Booster station electrical test live detection system and method

Abstract

The invention discloses a booster station electrical test live detection system and method. According to the invention, the data analysis and judgment module in the intelligent analysis and diagnosis module can quickly identify abnormal data in the operation of the electrical equipment through threshold alarm and an isolated forest anomaly detection algorithm, and an accurate data basis is provided for subsequent fault diagnosis. The model building module continuously adjusts model parameters through an Adam optimization algorithm, so that the model is gradually converged in the training process, and the accuracy of fault diagnosis is improved. And finally, the model evaluation and optimization module evaluates the model performance through indexes such as accuracy, recall rate and F1 score, adjusts the model structure or training strategy according to the evaluation result, and further optimizes the diagnosis effect. Therefore, accurate prediction and diagnosis of faults of the electrical equipment are realized, and misdiagnosis or missed diagnosis caused by artificial judgment errors or improper data processing in a traditional method is effectively avoided.

Description

Technical Field

[0001] The invention belongs to the technical field of electrical detection of booster stations, and in particular relates to a live detection system and method for electrical testing of booster stations. Background Art

[0002] Electrical testing and inspection of booster stations is an important safeguard for the safe operation of power systems. Its main purpose is to ensure that the performance, parameters and operating status of electrical equipment in booster stations meet national standards and design requirements. During the test and inspection process, professionals will evaluate the status of the equipment based on the actual condition of the equipment and the test results, and formulate reasonable maintenance and overhaul plans to ensure the reliability and economy of the power system. With the rapid development of the power industry, booster stations are an important part of the power system, and the stable operation of their electrical equipment is crucial to the safety of the entire power grid. However, since electrical equipment is in a high-voltage, high-current working environment for a long time, various faults will inevitably occur. If these faults are not detected and handled in time, they may cause serious safety accidents, resulting in huge economic losses and social impacts. Electrical testing and inspection of booster stations is an important part of the maintenance and management of power facilities, and is of vital importance to improving the operation level of the power system and preventing accidents.

[0003] However, the detection systems in the prior art rely on manual experience and simple algorithms for fault diagnosis, which makes it difficult to accurately predict and identify complex faults. At the same time, manual data analysis is time-consuming and labor-intensive, and it is difficult to achieve timely monitoring and rapid response to real-time faults. Summary of the invention

[0004] The purpose of the present invention is to provide a live detection system and method for electrical testing of a booster station in order to solve the above-mentioned problems.

[0005] The technical solution adopted by the present invention is as follows: a live detection system for electrical testing of a booster station, the system comprising: a data acquisition module, a signal processing module, a communication module, a control module, an intelligent analysis and diagnosis module, a display module and a power supply module;

[0006] The intelligent analysis and diagnosis module is internally provided with a data analysis and judgment module, a model construction module, a model training module and a model evaluation and optimization module;

[0007] The data acquisition module first collects the current, voltage and temperature key parameters of the electrical equipment in real time through high-precision sensors, and transmits these raw data to the signal processing module;

[0008] The signal processing module filters, amplifies, and performs analog-to-digital conversion on the data to ensure the accuracy and reliability of the data, and then sends the processed digital signal to the intelligent analysis and diagnosis module;

[0009] The intelligent analysis and diagnosis module uses artificial intelligence technology to perform deep learning and analysis on data to achieve fault prediction and diagnosis, and transmits the analysis results and diagnosis reports to the control module and the display module;

[0010] The control module executes the corresponding control strategy according to the analysis results, and feeds back the control instructions to the data acquisition module and the signal processing module; the display module displays the real-time data, analysis results and system status to the operator in the form of a graphical interface, providing intuitive information display and alarm prompts; the communication module is responsible for transmitting the data and analysis results in the system to the remote monitoring center or cloud platform to realize remote monitoring and big data analysis, and at the same time receiving control instructions and update programs from the remote; the power module provides a stable and reliable power supply for the entire system to ensure the normal operation of each module.

[0011] In a preferred embodiment, the data acquisition module is internally provided with:

[0012] Sensor unit: including current sensor, voltage sensor, temperature sensor, humidity sensor, used to measure different physical quantities of electrical equipment;

[0013] Signal conditioning circuit: amplifies, filters, and isolates the weak signal output by the sensor to meet the needs of subsequent circuits;

[0014] Data acquisition card: Use high-performance data acquisition card to achieve multi-channel, high sampling rate data acquisition to ensure data integrity and accuracy.

