Transformer substation intelligent monitoring system based on digital spasm and method thereof

Abstract

The invention discloses a transformer substation intelligent monitoring system and method based on digital spasm, and relates to the technical field of digital spasm, the transformer substation intelligent monitoring system comprises a transformer substation information exchange hub, the transformer substation information exchange hub is in communication connection with an operation parameter monitoring module, a fault early warning module and a data visualization module; the parameter monitoring module collects temperature and current data of a transformer substation, the fault early warning module combines a simulation model and historical and real-time monitoring data, an LSTM model is used for early warning a fault, the data visualization module displays performance, trend and alarm, an LSTM is called to predict a load, and an algorithm is used for optimizing energy consumption, cost and load distribution. Intelligent monitoring of the transformer substation is achieved through the digital spasm technology, the accuracy and the real-time performance of monitoring are improved, the fault early warning precision and the load prediction capacity are improved through the LSTM model, meanwhile, the energy consumption, the cost and the load distribution are optimized through the algorithm, and the operation efficiency and the economic benefits of the transformer substation are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twins, and in particular to a substation intelligent monitoring system and method based on digital twins. Background Art

[0002] The substation monitoring system is a comprehensive security management system designed for users in the power industry. It combines industrial control, security management, and digital video technologies, utilizes the existing network resources of the power grid, and integrates multifunctional subsystems such as remote viewing system, access control system, fire protection system, environment and power monitoring system. The system can monitor the operating status and parameters of various power equipment in the substation in real time, including voltage, current, power, temperature, etc., as well as the environmental conditions of the substation, such as oxygen content, humidity, etc. This information is collected through data acquisition devices and sensors and transmitted to the monitoring center for processing and analysis. Once an abnormal situation or fault is detected, the system will immediately issue an alarm signal and notify the operator to handle it. In addition, the substation monitoring system also has a remote control operation function, which can remotely dispatch and control the equipment in the substation and improve the automation level of the power system. At the same time, the system also provides a friendly human-computer interaction interface, which is convenient for operation and maintenance personnel to view equipment status, operate equipment, and query data.

[0003] In order to solve the problem that the accuracy and real-time performance of fault warning and diagnosis in substation monitoring systems need to be improved, the existing technology uses traditional threshold judgment and simple data analysis to process. However, this method often relies on fixed warning thresholds and manually set rules, which makes it difficult to accurately capture the complex characteristics and potential risks of substation equipment failures, and may also result in false alarms, missed alarms, and delayed warnings. As a result, operation and maintenance personnel are unable to obtain accurate fault information in a timely manner, and it is difficult to take quick and effective countermeasures, which not only increases the risk of power grid operation, but may also cause serious consequences such as equipment damage and power outages. In order to overcome this defect, a substation intelligent monitoring system and method based on digital twins are proposed. Summary of the invention

[0004] The purpose of the present invention is to provide a substation intelligent monitoring system and method based on digital twin to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the technical solutions adopted by the present invention are: in the first aspect, a substation intelligent monitoring system based on digital twins includes a substation information exchange hub, and the substation information exchange hub is communicatively connected with an operation parameter monitoring module, a fault warning module and a data visualization module;

[0006] The operating parameter monitoring module collects temperature data on the surface of the substation equipment through a temperature sensor and collects current data in the substation circuit through a current transformer;

[0007] The fault warning module introduces a physical simulation model of the substation, simulates the behavior characteristics of each component in the substation, combines historical data and real-time monitoring data to evaluate the health status of substation equipment, builds an LSTM model, and realizes substation fault warning;

[0008] The data visualization module displays the main performance indicators, trend charts and alarm information of the substation through a visualization interface, calls the LSTM model to realize the substation load prediction, minimizes energy consumption and cost with the help of mixed integer linear programming algorithm, and adjusts the substation load distribution in real time in combination with the reinforcement learning algorithm.

[0009] A further improvement of the technical solution of the present invention is that: in the operating parameter monitoring module, the process of collecting temperature data on the surface of the substation equipment through the temperature sensor includes:

[0010] Temperature sensors are deployed on the surfaces of transformers, circuit breakers, switch cabinets, busbars, cable joints, and capacitor banks. The temperature sensors sense the temperature changes on the surface of the equipment and convert the physical quantity into an electrical signal.

[0011] Preprocessing the data obtained from the temperature sensor, wherein the preprocessing operation includes removing high-frequency noise in the temperature data by using a low-pass filter and a sliding average filter technique, filling missing values ​​by using linear interpolation, and performing Z-score normalization processing;

[0012] Extracting features from the preprocessed temperature data, the feature extraction process includes calculating the average value, maximum and minimum values, skewness, kurtosis and autocorrelation coefficient to extract time domain features, evaluating the temperature change trend, applying wavelet transform to extract frequency domain features, decomposing the time domain signal into frequency components, capturing periodic and non-periodic components to reflect the working status of the equipment;

[0013] The temperature data is packaged and segmented, lossless compression technology is used to reduce the data volume, encryption protocol and identity authentication are used, and the data packets are sent to the fault warning module through the TCP / IP transmission protocol, and the log of each successful transmission is recorded.

[0014] A further improvement of the technical solution of the present invention is that: in the operating parameter monitoring module, the process of collecting current data in the substation circuit through the current transformer includes:

[0015] Current transformers are deployed in the input and output circuits of the main transformer, feeder circuits, busbar connection points, circuit breakers, inside switch cabinets, and capacitor bank circuits. Based on the principle of electromagnetic induction, current transformers proportionally reduce the large current on the primary side to a small current on the secondary side. When current flows through the primary circuit, a magnetic field is generated in the iron core of the current transformer, which induces a corresponding current in the secondary winding.

