagnosis system and a control method thereof. Background Art

[0002] The medium voltage ground fault intelligent diagnosis system is an advanced monitoring and analysis system specially designed to detect, locate and diagnose ground faults in medium voltage distribution networks. This system usually integrates multiple sensors, signal processing technologies, data processing algorithms and intelligent diagnostic tools to improve the accuracy of fault detection and response speed;

[0003] Medium-voltage DC distribution networks are widely used in scenarios such as photovoltaic energy access. However, in actual operation, medium-voltage lines may be affected by environmental factors, such as the increase of transition resistance, which leads to the occurrence of high-resistance grounding faults. The traditional fault diagnosis method has low detection accuracy for high-resistance grounding faults, especially under various complex environmental conditions. It is difficult to achieve refined fault identification. Therefore, a medium-voltage grounding fault intelligent diagnosis system and its control method are proposed to address the above problems. Summary of the invention

[0004] The purpose of the present invention is to provide a medium-voltage grounding fault intelligent diagnosis system and a control method thereof to solve the problem that the medium-voltage line may be affected by environmental factors, such as an increase in transition resistance, resulting in the occurrence of high-resistance grounding faults. The traditional fault diagnosis method has low detection accuracy for high-resistance grounding faults, especially under various complex environmental conditions, and it is difficult to achieve refined fault identification.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A medium voltage ground fault intelligent diagnosis system and control method thereof, comprising: a data acquisition module, a signal processing module, a feature extraction module, a fault diagnosis module, a display and alarm module;

[0007] The data acquisition module is used to collect transient current signals in the medium voltage distribution line. The data acquisition module adopts a Rogowski coil and is equipped with a temperature compensation function.

[0008] The signal processing module is used to decompose the collected transient current signal through multi-resolution wavelet transform, and the signal processing module is used to extract the time-frequency characteristics of the signal;

[0009] The feature extraction module is used to calculate the energy moment Q at each decomposition scale. j , to characterize the energy distribution of fault signals in different frequency bands;

[0010] The fault diagnosis module adopts the least squares multi-level support vector machine, takes the extracted time-frequency features as input, performs classification diagnosis of high-resistance grounding faults, and outputs fault diagnosis results;

[0011] The display and alarm module is used to display the fault diagnosis results in real time and trigger an alarm when a high-resistance grounding fault is detected. The alarm module includes multiple alarm modes such as sound, light, text message or APP push.

[0012] As a further optimized content of the present invention, the signal processing module uses orthogonal wavelet transform to perform multi-scale decomposition on the transient current signal, the decomposition level is J, and the signal obtained at each decomposition scale includes a detail signal and a discrete smooth approximation value.

[0013] As a further optimization of the present invention, the feature extraction module calculates the energy moment Q at each decomposition scale by the following formula: j :

[0014]

[0015] Where M is the number of sampling points, D j [n] is the j Discrete detail signals at decomposition scales, γ j is the corresponding weight factor, and △t is the sampling time interval.

[0016] As a further optimization of the present invention, the least squares multi-level support vector machine in the fault diagnosis module adopts the kernel function K(x i ,x j ), and optimize its parameters by Lagrangian optimization method, the optimization goal is:

[0017]

[0018] In the formula, w is the weight vector, C is the penalty parameter, ξ i is the relaxation factor.

[0019] As a further optimization of the present invention, the fault diagnosis module uses historical fault data to train the least squares multi-level support vector machine, and updates its model parameters through real-time signals to achieve online learning and dynamic fault diagnosis.

[0020] As a further optimized content of the present invention, the display and alarm module includes an alarm threshold setting unit, and when the predicted probability of failure exceeds a preset threshold, the system automatically triggers an alarm signal.

[0021] As a further optimized content of the present invention, the following steps are included:

[0022] S1: Signal acquisition: The data acquisition module is used to collect transient current signals in the medium voltage distribution line in real time;

[0023] S2: Signal processing: Use orthogonal wavelet transform to perform multi-scale decomposition on the collected transient current signal to obtain the detail signal and discrete smooth approximation value at each decomposition scale;

[0024] S3: Feature extraction: extract the time-frequency characteristics of the fault signal by calculating the energy moment at each decomposition scale;

[0025] S4: Fault diagnosis: The extracted time-frequency features are input into the least squares multi-level support vector machine for fault classification diagnosis to obtain the probability of fault occurrence or classification results;

[0026] S5: Result display and alarm: Real-time display and alarm triggering according to the fault diagnosis results.