[0015] In a preferred embodiment, the signal processing module performs in-depth processing and analysis on the original signal output by the data acquisition module to extract useful information; first, the signal is filtered through a filtering circuit to remove noise and interference and improve the signal-to-noise ratio of the signal; the filtered signal is amplified by a signal amplification circuit to meet the input requirements of a subsequent analysis circuit;

[0016] The communication module transmits the data of the signal processing module to the intelligent analysis and diagnosis module, and realizes seamless connection with the remote monitoring system; the communication module provides a variety of communication interfaces, including Ethernet, serial port and wireless communication, and the data transmission protocol of the communication module adopts international standards;

[0017] The control module performs corresponding control operations according to the analysis results of the signal processing module and the intelligent analysis and diagnosis module to adjust the operating status of the electrical equipment or respond to fault conditions; the control module adopts PID control or fuzzy control to perform real-time control of the electrical equipment according to preset parameters and algorithms; the actuator is the execution unit of the control module, including relays and contactors, which are used to receive control commands and perform corresponding actions.

[0018] In a preferred embodiment, the data analysis and judgment module first performs a format check and preliminary screening of the booster station electrical test data to ensure the accuracy and completeness of the booster station electrical test data;

[0019] The threshold alarm system sets the thresholds of each monitoring indicator according to the normal electrical operation range and historical data of the booster station; the current threshold is set to I_min to I_max, the voltage threshold is set to V_min to V_max, and the temperature threshold is set to T_min to T_max; the central processing unit compares the real-time monitoring data with the preset thresholds; if the data exceeds the threshold range, the alarm mechanism is immediately triggered to notify the operation and maintenance personnel;

[0020] Use the machine learning-based isolation forest anomaly detection algorithm to conduct deep data analysis to identify potential failure modes;

[0021] Model training: Use historical normal data to train the isolation forest model; model parameters include:

[0022] n_estimators: the number of isolated trees, which affects the complexity of the model and the detection effect;

[0023] max_samples: The number of samples used by each tree, set as a fraction of the total number of samples;

[0024] Contamination: An estimate of the proportion of outliers in the data, used to adjust the sensitivity of the model;

[0025] Real-time detection: Input the real-time normalized booster station electrical test data into the trained isolation forest model and calculate the anomaly score of each data point;

[0026] The anomaly score calculation formula is:

[0027] anomaly_score=2^-(E(h(x)) / c(n))

[0028] Anomaly judgment: Determine whether it is an abnormal point based on the anomaly score and the preset anomaly threshold; the higher the anomaly score, the more likely the booster station electrical test data point is to be an abnormal point; where E(h(x)) is the average path length of the data point x in the isolation forest, c(n) is the expected value of the path length, and n is the number of samples.

[0029] In a preferred embodiment, the model building module selects a convolutional neural network (CNN) as a fault diagnosis algorithm; specifically includes:

[0030] Input layer: input preprocessed feature data;

[0031] Convolutional layer: Use multiple convolution kernels to extract local features in the data;

[0032] Pooling layer: reduces the dimension of the features output by the convolutional layer to reduce the amount of calculation;

[0033] Fully connected layer: converts the feature vector output by the pooling layer into the probability distribution of the fault type;

[0034] Output layer: Outputs the prediction results of fault types.

[0035] In a preferred embodiment, the model training module uses a cross entropy loss function to measure the difference between the model prediction result and the true label, and the calculation formula is:

[0036] ,in:

[0037] H(p,q) represents the cross entropy loss;

[0038] p(xi) represents the probability of the i-th type of fault in the true label. For one-hot encoding, the probability of the true category is 1 and the rest are 0;

[0039] q(xi) represents the probability of the i-th type of failure predicted by the model;

[0040] ∑i represents the sum of all fault categories;

[0041] The model training module uses the Adam optimization algorithm to update the model parameters and speed up the convergence speed, wherein the calculation process includes:

[0042] Initialization: Set the initial learning rate α, decay rates β1 and β2, the minimum constant ∈, and the initial value θ of the parameter;

[0043] Calculate the gradient: Calculate the gradient of the loss function with respect to the parameter θ;

[0044] Update the first and second moments, where:

[0045] The formula for calculating the first-order moment momentum is: mt =β1m t-1 +(1-β1)g t

[0046] The calculation formula for the second-order moment square gradient is:

[0047] v t =β2v t-1 +(1-β2)gt 2 Among them, g t is the gradient at time step t, m t and v t are the first and second order moments of time step t respectively;

[0048] Bias correction: bias correction is performed on the first-order moment and the second-order moment to offset the effect of their initial values ​​being 0;

[0049] The bias correction calculation formula is:

[0050]

[0051]

[0052] In the process of updating parameters, the parameters are updated according to the corrected first-order moment and second-order moment. The calculation formula is:

[0053] Where α is the learning rate;

[0054] β1 and β2 are the decay rates, set to 0.9 and 0.999;

[0055] ∈ is a very small constant to prevent division by 0, set to 1e-8;

[0056] θ is the model parameter;

[0057] g t is the gradient at time step t, m t and v t are the first and second order moments of time step t respectively;