[0016] Preprocessing the data obtained from the temperature sensor, wherein the preprocessing operation includes removing high-frequency noise in the current data by using a low-pass filter and a sliding average filter technique, filling missing values ​​by using linear interpolation, and performing Z-score normalization processing;

[0017] Extracting features from the preprocessed current data, wherein the feature extraction process includes calculating the average value, the maximum and minimum values, and the variance to describe the trend of the current change over time, applying wavelet transform to extract frequency domain features from the time domain signal, capturing the periodic and non-periodic components in the current change, and reflecting the working state of the circuit;

[0018] The current data is packaged into data packets containing current information and timestamps, encoded, compressed and encrypted, and sent to the fault warning module.

[0019] A further improvement of the technical solution of the present invention is that: in the fault warning module, a physical simulation model of a substation is introduced, and the process of simulating the behavior characteristics of each component in the substation includes:

[0020] Constructing a physical simulation model of a substation, wherein the physical simulation model of the substation includes a transformer simulation model, a circuit breaker simulation model and a comprehensive power system simulation model;

[0021] The transformer simulation model uses Maxwell's equations and heat conduction equations to simulate the changes in transformer winding, core and oil temperature. The transformer equipment surface temperature data and the current data of the main transformer input and output circuits are input into the simulation model and compared with the results predicted by the transformer simulation model. By analyzing the temperature and current change trends, the load capacity and health status of the transformer are evaluated. The calculation process is as follows:

[0022] Heat conduction equation:

[0023] Law of Electromagnetic Induction:

[0024] Where T is temperature, t is time, α is thermal diffusion coefficient, is the Laplace operator, Q(x,t) represents the heat source term, V is the induced electromotive force, L is the self-inductance coefficient, and I is the current;

[0025] The circuit breaker simulation model builds a dynamic model of the arc between the circuit breaker contacts, inputs the circuit breaker equipment surface temperature data and the current data inside the circuit breaker into the simulation model, compares them with the results predicted by the circuit breaker simulation model, monitors the circuit breaker's operating temperature, predicts the arc formation and extinction process at the moment of opening and closing, and optimizes the design parameters of the circuit breaker. The calculation process is as follows:

[0026] Arc voltage between contacts: U arc =R arc I;

[0027] Among them, U arc is the arc voltage, R arc is the arc resistance;

[0028] The transformer simulation model and the circuit breaker simulation model are integrated to form a comprehensive power system simulation model. The model uses the collected temperature data and current data to evaluate the load distribution, stability and accident response at the whole network level, optimize the load distribution strategy, simulate the response of the substation intelligent monitoring system under extreme conditions, evaluate the stability of the power grid, and conduct regular accident rehearsals and test protection measures.

[0029] A further improvement of the technical solution of the present invention is that: in the fault warning module, the process of evaluating the health status of substation equipment by combining historical data and real-time monitoring data includes:

[0030] If the temperature sensor shows that the transformer surface temperature continues to rise and exceeds the predicted value of the transformer simulation model, it is considered that the internal windings and core of the transformer are overheated, resulting in an internal overload of the transformer. If the temperature sensor shows that the transformer surface temperature is lower than the predicted value of the transformer simulation model, it is considered that the low temperature causes the fluidity of the transformer oil to deteriorate, and an alarm is issued to remind the operation and maintenance personnel to use a heating device to make the transformer operate normally. If the current transformer detects that the current of the input and output circuits of the main transformer exceeds the predicted value of the transformer simulation model, it is considered that the transformer is in an overloaded state and there is a risk of heating, which may cause aging and damage to the insulation material. The substation intelligent monitoring system generates an alarm, reduces the load and checks whether there is a short circuit. If the current transformer detects that the current of the input and output circuits of the main transformer is lower than the predicted value of the transformer simulation model, it is considered that the transformer has insufficient load and open circuit problems.

[0031] In the circuit breaker simulation model, if the temperature sensor shows that the circuit breaker surface temperature exceeds the value predicted by the circuit breaker simulation model, it is considered that the circuit breaker has problems with contact wear and poor contact, and the increased resistance causes heat. The substation intelligent monitoring system generates an alarm to remind the operation and maintenance personnel to replace the contacts. If the temperature sensor shows that the circuit breaker surface temperature is lower than the value predicted by the circuit breaker simulation model, it is considered that the low temperature causes the fluidity of the circuit breaker oil to deteriorate, and an alarm is issued to remind the operation and maintenance personnel to use a heating device to make the circuit breaker operate normally. If the current transformer detects that the internal current of the circuit breaker exceeds the value predicted by the circuit breaker simulation model, it is considered that the circuit is open, and the operation and maintenance personnel are reminded to check the circuit immediately to prevent the circuit breaker from overheating and damage. If the current transformer detects that the internal current of the circuit breaker is lower than the value predicted by the circuit breaker simulation model, it is considered that the circuit breaker has problems with insufficient load and open circuit.

[0032] In the integrated power system simulation model, if the temperature sensor shows that the surface temperature of the switchgear, busbar, cable joint and capacitor group equipment exceeds the predicted value of the integrated power system simulation model, it is considered that the substation has problems of excessive load and local short circuit. If the temperature sensor shows that the surface temperature of the switchgear, busbar, cable joint and capacitor group equipment is lower than the predicted value of the integrated power system simulation model, it is considered that the low temperature causes the conductor to shrink and the lubricant to solidify. If the current transformer detects that the feeder loop current exceeds the predicted value of the integrated power system simulation model, it is considered that the feeder loop is in an overloaded state. If the current transformer detects that the feeder loop current is lower than the predicted value of the integrated power system simulation model, it is considered that the feeder loop has problems of insufficient load and open circuit.