[0027] As a further optimization of the present invention, in the feature extraction step, the energy moment Q is calculated j The energy distribution of the signal in each frequency band is characterized and combined with the nonlinear characteristics of the signal to identify high-resistance grounding faults.

[0028] As a further optimization of the present invention, the fault diagnosis step adjusts the parameters of the support vector machine through the Lagrangian optimization method to achieve refined classification diagnosis of different fault types.

[0029] As further optimized content of the present invention, among others: the system utilizes the data collected in real time to perform online learning and model updating, thereby ensuring the timeliness and accuracy of fault diagnosis.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] 1. In the present invention, the time-frequency characteristics of transient signals are extracted by wavelet transform, and combined with the least squares multi-level support vector machine, high-resistance grounding faults in complex environments can be effectively identified, and the accuracy and reliability of diagnosis are improved. The least squares method and Lagrangian optimization technology are used to optimize the parameters of the support vector machine, so that the diagnosis system can maintain efficient fault identification capabilities under various environmental conditions, and the real-time signal processing and online learning mechanism can quickly respond to and process new fault signals, ensuring that the distribution network can timely and accurately diagnose and alarm when a fault occurs, and ensuring the safe operation of the system;

[0032] 2. In the present invention, by integrating multi-resolution wavelet transform and least squares multi-level support vector machine, the high-resistance grounding fault signal in the medium-voltage DC distribution network is effectively extracted and analyzed. By using the multi-scale analysis capability of wavelet transform, the system can capture the time-frequency characteristics of the fault signal and realize accurate diagnosis of complex nonlinear faults.

[0033] 3. In the present invention, the system can adjust the diagnostic model in real time, dynamically adapt to the changing fault modes and operating conditions, ensure that efficient fault identification capabilities can be maintained in different environments, reduce the downtime of the power system, reduce operation and maintenance costs, and extend the service life of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 The present invention is a flow chart of a medium voltage grounding fault intelligent diagnosis system and a control method thereof. DETAILED DESCRIPTION

[0035] See also Figure 1 , the present invention provides a technical solution:

[0036] A medium voltage ground fault intelligent diagnosis system and control method thereof, comprising: a data acquisition module, a signal processing module, a feature extraction module, a fault diagnosis module, a display and alarm module;

[0037] The data acquisition module is used to collect transient current signals in the medium voltage distribution line. The data acquisition module adopts a Rogowski coil and is equipped with a temperature compensation function.

[0038] The signal processing module is used to decompose the collected transient current signal through multi-resolution wavelet transform, and the signal processing module is used to extract the time-frequency characteristics of the signal;

[0039] The feature extraction module is used to calculate the energy moment Q at each decomposition scale. j , to characterize the energy distribution of fault signals in different frequency bands;

[0040] The fault diagnosis module adopts the least squares multi-level support vector machine, takes the extracted time-frequency features as input, performs classification diagnosis of high-resistance grounding faults, and outputs fault diagnosis results;

[0041] The display and alarm module is used to display the fault diagnosis results in real time and trigger an alarm when a high-resistance grounding fault is detected. The alarm module includes multiple alarm modes such as sound, light, text message or APP push; through the coordinated work of various modules, the system can realize fast and accurate diagnosis of medium-voltage grounding faults to ensure the safe operation of the power system.

[0042] As a technical solution for further implementing this solution, the signal processing module uses orthogonal wavelet transform to perform multi-scale decomposition on the transient current signal, the decomposition level is J, and the signal obtained at each decomposition scale includes a detail signal and a discrete smooth approximation value. The orthogonal wavelet transform provides a multi-scale analysis of the fault signal, which can effectively capture the fault characteristics in different frequency bands and enhance the accuracy and reliability of diagnosis;

[0043] As a technical solution for further implementing this solution, the feature extraction module calculates the energy moment Q at each decomposition scale by the following formula: j :

[0044]

[0045] Where M is the number of sampling points, D j [n] is the j Discrete detail signals at decomposition scales, γ j is the corresponding weight factor, △t is the sampling time interval, and the calculation of energy moment can quantify the energy distribution of the signal in each frequency band, providing refined fault feature extraction and supporting more accurate fault diagnosis;

[0046] As a further technical solution for implementing this solution, the least squares multi-level support vector machine in the fault diagnosis module adopts the kernel function K(x i ,x j ), and optimize its parameters by Lagrangian optimization method, the optimization goal is:

[0047]

[0048] In the formula, w is the weight vector, C is the penalty parameter, ξ i is the relaxation factor. Through the Lagrangian optimization method, the system can optimize the decision boundary of the support vector machine to ensure the accuracy and robustness of fault classification;