[0058] The model training module divides the entire training data set into multiple small batches during batch training, each batch contains a certain number of samples, and the calculation process includes:

[0059] S1. Divide into batches: Divide the entire substation electrical training dataset into multiple batches, each batch contains NN samples;

[0060] S2. Forward propagation: forward propagation of each batch of booster station electrical data to calculate the output of the model;

[0061] S3. Calculate loss: Based on the model output and the true label, use the cross entropy loss function to calculate the loss of each batch;

[0062] S4. Back propagation: Back propagate the loss of each batch and calculate the gradient of the parameters;

[0063] S5. Update parameters: Use the Adam optimization algorithm to update the model parameters according to the calculated gradients;

[0064] S6. Repeat step: Repeat steps S2-S5 until all batches are traversed or the set number of training rounds is reached.

[0065] In a preferred embodiment, the model evaluation and optimization module uses accuracy, recall, and F1 score indicators to evaluate model performance; adjusts the model structure, parameters, or training strategy based on the evaluation results to improve diagnostic accuracy;

[0066] Among them: The calculation formula of the cross entropy loss function is:

[0067] L=-Σ(y_log(p)+(1-y)_log(1-p));

[0068] Among them, y is the true label (0 or 1), and p is the probability predicted by the model;

[0069] The calculation formula of the Adam optimization algorithm is:

[0070] m_t=β1*m_(t-1)+(1-β1)*g_t

[0071] v_t=β2*v_(t-1)+(1-β2)*g_t^2

[0072] m_t_hat=m_t / (1-β1^t)

[0073] v_t_hat=v_t / (1-β2^t)

[0074] θ_t=θ_(t-1)-α*m_t_hat / (sqrt(v_t_hat)+ε);

[0075] Among them, m_t and v_t are estimates of the first-order moment and the second-order moment respectively, β1 and β2 are decay rates, g_t is the gradient, α is the learning rate, ε is a small constant, and θ_t is the updated parameter;

[0076] The convolution layer parameters include: convolution kernel size: used to extract local features; step size: represents the step size of the convolution kernel moving on the data; padding: used to control the size of the output feature map.

[0077] In a preferred embodiment, the display module displays the system's operating status, detection data, and fault information to the operator in an intuitive and easy-to-understand manner; a high-resolution LCD screen is used to display clear graphical interfaces and text information; the display content is rich, including real-time data curves, equipment status indications, and fault alarm prompts; when a fault occurs, the display module will immediately pop up an alarm window to display the fault type, location, and suggested solutions to help operators quickly respond to and handle the fault.

[0078] In a preferred embodiment, the power module converts the input AC power into various DC voltages required by the system; the module adopts an electromagnetic compatibility design to reduce the impact of power supply noise and electromagnetic interference on the system; the power monitoring circuit of the power module monitors the power supply voltage and current in real time, and once an abnormality is found, an alarm is immediately triggered and fault information is recorded.

[0079] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0080] 1. In the present invention, the intelligent analysis and diagnosis module significantly improves the accuracy of the live detection system of the electrical test of the booster station in fault prediction and diagnosis by integrating submodules such as data analysis and judgment, model construction, model training, and model evaluation and optimization. The data analysis and judgment module can quickly identify abnormal data in the operation of electrical equipment through threshold alarm and isolated forest anomaly detection algorithm, providing an accurate data basis for subsequent fault diagnosis. The model construction module adopts the advanced artificial intelligence technology of convolutional neural network CNN, which can deeply mine the feature information in the data, so as to more accurately identify the fault type. The model training module uses historical data and cross entropy loss function, and continuously adjusts the model parameters through the Adam optimization algorithm, so that the model gradually converges during the training process, and improves the accuracy of fault diagnosis. Finally, the model evaluation and optimization module evaluates the model performance through indicators such as accuracy, recall rate, F1 score, and adjusts the model structure or training strategy according to the evaluation results to further optimize the diagnostic effect. Thereby, accurate prediction and diagnosis of electrical equipment faults are achieved, effectively avoiding the misdiagnosis or missed diagnosis caused by human judgment errors or improper data processing in traditional methods.

[0081] 2. In the present invention, the intelligent analysis and diagnosis module greatly reduces the need for manual intervention and improves the operating efficiency of the system through automated data analysis and intelligent fault diagnosis. In addition, the module can also monitor the operating status of electrical equipment in real time, detect potential faults in a timely manner and issue warnings, achieve preventive maintenance, and effectively extend the service life of the equipment. At the same time, the module's support for remote monitoring allows operators to monitor and manage the system anytime and anywhere, further improving the convenience and practicality of the system. In general, the intelligent analysis and diagnosis module not only improves the technical level of the live detection system for electrical tests in the booster station in fault prediction and diagnosis, but also promotes the development of the entire system towards intelligence and automation, providing strong technical support for the stable operation and efficient management of the booster station. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 is the overall system block diagram of the present invention;

[0083] Figure 2 This is a system block diagram of the intelligent analysis and diagnosis module in the present invention. DETAILED DESCRIPTION

[0084] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with 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.