[0033] A further improvement of the technical solution of the present invention is that in the fault warning module, the process of training the LSTM model includes:

[0034] Constructing an LSTM model, wherein the LSTM model includes an input layer, a hidden layer, and an output layer;

[0035] The input layer receives the temperature and current time series data after preprocessing and feature extraction and the prediction results of the physical simulation model, and divides the time series data and the prediction results into a time window containing the past T time steps and the temperature, current and prediction result features, forming a multidimensional array X = [x1, x2, ..., x t ], where x t Represents the data vector at the t-th time step. Each feature is normalized to make the numerical ranges of different features consistent. The calculation process is as follows:

[0036]

[0037] Among them, x′ t represents the normalized eigenvalue, μ represents the eigenmean, and σ represents the standard deviation;

[0038] Map the data vector of each time window to the input space of the LSTM model to form a three-dimensional tensor X input ∈R T ×F , where T represents the time window length, F represents the number of features, and for each new time window, the hidden state h0 and cell state C0 of the LSTM unit are initialized to zero vectors;

[0039] The hidden layer uses the gating mechanism in its internal long short-term memory cell state to capture its long-term dependencies over time, and the gating mechanism includes a forget gate, an input gate, a cell state update, and an output gate;

[0040] The forget gate is based on the historical hidden state h t-1 and the current time step x t The input is used to evaluate and decide whether to retain the past state information. If the temperature data and current data fluctuate abnormally, the forget gate chooses to retain the historical abnormal information. If the temperature data and current data change regularly, the forget gate chooses to discard the historical abnormal information. The calculation process is as follows:

[0041] f t =σ(W f ·[h t-1 ;x t ]+b f );

[0042] Among them, x t represents the input vector of the current time step, including temperature, current, and prediction results of the physical simulation model, h t-1 represents the output state of the previous time step, W f is the weight matrix corresponding to the forget gate, b f is the bias term, σ is the sigmoid activation function, which maps the result to between 0 and 1, indicating the degree of retention and discarding. t If f is close to 1, the information is retained. t If it is close to 0, the information is discarded;

[0043] The input gate evaluates the input x at the current time step t Whether to include temperature and current mutations and allow the changes to enter cell state C t , the calculation process is as follows:

[0044] i t =σ(W i ·[h t-1 ,x t ]+b i );

[0045]

[0046] Among them, i t Represents the input factor, using the Sigmoid activation function to determine whether each element is added to the cell state. The Tanh activation function is used to generate candidate cell states ranging from -1 to 1, W i and b i Determines the selection mechanism of the input gate, W C and b C Controls the generation of candidate cell states;

[0047] Combining the results of the forget gate and the input gate, the cell state at the previous moment is bitwise multiplied by the output of the forget gate, and then added to the candidate state generated by the input gate. If there is a sudden change in temperature and current, the abnormal situation is reflected by updating the cell state that stores long-term information. The calculation process is as follows:

[0048]

[0049] The output gate extracts and outputs the information related to the temperature and current mutation from the cell state as the hidden state h t , evaluate the working status of each device in the substation. If the temperature and current data of the device are normal, the output gate transfers the hidden state to the next time step. The calculation process is as follows:

[0050] o t =σ(W o ·[h t-1 ,x t ]+b o );

[0051] h t =o t ⊙tanh(C t );

[0052] Among them, t is the output factor, and the Sigmoid activation function is used to determine whether each element is output. Tanh(C t ) makes the output value fall between -1 and 1, reflecting the change of the current cell state;

[0053] For each time step t in each time window, the calculation process of the forget gate, input gate and output gate is repeated to gradually update the hidden state h t and cell state C t ,capture the dynamic changes in the time series;

[0054] The output layer receives the hidden state h from the hidden layer output t, it is mapped to the target output dimension through the fully connected layer inside the output layer, and the probability of substation failure is output. By minimizing the binary cross entropy activation function, the difference between the predicted probability and the actual fault label is measured, and the LSTM model parameters are adjusted. The calculation process is as follows:

[0055] P (故障) =σ(W y ·h(t)+b y );

[0056]

[0057] Among them, P (故障) represents the probability of failure occurring at time t, W y represents the weight matrix connecting the hidden state to the output layer, b y is the bias term, N is the number of samples, y(t) is the actual fault label, if a fault occurs, y(t) = 1, if no fault occurs, y(t) = 0, L is the loss function value, which reflects the average difference between the predicted probability and the actual fault label.

[0058] A further improvement of the technical solution of the present invention is that in the fault warning module, based on the LSTM model, the process of realizing the substation fault warning includes:

[0059] If the calculated fault probability value reaches 0.8, it is considered that a fault is about to occur, and the operation and maintenance personnel are reminded to check and handle it in time. If the fault probability value is lower than 0.8, it is considered that the substation is in normal working condition. The simulation model is combined for correction. The temperature curve predicted by the transformer simulation model established based on the heat conduction equation is used to learn the transformer temperature change pattern. The current characteristics under standard operating conditions generated by the arc dynamic model are used to learn the circuit breaker current change pattern.

[0060] A further improvement of the technical solution of the present invention is that in the data visualization module, the process of displaying the main performance indicators, trend charts and alarm information of the substation through the visualization interface includes:

[0061] Build a visual interface to display the surface temperature and loop current values ​​of substation equipment in real time, and compare them with the prediction results of the physical simulation model;

[0062] Observe the periodic and non-periodic components of temperature data, current data and their time and frequency domain characteristics through trend charts, evaluate the accuracy of the prediction results of the physical simulation model, monitor the load distribution over a long time span, and assist in optimizing the power grid operation strategy;

[0063] When the failure probability reaches 0.8, the visualization interface displays an alarm prompt including the specific fault location, type and severity, and proposes countermeasures.