[0049] As a technical solution for further implementation of this solution, the fault diagnosis module uses historical fault data to train the least squares multi-level support vector machine, and updates its model parameters through real-time signals to achieve online learning and dynamic fault diagnosis. Through online learning and model updating, the system can adapt to the ever-changing operating environment and improve the real-time and accuracy of fault diagnosis;

[0050] As a technical solution for further implementation of this scheme, the display and alarm module includes an alarm threshold setting unit. When the predicted probability of failure exceeds the preset threshold, the system automatically triggers an alarm signal. The set alarm threshold can ensure that an alarm is issued in time when a high probability failure occurs, thereby reducing system downtime and reducing accident risks.

[0051] Embodiment 1

[0052] Background: A medium-voltage DC distribution network faces the risk of high-resistance grounding faults due to the access of photovoltaic energy. Especially when environmental conditions change, traditional fault detection methods are difficult to accurately identify such faults. In order to improve the accuracy and timeliness of fault detection, a medium-voltage grounding fault intelligent diagnosis system based on the present invention is implemented;

[0053] Implementation steps:

[0054] S1: Signal acquisition: High-precision current sensors are installed at key nodes of the distribution network to collect transient current signals in real time. The data acquisition module transmits these signals to the signal processing module;

[0055] S2: Signal processing: Use orthogonal wavelet transform to perform multi-scale decomposition on the collected transient current signal. Assuming that the signal is decomposed into 4 scales, the detail signal and discrete smooth approximation value at each decomposition scale are obtained. Through this process, the high-frequency and low-frequency features in the signal are effectively separated;

[0056] S3: Feature extraction: Calculate the energy moment Q for each decomposition scale j , these energy moments Q j It reflects the energy distribution of the signal in each frequency band. The extracted features show that there is a significant energy concentration at the third decomposition scale, indicating that there may be a high-resistance grounding fault.

[0057] S4: Fault diagnosis: The extracted time-frequency features are input into the least squares multi-level support vector machine for classification diagnosis. Through the optimized kernel function and weight, the system accurately identifies the high-resistance grounding fault and gives a fault probability of 85%;

[0058] S5: Result display and alarm: Since the fault probability exceeds the 70% alarm threshold preset by the system, the display and alarm module immediately triggers an alarm signal, displays the fault location and possible fault type on the monitoring interface, and notifies the operation and maintenance personnel to handle it;

[0059] Results: Through the above implementation, the system successfully detected and diagnosed high-resistance grounding faults in the medium-voltage distribution network, ensuring that the faults were identified and handled at an early stage, avoiding more serious power accidents, which proved the efficiency and accuracy of the present invention in high-resistance grounding fault detection.

[0060] Embodiment 2

[0061] Background: In another project, an industrial park adopted a medium-voltage DC distribution network to provide power support for multiple production workshops. Due to the high-frequency start and stop operations of the production equipment, transient current fluctuations often occur in the distribution network, increasing the risk of high-resistance grounding faults. This embodiment applies the intelligent diagnostic system of the present invention to ensure the stable operation of the distribution network.

[0062] Implementation steps:

[0063] S1: Signal acquisition: The installed sensor network covers the key power nodes of the entire industrial park, collecting transient current signals of each node in real time. The data acquisition module collects these signals to the central processing unit;

[0064] S2: Signal processing: Orthogonal wavelet transform is used to decompose the collected signal into 6 layers. Through decomposition, the system successfully extracts the signal details including high-frequency and low-frequency features, and pays special attention to the abnormal signals appearing in the 5th and 6th layers of decomposition;

[0065] S3: Feature extraction: Calculate the energy moment Q for each decomposition scale signal j During the analysis, it was found that the energy moment Q of the 5th layer j Significantly higher than other layers, suggesting the possibility of high-resistance ground fault in this frequency band;

[0066] S4: Fault diagnosis: The energy moment Q extracted above is j The features were input into the least squares multi-level support vector machine for fault diagnosis. After online learning, the system has the ability to classify complex fault types. The diagnosis results showed that there was a high-resistance grounding fault on the power line of a specific workshop, with a fault probability of 92%.

[0067] S5: Result display and alarm: Based on the diagnosis results, the system will sound an alarm in the control room and highlight the specific location and type of the fault on the monitoring screen. The maintenance personnel will quickly locate and solve the fault according to the system prompts, ensuring the continuity of production;

[0068] Results: In practical applications, the intelligent diagnosis system of the present invention successfully identified and diagnosed high-resistance grounding faults in complex environments, ensuring the stability of the distribution network in the industrial park, which further demonstrated the applicability and reliability of the present invention under various environmental conditions.