[0085] Reference Figure 1-2 ,

[0086] Example:

[0087] A live detection system for electrical testing of a booster station, the system comprising: a data acquisition module, a signal processing module, a communication module, a control module, an intelligent analysis and diagnosis module, a display module and a power module;

[0088] The intelligent analysis and diagnosis module is internally configured with a data analysis and judgment module, a model building module, a model training module, and a model evaluation and optimization module;

[0089] The data acquisition module first collects key parameters of electrical equipment such as current, voltage, temperature, etc. in real time through high-precision sensors, and transmits these raw data to the signal processing module;

[0090] The signal processing module filters, amplifies, and performs analog-to-digital conversion on the data to ensure the accuracy and reliability of the data, and then sends the processed digital signal to the intelligent analysis and diagnosis module;

[0091] The intelligent analysis and diagnosis module uses artificial intelligence technology to conduct deep learning and analysis of data to achieve fault prediction and diagnosis, and transmits the analysis results and diagnosis reports to the control module and display module;

[0092] The control module executes the corresponding control strategy according to the analysis results, and feeds back the control instructions to the data acquisition module and the signal processing module; the display module presents the real-time data, analysis results and system status to the operator in the form of a graphical interface, providing intuitive information display and alarm prompts; the communication module is responsible for transmitting the data and analysis results in the system to the remote monitoring center or cloud platform to realize remote monitoring and big data analysis, and at the same time receive control instructions and update programs from the remote; the power module provides a stable and reliable power supply for the entire system to ensure the normal operation of each module.

[0093] The internal settings of the data acquisition module are:

[0094] Sensor unit: includes current sensor, voltage sensor, temperature sensor, humidity sensor, etc., which are used to measure different physical quantities of electrical equipment.

[0095] Signal conditioning circuit: amplifies, filters, isolates, and processes the weak signals output by the sensor to meet the needs of subsequent circuits.

[0096] Data acquisition card: Use high-performance data acquisition card to achieve multi-channel, high sampling rate data acquisition to ensure data integrity and accuracy.

[0097] The signal processing module processes and analyzes the original signal output by the data acquisition module in depth to extract useful information. First, the signal is filtered through the filter circuit to remove noise and interference and improve the signal-to-noise ratio. The filtered signal is amplified by the signal amplifier circuit to meet the input requirements of the subsequent analysis circuit;

[0098] The communication module transmits the data of the signal processing module to the intelligent analysis and diagnosis module, and realizes seamless connection with the remote monitoring system. The communication module provides a variety of communication interfaces, including Ethernet, serial port and wireless communication. The data transmission protocol of the communication module adopts international standards, such as MODBUS, TCP / IP, etc., to ensure the reliability of data transmission and compatibility between systems;

[0099] The control module performs corresponding control operations according to the analysis results of the signal processing module and the intelligent analysis and diagnosis module to adjust the operating status of the electrical equipment or respond to fault conditions. The control module adopts PID control or fuzzy control to control the electrical equipment in real time according to preset parameters and algorithms. The actuator is the execution unit of the control module, including relays, contactors, etc., which are used to receive control commands and perform corresponding actions, such as switch switching, current regulation, etc. In order to ensure the safe operation of the system under abnormal conditions, the control module is designed with a variety of protection circuits, such as overcurrent protection, overvoltage protection, etc., which can quickly cut off the power supply or take other protective measures when abnormal conditions are detected.

[0100] The data analysis and judgment module first performs format verification and preliminary screening of the booster station electrical test data to ensure the accuracy and completeness of the booster station electrical test data;

[0101] The threshold alarm system sets the thresholds of each monitoring indicator according to the normal electrical operation range and historical data of the booster station; the current threshold is set to I_min to I_max, the voltage threshold is set to V_min to V_max, and the temperature threshold is set to T_min to T_max; the central processing unit compares the real-time monitoring data with the preset thresholds; if the data exceeds the threshold range, the alarm mechanism is immediately triggered to notify the operation and maintenance personnel;

[0102] Use the machine learning-based isolation forest anomaly detection algorithm to conduct deep data analysis to identify potential failure modes;

[0103] Model training: Use historical normal data to train the isolation forest model; model parameters include:

[0104] n_estimators: the number of isolated trees, which affects the complexity of the model and the detection effect;

[0105] max_samples: The number of samples used by each tree, set as a fraction of the total number of samples;

[0106] Contamination: An estimate of the proportion of outliers in the data, used to adjust the sensitivity of the model;

[0107] Real-time detection: Input the real-time normalized booster station electrical test data into the trained isolation forest model and calculate the anomaly score of each data point;

[0108] The anomaly score calculation formula is:

[0109] anomaly_score=2^-(E(h(x)) / c(n))

[0110] Anomaly judgment: Determine whether it is an abnormal point based on the anomaly score and the preset anomaly threshold; the higher the anomaly score, the more likely the booster station electrical test data point is to be an abnormal point; where E(h(x)) is the average path length of the data point x in the isolation forest, c(n) is the expected value of the path length, and n is the number of samples.