[0064] A further improvement of the technical solution of the present invention is that in the data visualization module, the LSTM model is called to realize the substation load prediction, the mixed integer linear programming algorithm is used to minimize the energy consumption and cost, and the process of real-time adjustment of the substation load distribution in combination with the reinforcement learning algorithm includes:

[0065] The long short-term memory network model is called to capture long-term dependencies based on historical temperature and current data and physical simulation model prediction results to realize substation load forecasting, evaluate the equipment operating status, and use the mixed integer linear programming algorithm to define the objective function and constraints, which include load balance constraints, equipment capacity limitations, and operating rule constraints. The constraints minimize energy consumption and costs while meeting load demand, solve the optimal load distribution plan, introduce reinforcement learning algorithms for real-time adjustments, optimize scheduling decisions based on the current power grid status and historical experience, and improve scheduling strategies through cumulative reward mechanisms.

[0066] In the second aspect, a substation intelligent monitoring method based on digital twin is used to implement the above-mentioned substation intelligent monitoring system based on digital twin, and is composed of the following steps:

[0067] S1. Collect temperature data on the surface of substation equipment through temperature sensors;

[0068] S2, collecting current data in the substation circuit through current transformer;

[0069] S3. Introduce a physical simulation model of the substation to simulate the behavior characteristics of each component in the substation, and combine historical data and real-time monitoring data to evaluate the health status of substation equipment;

[0070] S4, build LSTM model to realize substation fault warning;

[0071] S5. Display the main performance indicators, trend charts and alarm information of the substation through a visual interface;

[0072] S6. Call the LSTM model to realize substation load forecasting. With the help of mixed integer linear programming algorithm, minimize energy consumption and cost while meeting load demand, and combine reinforcement learning algorithm to adjust substation load distribution in real time.

[0073] Due to the adoption of the above technical solution, the present invention has the following technical advances compared with the prior art:

[0074] 1. The present invention provides a substation intelligent monitoring system and method based on digital twins, which significantly improves the accuracy and real-time performance of fault warning. By introducing the substation physical simulation model and LSTM model, the system can simulate the behavioral characteristics of each component in the substation, and conduct a comprehensive analysis based on historical data and real-time monitoring data to achieve an accurate assessment of the health status of the equipment, thereby timely warning of potential faults and reducing the risk of equipment damage and power outages.

[0075] 2. The present invention provides a substation intelligent monitoring system and method based on digital twins, which enhances data visualization and intelligent management capabilities. The system uses a data visualization module to display the main performance indicators, trend charts, and alarm information of the substation to operation and maintenance personnel in an intuitive way, so that they can quickly understand the operation status of the power grid. At the same time, the LSTM model is combined for load forecasting, and the mixed integer linear programming algorithm and reinforcement learning algorithm are used to optimize energy consumption and cost and adjust load distribution, thereby improving the efficiency and economy of power grid operation.

[0076] 3. The present invention provides a substation intelligent monitoring system and method based on digital twins, which improves the efficiency and quality of operation and maintenance work. The system realizes real-time monitoring and intelligent diagnosis of substation equipment, provides comprehensive fault information and diagnosis results for operation and maintenance personnel, and reduces the workload of manual inspection and data analysis. At the same time, accurate fault warning and load prediction capabilities enable operation and maintenance personnel to formulate response measures in advance, improving the initiative and flexibility of operation and maintenance work. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0078] Figure 1 It is a block diagram of the intelligent monitoring system for substation based on digital twin of the present invention;

[0079] Figure 2 The present invention is a flow chart of the intelligent monitoring method for substation based on digital twin. DETAILED DESCRIPTION

[0080] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0081] Embodiment 1, as Figure 1 As shown, the present invention provides a substation intelligent monitoring system based on digital twin, including a substation information exchange hub, wherein the substation information exchange hub is communicatively connected with an operation parameter monitoring module, a fault warning module and a data visualization module;

[0082] The operating parameter monitoring module collects temperature data on the surface of substation equipment through temperature sensors, collects current data in substation circuits through current transformers, and deploys temperature sensors on the surfaces of transformers, circuit breakers, switch cabinets, busbars, cable joints and capacitor banks. The temperature sensors convert physical quantities into electrical signals by sensing temperature changes on the surface of the equipment, and preprocess the data obtained from the temperature sensors. The preprocessing operation includes removing high-frequency noise in the temperature data through low-pass filters and sliding average filtering techniques, filling missing values ​​with linear interpolation, and performing Z-score standardization processing. Feature extraction is performed on the preprocessed temperature data. The feature extraction process includes calculating the average value, maximum and minimum values, skewness, kurtosis and autocorrelation coefficient to extract time domain features, evaluate temperature change trends, apply wavelet transform to extract frequency domain features, decompose time domain signals into frequency components, capture periodic and non-periodic components to reflect the working status of the equipment, package and segment the temperature data, use lossless compression technology to reduce the amount of data, use encryption protocols and identity authentication, and send data packets to faults through TCP / IP transmission protocols. Fault warning module, records the log of each successful transmission, deploys current transformers in the input and output circuits of the main transformer, feeder circuits, busbar connection points, circuit breakers, inside the switch cabinet and capacitor bank circuits. Based on the principle of electromagnetic induction, the current transformer reduces the large current on the primary side to the small current on the secondary side in proportion. When the current flows through the primary circuit, a magnetic field is generated in the iron core of the current transformer, and the magnetic field induces a corresponding current in the secondary winding. The data obtained from the temperature sensor is preprocessed. The preprocessing operation includes removing high-frequency noise in the current data through a low-pass filter and a sliding average filter technology, using linear interpolation to fill missing values, and performing Z-score standardization. Feature extraction is performed from the preprocessed current data. The feature extraction process includes calculating the average value, maximum and minimum values, and variance to describe the trend of current changes over time. Wavelet transform is applied to extract frequency domain features from time domain signals to capture periodic and non-periodic components in current changes to reflect the working status of the circuit. The current data is packaged into a data packet containing current information and a timestamp, and encoded, compressed and encrypted, and sent to the fault warning module;