[0069] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only used to help understand the methods and core ideas of the present invention. The above is only a preferred implementation of the present invention. It should be pointed out that due to the limitations of textual expression and the objective existence of infinite specific structures, ordinary technicians in this technical field can make several improvements, modifications or changes without departing from the principles of the present invention, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes or combinations, or the direct application of the inventive concept and technical solution to other occasions without improvement, should be regarded as the protection scope of the present invention.

Claims

1. A medium voltage grounding fault intelligent diagnosis system, characterized in that: include: Data acquisition module, signal processing module, feature extraction module, fault diagnosis module, display and alarm module; The data acquisition module is used to collect transient current signals in the medium voltage distribution line. The data acquisition module adopts a Rogowski coil and is equipped with a temperature compensation function. The signal processing module is used to decompose the collected transient current signal through multi-resolution wavelet transform, and the signal processing module is used to extract the time-frequency characteristics of the signal; The feature extraction module is used to calculate the energy moment Q at each decomposition scale. j , to characterize the energy distribution of fault signals in different frequency bands; The fault diagnosis module adopts the least squares multi-level support vector machine, takes the extracted time-frequency features as input, performs classification diagnosis of high-resistance grounding faults, and outputs fault diagnosis results; The display and alarm module is used to display the fault diagnosis results in real time and trigger an alarm when a high-resistance grounding fault is detected. The alarm module includes multiple alarm modes such as sound, light, text message or APP push.

2. A medium voltage ground fault intelligent diagnosis system according to claim 1, characterized in that: The signal processing module uses orthogonal wavelet transform to perform multi-scale decomposition on the transient current signal, the number of decomposition levels is J, and the signal obtained at each decomposition scale includes a detail signal and a discrete smooth approximation value.

3. A medium voltage ground fault intelligent diagnosis system according to claim 1, characterized in that: The feature extraction module calculates the energy moment Q at each decomposition scale by the following formula j : Where M is the number of sampling points, D j [n] is the j Discrete detail signals at decomposition scales, γ j is the corresponding weight factor, and △t is the sampling time interval.

4. A medium voltage ground fault intelligent diagnosis system according to claim 1, characterized in that: The least squares multi-level support vector machine in the fault diagnosis module adopts the kernel function K(x i ,x j ), and optimize its parameters by Lagrangian optimization method, the optimization goal is: In the formula, w is the weight vector, C is the penalty parameter, ξ i is the relaxation factor.

5. The medium voltage ground fault intelligent diagnosis system according to claim 1, characterized in that: The fault diagnosis module uses historical fault data to train the least squares multi-level support vector machine and updates its model parameters through real-time signals to achieve online learning and dynamic fault diagnosis.

6. A medium voltage ground fault intelligent diagnosis system according to claim 1, characterized in that: The display and alarm module includes an alarm threshold setting unit. When the predicted probability of a fault occurring exceeds a preset threshold, the system automatically triggers an alarm signal.

7. A control method for a medium voltage ground fault intelligent diagnosis system according to any one of claims 1 to 6, characterized in that: The following steps are involved: S1: Signal acquisition: The data acquisition module is used to collect transient current signals in the medium voltage distribution line in real time; S2: Signal processing: Use orthogonal wavelet transform to perform multi-scale decomposition on the collected transient current signal to obtain the detail signal and discrete smooth approximation value at each decomposition scale; S3: Feature extraction: extract the time-frequency characteristics of the fault signal by calculating the energy moment at each decomposition scale; S4: Fault diagnosis: The extracted time-frequency features are input into the least squares multi-level support vector machine for fault classification diagnosis to obtain the probability of fault occurrence or classification results; S5: Result display and alarm: Real-time display and alarm triggering according to the fault diagnosis results.

8. The control method of a medium voltage ground fault intelligent diagnosis system according to claim 7, characterized in that: In the feature extraction step, the energy moment Q is calculated j The energy distribution of the signal in each frequency band is characterized and combined with the nonlinear characteristics of the signal to identify high-resistance grounding faults.

9. The control method of a medium voltage ground fault intelligent diagnosis system according to claim 7, characterized in that: The fault diagnosis step adjusts the parameters of the support vector machine through the Lagrangian optimization method to achieve refined classification diagnosis of different fault types.

10. The control method of a medium voltage ground fault intelligent diagnosis system according to claim 7, characterized in that: The system uses real-time collected data for online learning and model updating to ensure the timeliness and accuracy of fault diagnosis.

Images