[0111] The model building module selects convolutional neural network (CNN) as the fault diagnosis algorithm. Specifically, it includes:

[0112] Input layer: input preprocessed feature data;

[0113] Convolutional layer: Use multiple convolution kernels to extract local features in the data;

[0114] Pooling layer: reduces the dimension of the features output by the convolutional layer to reduce the amount of calculation;

[0115] Fully connected layer: converts the feature vector output by the pooling layer into the probability distribution of the fault type;

[0116] Output layer: Outputs the prediction results of fault types.

[0117] The model training module uses the cross entropy loss function to measure the difference between the model prediction results and the true label. The calculation formula is:

[0118] ,in:

[0119] H(p,q) represents the cross entropy loss.

[0120] p(xi) represents the probability of the i-th category fault in the true label. For one-hot encoding, the probability of the true category is 1 and the rest are 0.

[0121] q(xi) represents the probability of the i-th type of failure predicted by the model.

[0122] ∑i represents the sum of all fault categories;

[0123] The model training module uses the Adam optimization algorithm to update model parameters and speed up convergence. The calculation process includes:

[0124] Initialization: Set the initial learning rate α, decay rates β1 and β2, the minimum constant ∈, and the initial value θ of the parameter.

[0125] Calculate the gradient: Calculate the gradient of the loss function with respect to the parameter θ.

[0126] Update the first and second moments, where:

[0127] The formula for calculating the first-order moment momentum is: m t =β1m t-1+(1-β1)g t

[0128] The calculation formula for the second-order moment square gradient is:

[0129] v t =β2v t-1 +(1-β2)gt 2 Among them, g t is the gradient at time step t, m t and v t are the first and second order moments of time step t, respectively.

[0130] Bias correction: Bias correction is performed on the first-order moment and the second-order moment to offset the effect of their initial values ​​being 0.

[0131] The bias correction calculation formula is:

[0132]

[0133]

[0134] In the process of updating parameters, the parameters are updated according to the corrected first-order moment and second-order moment. The calculation formula is:

[0135] Where α is the learning rate.

[0136] β1 and β2 are the decay rates, which are set to 0.9 and 0.999.

[0137] ∈ is a very small constant to prevent division by 0, usually set to 1e-8.

[0138] θ is the model parameter.

[0139] g t is the gradient at time step t, m t and v t are the first and second order moments of time step t respectively;

[0140] The model training module divides the entire training data set into multiple small batches during batch training. Each batch contains a certain number of samples. The calculation process includes:

[0141] S1. Divide into batches: Divide the entire training dataset into multiple batches, each batch contains NN samples.

[0142] S2. Forward propagation: Perform forward propagation on each batch of data to calculate the output of the model.

[0143] S3. Calculate loss: Calculate the loss of each batch using the cross entropy loss function based on the model’s output and the true label.

[0144] S4. Backpropagation: Backpropagate the loss of each batch and calculate the gradient of the parameters.

[0145] S5. Update parameters: Use the Adam optimization algorithm to update the model parameters according to the calculated gradients.

[0146] S6. Repeat step: Repeat steps S2-S5 until all batches are traversed or the set number of training rounds is reached.

[0147] The model evaluation and optimization module uses accuracy, recall, and F1 score indicators to evaluate model performance; adjust the model structure, parameters, or training strategy based on the evaluation results to improve diagnostic accuracy;

[0148] Among them: The calculation formula of the cross entropy loss function is:

[0149] L=-Σ(y_log(p)+(1-y)_log(1-p));

[0150] Among them, y is the true label (0 or 1), and p is the probability predicted by the model;

[0151] The calculation formula of the Adam optimization algorithm is:

[0152] m_t=β1*m_(t-1)+(1-β1)*g_t

[0153] v_t=β2*v_(t-1)+(1-β2)*g_t^2

[0154] m_t_hat=m_t / (1-β1^t)

[0155] v_t_hat=v_t / (1-β2^t)

[0156] θ_t=θ_(t-1)-α*m_t_hat / (sqrt(v_t_hat)+ε);

[0157] Among them, m_t and v_t are estimates of the first-order moment and the second-order moment respectively, β1 and β2 are decay rates, g_t is the gradient, α is the learning rate, ε is a small constant, and θ_t is the updated parameter;

[0158] The convolution layer parameters include: convolution kernel size: used to extract local features; step size: represents the step size of the convolution kernel moving on the data; padding: used to control the size of the output feature map.