[0083] The fault warning module introduces a physical simulation model of a substation, simulates the behavior characteristics of each component in the substation, combines historical data and real-time monitoring data to evaluate the health status of substation equipment, builds an LSTM model, realizes substation fault warning, and builds a physical simulation model of a substation. The substation physical simulation model includes a transformer simulation model, a circuit breaker simulation model, and a comprehensive power system simulation model. The transformer simulation model uses Maxwell's equations and heat conduction equations to simulate the changes in transformer windings, cores, and oil temperatures. The transformer equipment surface temperature data and the current data of the main transformer input and output circuits are input into the simulation model and compared with the results predicted by the transformer simulation model. By analyzing the temperature and current change trends, the load capacity and health status of the transformer are evaluated. The calculation process is as follows:

[0084] Heat conduction equation:

[0085] Law of Electromagnetic Induction:

[0086] Where T is temperature, t is time, α is thermal diffusion coefficient, is the Laplace operator, Q(x, t) represents the heat source term, V is the induced electromotive force, L is the self-inductance coefficient, and I is the current. The circuit breaker simulation model builds a dynamic model of the arc between the contacts of the circuit breaker. The surface temperature data of the circuit breaker equipment and the current data inside the circuit breaker are input into the simulation model and compared with the results predicted by the circuit breaker simulation model. The operating temperature of the circuit breaker is monitored, the arc formation and extinction process at the moment of opening and closing is predicted, and the design parameters of the circuit breaker are optimized. The calculation process is as follows:

[0087] Arc voltage between contacts: U arc =R arc I;

[0088] Among them, U arc is the arc voltage, R arcFor arc resistance, the transformer simulation model and the circuit breaker simulation model are integrated to form a comprehensive power system simulation model. The model uses the collected temperature data and current data to evaluate the load distribution, stability and accident response at the whole network level, optimize the load distribution strategy, simulate the response of the substation intelligent monitoring system under extreme conditions, evaluate the stability of the power grid, and conduct accident rehearsals regularly to test protection measures. If the temperature sensor shows that the surface temperature of the transformer continues to rise and exceeds the predicted value of the transformer simulation model, it is considered that the internal winding and core of the transformer are overheated, resulting in internal overload of the transformer. If the temperature sensor shows that the surface temperature of the transformer is lower than the predicted value of the transformer simulation model, it is considered that the low temperature causes the fluidity of the transformer oil to deteriorate, and an alarm is issued to remind the operation and maintenance personnel to use the heating device The transformer is configured to operate normally. If the current transformer detects that the current of the input and output circuits of the main transformer exceeds the predicted value of the transformer simulation model, the transformer is considered to be in an overloaded state and there is a risk of heating, which will cause aging and damage to the insulation material. The substation intelligent monitoring system generates an alarm, reduces the load and checks whether there is a short circuit. If the current transformer detects that the current of the input and output circuits of the main transformer is lower than the predicted value of the transformer simulation model, the transformer is considered to have insufficient load and open circuit problems. In the circuit breaker simulation model, if the temperature sensor shows that the surface temperature of the circuit breaker exceeds the predicted value of the circuit breaker simulation model, the circuit breaker is considered to have contact wear and poor contact problems, and the increased resistance causes heating. The substation intelligent monitoring system generates an alarm to remind the operation and maintenance personnel to change Change contacts. If the temperature sensor shows that the surface temperature of the circuit breaker is lower than the predicted value of the circuit breaker simulation model, it is considered that the low temperature causes the fluidity of the circuit breaker oil to deteriorate, and an alarm is issued to remind the operation and maintenance personnel to use a heating device to make the circuit breaker operate normally. If the current transformer detects that the internal current of the circuit breaker exceeds the predicted value of the circuit breaker simulation model, it is considered that the circuit is open, and the operation and maintenance personnel are reminded to check the circuit immediately to prevent the circuit breaker from overheating and damage. If the current transformer detects that the internal current of the circuit breaker is lower than the predicted value of the circuit breaker simulation model, it is considered that the circuit breaker has insufficient load and open circuit problems. In the comprehensive power system simulation model, if the temperature sensor shows that the surface temperature of the switch cabinet, busbar, cable joint and capacitor group equipment exceeds the predicted value of the comprehensive power system simulation model, it is considered that the switch cabinet, busbar, cable joint and capacitor group equipment are under-loaded and under-loaded. The power station has problems of overload and local short circuit. If the temperature sensor shows that the surface temperature of the switch cabinet, busbar, cable joint and capacitor group equipment is lower than the predicted value of the comprehensive power system simulation model, it is considered that the low temperature causes the conductor to shrink and the lubricant to solidify. If the current transformer detects that the feeder loop current exceeds the predicted value of the comprehensive power system simulation model, it is considered that the feeder loop is in an overload state. If the current transformer detects that the feeder loop current is lower than the predicted value of the comprehensive power system simulation model, it is considered that the feeder loop has insufficient load and circuit open problems. An LSTM model is established. The LSTM model includes an input layer, a hidden layer and an output layer. The input layer receives the temperature and current time series data after preprocessing and feature extraction and the prediction results of the physical simulation model.The time series data and prediction results are divided into time windows containing the past T time steps and the temperature, current and prediction result features, forming a multidimensional array X = [x1, x2, ..., x, t ], where x t Represents the data vector at the t-th time step. Each feature is normalized to make the numerical ranges of different features consistent. The calculation process is as follows:

[0089]

[0090] Among them, x′ t represents the normalized eigenvalue, μ represents the feature mean, and σ represents the standard deviation. The data vector of each time window is mapped to the input space of the LSTM model to form a three-dimensional tensor X input ∈R T×F , where T represents the time window length, F represents the number of features, and for each new time window, the hidden state h0 and cell state C0 of the LSTM unit are initialized to zero vectors. The hidden layer uses the gating mechanism in its internal long short-term memory unit state to capture its long-term dependencies over time. The gating mechanism includes a forget gate, an input gate, a cell state update, and an output gate. The forget gate is based on the historical hidden state h t-1 and the current time step x t The input is used to evaluate and decide whether to retain the past state information. If the temperature data and current data fluctuate abnormally, the forget gate chooses to retain the historical abnormal information. If the temperature data and current data change regularly, the forget gate chooses to discard the historical abnormal information. The calculation process is as follows:

[0091] f t =σ(W f ·[h t-1 ;x t ]+b f );

[0092] Among them, x t represents the input vector of the current time step, including temperature, current, and prediction results of the physical simulation model, h t-1 represents the output state of the previous time step, W f is the weight matrix corresponding to the forget gate, b f is the bias term, σ is the sigmoid activation function, which maps the result to between 0 and 1, indicating the degree of retention and discarding. t If f is close to 1, the information is retained. t If it is close to 0, the information is discarded and the input gate evaluates the input x at the current time step. t Whether to include temperature and current mutations and allow the changes to enter cell state C t , the calculation process is as follows:

[0093] i t =σ(W i ·[h t-1 ,x t ]+b i );

[0094]

[0095] Among them, i t Represents the input factor, using the Sigmoid activation function to determine whether each element is added to the cell state. The Tanh activation function is used to generate candidate cell states ranging from -1 to 1, W i and b i Determines the selection mechanism of the input gate, W C and b C Control the generation of candidate cell states, combine the results of the forget gate and the input gate, multiply the cell state at the previous moment by the output of the forget gate, and add the candidate state generated by the input gate. If there is a sudden change in temperature and current, the abnormal situation is reflected by updating the cell state that stores long-term information. The calculation process is as follows:

[0096]

[0097] The output gate extracts and outputs the information related to the temperature and current mutation from the cell state as the hidden state h t , evaluate the working status of each device in the substation. If the temperature and current data of the device are normal, the output gate transfers the hidden state to the next time step. The calculation process is as follows:

[0098] o t =σ(W o ·[h t-1 ,x t ]+b o );

[0099] h t =o t ⊙tanh(C t );

[0100] Among them, t is the output factor, and the Sigmoid activation function is used to determine whether each element is output. Tanh(C t ) makes the output value fall between -1 and 1, reflecting the change of the current cell state. For each time step t in each time window, the calculation process of the forget gate, input gate and output gate is repeated to gradually update the hidden state h t and cell state C t, capturing the dynamic changes in the time series, the output layer receives the hidden state h from the hidden layer output t , it is mapped to the target output dimension through the fully connected layer inside the output layer, and the probability of substation failure is output. By minimizing the binary cross entropy activation function, the difference between the predicted probability and the actual fault label is measured, and the LSTM model parameters are adjusted. The calculation process is as follows:

[0101] P (故障) =σ(W y ·h(t)+b y );

[0102]

[0103] Among them, P (故障) represents the probability of failure occurring at time t, W y represents the weight matrix connecting the hidden state to the output layer, b y is the bias term, N is the number of samples, y(t) is the actual fault label, if a fault occurs, y(t) = 1, if no fault occurs, y(t) = 0, L is the loss function value, reflecting the average difference between the predicted probability and the actual fault label, if the calculated fault probability value reaches 0.8, it is considered that a fault is about to occur, and the operation and maintenance personnel are reminded to check and handle it in time, if the fault probability value is lower than 0.8, it is considered that the substation is in normal working condition, and correction is performed in combination with the simulation model. The temperature curve predicted by the transformer simulation model established based on the heat conduction equation is used to learn the transformer temperature change mode. The current characteristics under standard operating conditions generated by the arc dynamic model are used to learn the circuit breaker current change mode. The data visualization module displays the main performance indicators, trend charts and alarm information of the substation through a visual interface, calls the LSTM model to realize the substation load prediction, minimizes energy consumption and cost with the help of a mixed integer linear programming algorithm, and adjusts the substation load distribution in real time in combination with a reinforcement learning algorithm.

[0104] Embodiment 2, as Figure 2 As shown, on the basis of Example 1, the present invention further provides a technical solution: a substation intelligent monitoring method based on digital twin, which is used to realize a substation intelligent monitoring system based on digital twin, and is composed of the following steps:

[0105] S1. Collect temperature data on the surface of substation equipment through temperature sensors;

[0106] S2, collecting current data in the substation circuit through current transformer;

[0107] S3. Introduce a physical simulation model of the substation to simulate the behavior characteristics of each component in the substation, and combine historical data and real-time monitoring data to evaluate the health status of substation equipment;

[0108] S4, build LSTM model to realize substation fault warning;

[0109] S5. Display the main performance indicators, trend charts and alarm information of the substation through a visual interface;

[0110] S6. Call the LSTM model to realize substation load forecasting. With the help of mixed integer linear programming algorithm, minimize energy consumption and cost while meeting load demand, and combine reinforcement learning algorithm to adjust substation load distribution in real time.