[0159] The display module is the interactive interface between the live detection system of the electrical test of the booster station and the operator. It is responsible for displaying the system's working status, detection data, fault information, etc. to the operator in an intuitive and easy-to-understand way. The module uses a high-resolution LCD screen that can display clear graphical interfaces and text information. The display content is rich, including real-time data curves, equipment status indications, fault alarm prompts, etc., so that operators can understand the operation of the system at a glance. The human-computer interaction design focuses on user experience, provides a friendly operation interface and a simple operation process, so that operators can easily set parameters, query data and view historical records. The display module also supports multi-language display to meet the needs of operators in different countries and regions. In addition, the module has an automatic brightness adjustment function, which can automatically adjust the screen brightness according to the intensity of the ambient light, ensuring the display effect while saving energy. When a fault occurs, the display module will immediately pop up an alarm window to display the fault type, location and recommended solutions to help operators quickly respond to and handle the fault. The historical data backtracking function allows operators to view the detection data and analysis results over a period of time in the past, providing data support for equipment maintenance and fault analysis.

[0160] The power module provides a stable and reliable power supply for the live detection system of the electrical test of the booster station, which is the basic guarantee for the normal operation of the system. The module adopts a high-performance power converter to convert the input AC power into various DC voltages required by the system, such as 5V, 12V, 24V, etc., to meet the power requirements of different modules. The power module has overvoltage, undervoltage, overcurrent and short-circuit protection functions, and can automatically cut off the power supply in the case of power abnormality to protect the system from damage. In order to improve the stability and anti-interference ability of the power supply, the module adopts electromagnetic compatibility design to reduce the impact of power supply noise and electromagnetic interference on the system. The power monitoring circuit monitors the power supply voltage and current in real time. Once an abnormality is found, it immediately triggers an alarm and records the fault information. The backup power system includes batteries and inverters, which can automatically switch when the main power fails to ensure uninterrupted operation of the system. The power module also has an energy-saving design, which reduces energy consumption and operating costs by optimizing power management and reducing power consumption. The heat dissipation design of the module ensures the stable output of the power supply in a high temperature environment and extends the service life of the power module. In addition, the power module supports remote monitoring function, allowing operators to view power status and fault information in real time through the remote monitoring system, which facilitates remote maintenance and management.

[0161] A method for live detection of electrical tests for a substation is provided, and when used, the live detection system for electrical tests for a substation is operated according to the above-mentioned embodiment.

[0162] In the present invention, the intelligent analysis and diagnosis module significantly improves the accuracy of the live detection system of the electrical test of the booster station in fault prediction and diagnosis by integrating submodules such as data analysis and judgment, model construction, model training, and model evaluation and optimization. The data analysis and judgment module can quickly identify abnormal data in the operation of electrical equipment through threshold alarm and isolated forest anomaly detection algorithm, providing an accurate data basis for subsequent fault diagnosis. The model construction module adopts the advanced artificial intelligence technology of convolutional neural network (CNN), which can deeply mine the feature information in the data, so as to more accurately identify the fault type. The model training module uses historical data and cross entropy loss function, and continuously adjusts the model parameters through the Adam optimization algorithm, so that the model gradually converges during the training process, and improves the accuracy of fault diagnosis. Finally, the model evaluation and optimization module evaluates the model performance through indicators such as accuracy, recall rate, F1 score, and adjusts the model structure or training strategy according to the evaluation results to further optimize the diagnostic effect. Thereby, accurate prediction and diagnosis of electrical equipment faults are achieved, and misdiagnosis or missed diagnosis caused by human judgment errors or improper data processing in traditional methods is effectively avoided. 。

[0163] In the present invention, the intelligent analysis and diagnosis module greatly reduces the need for manual intervention and improves the operating efficiency of the system through automated data analysis and intelligent fault diagnosis. In addition, the module can also monitor the operating status of electrical equipment in real time, detect potential faults in a timely manner and issue warnings, achieve preventive maintenance, and effectively extend the service life of the equipment. At the same time, the module's support for remote monitoring allows operators to monitor and manage the system anytime and anywhere, further improving the convenience and practicality of the system. In general, the intelligent analysis and diagnosis module not only improves the technical level of the live detection system for electrical tests in booster stations in fault prediction and diagnosis, but also promotes the development of the entire system in the direction of intelligence and automation, providing strong technical support for the stable operation and efficient management of booster stations.

[0164] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprises" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0165] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A live detection system for electrical testing of a booster station, characterized in that: The system includes: a data acquisition module, a signal processing module, a communication module, a control module, an intelligent analysis and diagnosis module, a display module and a power supply module; The intelligent analysis and diagnosis module is internally provided with a data analysis and judgment module, a model construction module, a model training module and a model evaluation and optimization module; The data acquisition module first collects the current, voltage and temperature of the electrical equipment in real time through high-precision sensors, and transmits these raw data to the signal processing module; The signal processing module filters, amplifies, and performs analog-to-digital conversion on the data to ensure the accuracy and reliability of the data, and then sends the processed digital signal to the intelligent analysis and diagnosis module; The intelligent analysis and diagnosis module uses artificial intelligence technology to perform deep learning and analysis on data to achieve fault prediction and diagnosis, and transmits the analysis results and diagnosis reports to the control module and the display module; The control module executes the corresponding control strategy according to the analysis results, and feeds back the control instructions to the data acquisition module and the signal processing module; the display module displays the real-time data, analysis results and system status to the operator in the form of a graphical interface, providing intuitive information display and alarm prompts; the communication module is responsible for transmitting the data and analysis results in the system to the remote monitoring center or cloud platform to realize remote monitoring and big data analysis, and at the same time receiving control instructions and update programs from the remote; the power module provides a stable and reliable power supply for the entire system to ensure the normal operation of each module.

2. A booster station electrical test live detection system as claimed in claim 1, characterized in that: The data acquisition module is internally configured with: Sensor unit: including current sensor, voltage sensor, temperature sensor, humidity sensor, used to measure different physical quantities of electrical equipment; Signal conditioning circuit: amplifies, filters, and isolates the weak signal output by the sensor to meet the needs of subsequent circuits; Data acquisition card: realize multi-channel, high sampling rate data acquisition to ensure data integrity and accuracy.

3. A booster station electrical test live detection system as claimed in claim 1, characterized in that: The signal processing module performs in-depth processing and analysis on the original signal output by the data acquisition module to extract useful information; firstly, the signal is filtered through the filtering circuit to remove noise and interference and improve the signal-to-noise ratio of the signal; the filtered signal is amplified by the signal amplification circuit to meet the input requirements of the subsequent analysis circuit; The communication module transmits the data of the signal processing module to the intelligent analysis and diagnosis module, and realizes seamless connection with the remote monitoring system; the communication module provides a variety of communication interfaces, including Ethernet, serial port and wireless communication, and the data transmission protocol of the communication module adopts international standards; The control module performs corresponding control operations according to the analysis results of the signal processing module and the intelligent analysis and diagnosis module to adjust the operating status of the electrical equipment or respond to fault conditions; the control module adopts PID control or fuzzy control to perform real-time control of the electrical equipment according to preset parameters and algorithms; the actuator is the execution unit of the control module, including relays and contactors, which are used to receive control commands and perform corresponding actions.

4. A booster station electrical test live detection system as claimed in claim 1, characterized in that: The data analysis and judgment module first performs a format check and preliminary screening of the booster station electrical test data to ensure the accuracy and completeness of the booster station electrical test data; The threshold alarm system sets the thresholds of each monitoring indicator according to the normal electrical operation range and historical data of the booster station; the current threshold is set to I_min to I_max, the voltage threshold is set to V_min to V_max, and the temperature threshold is set to T_min to T_max; the central processing unit compares the real-time monitoring data with the preset thresholds; if the data exceeds the threshold range, the alarm mechanism is immediately triggered to notify the operation and maintenance personnel; Use the machine learning-based isolation forest anomaly detection algorithm to conduct deep data analysis to identify potential failure modes; Model training: Use historical normal data to train the isolation forest model; model parameters include: n_estimators: the number of isolated trees, which affects the complexity of the model and the detection effect; max_samples: The number of samples used by each tree, set as a fraction of the total number of samples; Contamination: An estimate of the proportion of outliers in the data, used to adjust the sensitivity of the model; Real-time detection: Input the real-time normalized booster station electrical test data into the trained isolation forest model and calculate the anomaly score of each data point; The anomaly score calculation formula is: anomaly_score=2^-(E(h(x)) / c(n)) Anomaly judgment: Determine whether it is an abnormal point based on the anomaly score and the preset anomaly threshold; the higher the anomaly score, the more likely the booster station electrical test data point is to be an abnormal point; where E(h(x)) is the average path length of the data point x in the isolation forest, c(n) is the expected value of the path length, and n is the number of samples.

5. A booster station electrical test live detection system as claimed in claim 1, characterized in that: The model building module selects a convolutional neural network (CNN) as a fault diagnosis algorithm; specifically includes: Input layer: input preprocessed feature data; Convolutional layer: Use multiple convolution kernels to extract local features in the data; Pooling layer: reduces the dimension of the features output by the convolutional layer to reduce the amount of calculation; Fully connected layer: converts the feature vector output by the pooling layer into the probability distribution of the fault type; Output layer: outputs the prediction results of fault type.