[0111] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. The intelligent monitoring system for substations based on digital twins is characterized by: It includes a substation information exchange hub, which is communicatively connected to an operation parameter monitoring module, a fault warning module and a data visualization module; The operating parameter monitoring module collects temperature data on the surface of the substation equipment through a temperature sensor and collects current data in the substation circuit through a current transformer; The fault warning module introduces a physical simulation model of the substation, simulates the behavior characteristics of each component in the substation, combines historical data and real-time monitoring data to evaluate the health status of substation equipment, builds an LSTM model, and realizes substation fault warning; The data visualization module displays the main performance indicators, trend charts and alarm information of the substation through a visualization interface, calls the LSTM model to realize the substation load prediction, minimizes energy consumption and cost with the help of mixed integer linear programming algorithm, and adjusts the substation load distribution in real time in combination with the reinforcement learning algorithm.

2. The intelligent monitoring system for substations based on digital twin according to claim 1 is characterized in that: In the operating parameter monitoring module, the process of collecting temperature data on the surface of substation equipment through a temperature sensor includes: Temperature sensors are deployed on the surfaces of transformers, circuit breakers, switch cabinets, busbars, cable joints, and capacitor banks. The temperature sensors sense the temperature changes on the surface of the equipment and convert the physical quantity into an electrical signal. Preprocessing the data obtained from the temperature sensor, wherein the preprocessing operation includes removing high-frequency noise in the temperature data by using a low-pass filter and a sliding average filter technique, filling missing values ​​by using linear interpolation, and performing Z-score normalization processing; Extracting features from the preprocessed temperature data, the feature extraction process includes calculating the average value, maximum and minimum values, skewness, kurtosis and autocorrelation coefficient to extract time domain features, evaluating the temperature change trend, applying wavelet transform to extract frequency domain features, decomposing the time domain signal into frequency components, capturing periodic and non-periodic components to reflect the working status of the equipment; The temperature data is packaged and segmented, lossless compression technology is used to reduce the data volume, encryption protocol and identity authentication are used, and the data packets are sent to the fault warning module through the TCP / IP transmission protocol, and the log of each successful transmission is recorded.

3. The intelligent monitoring system for substations based on digital twin according to claim 2 is characterized in that: In the operating parameter monitoring module, the process of collecting current data in the substation circuit through the current transformer includes: Current transformers are deployed in the input and output circuits of the main transformer, feeder circuits, busbar connection points, circuit breakers, inside switch cabinets, and capacitor bank circuits. Based on the principle of electromagnetic induction, current transformers proportionally reduce the large current on the primary side to a small current on the secondary side. When current flows through the primary circuit, a magnetic field is generated in the iron core of the current transformer, which induces a corresponding current in the secondary winding. Preprocessing the data obtained from the temperature sensor, wherein the preprocessing operation includes removing high-frequency noise in the current data by using a low-pass filter and a sliding average filter technique, filling missing values ​​by using linear interpolation, and performing Z-score normalization processing; Extracting features from the preprocessed current data, wherein the feature extraction process includes calculating the average value, the maximum and minimum values, and the variance to describe the trend of the current change over time, applying wavelet transform to extract frequency domain features from the time domain signal, capturing the periodic and non-periodic components in the current change, and reflecting the working state of the circuit; The current data is packaged into data packets containing current information and timestamps, encoded, compressed and encrypted, and sent to the fault warning module.

4. The intelligent monitoring system for substations based on digital twin according to claim 3 is characterized in that: In the fault warning module, a physical simulation model of a substation is introduced to simulate the behavior characteristics of each component in the substation, including: Constructing a physical simulation model of a substation, wherein the physical simulation model of the substation includes a transformer simulation model, a circuit breaker simulation model and a comprehensive power system simulation model; The transformer simulation model uses Maxwell's equations and heat conduction equations to simulate the changes in transformer winding, core and oil temperature. The transformer equipment surface temperature data and the current data of the main transformer input and output circuits are input into the simulation model and compared with the results predicted by the transformer simulation model. By analyzing the temperature and current change trends, the load capacity and health status of the transformer are evaluated. The calculation process is as follows: Heat conduction equation: Law of Electromagnetic Induction: Where T is temperature, t is time, α is thermal diffusion coefficient, is the Laplace operator, Q(x,t) represents the heat source term, V is the induced electromotive force, L is the self-inductance coefficient, and I is the current; The circuit breaker simulation model builds a dynamic model of the arc between the circuit breaker contacts, inputs the circuit breaker equipment surface temperature data and the current data inside the circuit breaker into the simulation model, compares them with the results predicted by the circuit breaker simulation model, monitors the circuit breaker's operating temperature, predicts the arc formation and extinction process at the moment of opening and closing, and optimizes the design parameters of the circuit breaker. The calculation process is as follows: Arc voltage between contacts: U arc =R arc I; Among them, U arc is the arc voltage, R arc is the arc resistance; The transformer simulation model and the circuit breaker simulation model are integrated to form a comprehensive power system simulation model. The model uses the collected temperature data and current data to evaluate the load distribution, stability and accident response at the whole network level, optimize the load distribution strategy, simulate the response of the substation intelligent monitoring system under extreme conditions, evaluate the stability of the power grid, and conduct regular accident rehearsals and test protection measures.