6. A booster station electrical test live detection system as claimed in claim 1, characterized in that: The model training module uses the cross entropy loss function to measure the difference between the model prediction result and the true label. The calculation formula is: ,in: H(p,q) represents the cross entropy loss; p(xi) represents the probability of the i-th type of fault in the true label. For one-hot encoding, the probability of the true category is 1 and the rest are 0; q(xi) represents the probability of the i-th type of failure predicted by the model; ∑i represents the sum of all fault categories; The model training module uses the Adam optimization algorithm to update the model parameters and speed up the convergence speed, wherein the calculation process includes: Initialization: Set the initial learning rate α, decay rates β1 and β2, the minimum constant ∈, and the initial value θ of the parameter; Calculate the gradient: Calculate the gradient of the loss function with respect to the parameter θ; Update the first and second moments, where: The formula for calculating the first-order moment momentum is: m t =β1m t-1 +(1-β1)g t The calculation formula for the second-order moment square gradient is: v t =β2v t-1 +(1-β2)gt 2 Among them, g t is the gradient at time step t, m t and v t are the first and second order moments of time step t respectively; Bias correction: bias correction is performed on the first-order moment and the second-order moment to offset the effect of their initial values ​​being 0; The bias correction calculation formula is: In the process of updating parameters, the parameters are updated according to the corrected first-order moment and second-order moment. The calculation formula is: Where α is the learning rate; β1 and β2 are the decay rates, set to 0.9 and 0.999; ∈ is a very small constant to prevent division by 0, set to 1e-8; θ is the model parameter; g t is the gradient at time step t, m t and v t are the first and second order moments of time step t respectively; The model training module divides the entire training data set into multiple small batches during batch training, each batch contains a certain number of samples, and the calculation process includes: S1. Divide into batches: Divide the entire substation electrical training dataset into multiple batches, each batch contains NN samples; S2. Forward propagation: forward propagation of each batch of booster station electrical data to calculate the output of the model; S3. Calculate loss: Based on the model output and the true label, use the cross entropy loss function to calculate the loss of each batch; S4. Back propagation: Back propagate the loss of each batch and calculate the gradient of the parameters; S5. Update parameters: Use the Adam optimization algorithm to update the model parameters according to the calculated gradients; S6. Repeat step: Repeat steps S2-S5 until all batches are traversed or the set number of training rounds is reached.

7. A booster station electrical test live detection system as claimed in claim 1, characterized in that: The model evaluation and optimization module uses accuracy, recall, and F1 score indicators to evaluate model performance; adjusts the model structure, parameters, or training strategy based on the evaluation results to improve diagnostic accuracy; Among them: The calculation formula of the cross entropy loss function is: L=-Σ(y_log(p)+(1-y)_log(1-p)); Among them, y is the true label (0 or 1), and p is the probability predicted by the model; The calculation formula of the Adam optimization algorithm is: m_t=β1*m_(t-1)+(1-β1)*g_t v_t=β2*v_(t-1)+(1-β2)*g_t^2 m_t_hat=m_t / (1-β1^t) v_t_hat=v_t / (1-β2^t) θ_t=θ_(t-1)-α*m_t_hat / (sqrt(v_t_hat)+ε); Among them, m_t and v_t are estimates of the first-order moment and the second-order moment respectively, β1 and β2 are decay rates, g_t is the gradient, α is the learning rate, ε is a small constant, and θ_t is the updated parameter; The convolution layer parameters include: convolution kernel size: used to extract local features; step size: represents the step size of the convolution kernel moving on the data; padding: used to control the size of the output feature map.

8. A booster station electrical test live detection system as claimed in claim 1, characterized in that: The display module displays the system's working status, detection data, and fault information to the operator in an intuitive and easy-to-understand manner; it uses a high-resolution LCD screen that can display clear graphical interfaces and text information; it displays rich content, including real-time data curves, equipment status indications, and fault alarm prompts; When a fault occurs, the display module will immediately pop up an alarm window to display the fault type, location and recommended solutions to help operators quickly respond to and handle the fault.

9. A booster station electrical test live detection system as claimed in claim 1, characterized in that: The power module converts the input AC power into various DC voltages required by the system; the module adopts an electromagnetic compatibility design to reduce the impact of power supply noise and electromagnetic interference on the system; the power monitoring circuit of the power module monitors the power supply voltage and current in real time, and once an abnormality is found, an alarm is immediately triggered and fault information is recorded.

10. A method for detecting live power in a booster station electrical test, characterized in that: When the method is used, the booster station electrical test live detection system as described in any one of claims 1 to 9 is operated.

Images