5. The intelligent monitoring system for substations based on digital twin according to claim 4 is characterized in that: In the fault warning module, the process of evaluating the health status of substation equipment by combining historical data and real-time monitoring data includes: If the temperature sensor shows that the transformer surface temperature continues to rise and exceeds the predicted value of the transformer simulation model, it is considered that the internal windings and core of the transformer are overheated, resulting in an internal overload of the transformer. If the temperature sensor shows that the transformer surface temperature is lower than the predicted value of the transformer simulation model, it is considered that the low temperature causes the fluidity of the transformer oil to deteriorate, and an alarm is issued to remind the operation and maintenance personnel to use a heating device to make the transformer operate normally. If the current transformer detects that the current of the input and output circuits of the main transformer exceeds the predicted value of the transformer simulation model, it is considered that the transformer is in an overloaded state and there is a risk of heating, which may cause aging and damage to the insulation material. The substation intelligent monitoring system generates an alarm, reduces the load and checks whether there is a short circuit. If the current transformer detects that the current of the input and output circuits of the main transformer is lower than the predicted value of the transformer simulation model, it is considered that the transformer has insufficient load and open circuit problems. In the circuit breaker simulation model, if the temperature sensor shows that the circuit breaker surface temperature exceeds the value predicted by the circuit breaker simulation model, it is considered that the circuit breaker has problems with contact wear and poor contact, and the increased resistance causes heat. The substation intelligent monitoring system generates an alarm to remind the operation and maintenance personnel to replace the contacts. If the temperature sensor shows that the circuit breaker surface temperature is lower than the value predicted by the circuit breaker simulation model, it is considered that the low temperature causes the fluidity of the circuit breaker oil to deteriorate, and an alarm is issued to remind the operation and maintenance personnel to use a heating device to make the circuit breaker operate normally. If the current transformer detects that the internal current of the circuit breaker exceeds the value predicted by the circuit breaker simulation model, it is considered that the circuit is open, and the operation and maintenance personnel are reminded to check the circuit immediately to prevent the circuit breaker from overheating and damage. If the current transformer detects that the internal current of the circuit breaker is lower than the value predicted by the circuit breaker simulation model, it is considered that the circuit breaker has problems with insufficient load and open circuit. In the integrated power system simulation model, if the temperature sensor shows that the surface temperature of the switchgear, busbar, cable joint and capacitor group equipment exceeds the predicted value of the integrated power system simulation model, it is considered that the substation has problems of excessive load and local short circuit. If the temperature sensor shows that the surface temperature of the switchgear, busbar, cable joint and capacitor group equipment is lower than the predicted value of the integrated power system simulation model, it is considered that the low temperature causes the conductor to shrink and the lubricant to solidify. If the current transformer detects that the feeder loop current exceeds the predicted value of the integrated power system simulation model, it is considered that the feeder loop is in an overloaded state. If the current transformer detects that the feeder loop current is lower than the predicted value of the integrated power system simulation model, it is considered that the feeder loop has problems of insufficient load and open circuit.

6. The intelligent monitoring system for substations based on digital twin according to claim 5 is characterized in that: In the fault warning module, the process of training the LSTM model includes: Constructing an LSTM model, wherein the LSTM model includes an input layer, a hidden layer, and an output layer; The input layer receives the temperature and current time series data after preprocessing and feature extraction and the prediction results of the physical simulation model, and divides the time series data and the prediction results into a time window containing the past T time steps and the temperature, current and prediction result features, forming a multidimensional array X = [x1, x2, ..., x t ], where x t Represents the data vector at the t-th time step. Each feature is normalized to make the numerical ranges of different features consistent. The calculation process is as follows: Among them, x′ t represents the normalized eigenvalue, μ represents the eigenmean, and σ represents the standard deviation; Map the data vector of each time window to the input space of the LSTM model to form a three-dimensional tensor X input ∈R T×F , where T represents the time window length, F represents the number of features, and for each new time window, the hidden state h0 and cell state C0 of the LSTM unit are initialized to zero vectors; The hidden layer uses the gating mechanism in its internal long short-term memory cell state to capture its long-term dependencies over time, and the gating mechanism includes a forget gate, an input gate, a cell state update, and an output gate; The forget gate is based on the historical hidden state h t-1 and the current time step x t The input is used to evaluate and decide whether to retain the past state information. If the temperature data and current data fluctuate abnormally, the forget gate chooses to retain the historical abnormal information. If the temperature data and current data change regularly, the forget gate chooses to discard the historical abnormal information. The calculation process is as follows: f t =σ(W f ·[h t-1 ;x t ]+b f ); Among them, x t represents the input vector of the current time step, including temperature, current, and prediction results of the physical simulation model, h t-1 represents the output state of the previous time step, W f is the weight matrix corresponding to the forget gate, b f is the bias term, σ is the sigmoid activation function, which maps the result to between 0 and 1, indicating the degree of retention and discarding. t If f is close to 1, the information is retained. t If it is close to 0, the information is discarded; The input gate evaluates the input x at the current time step t Whether to include temperature and current mutations and allow the changes to enter cell state C t , the calculation process is as follows: i t =σ(W i ·[h t-1 ,x t ]+b i ); Among them, i t Represents the input factor, using the Sigmoid activation function to determine whether each element is added to the cell state. The Tanh activation function is used to generate candidate cell states ranging from -1 to 1, W i and b i Determines the selection mechanism of the input gate, W C and b C Controls the generation of candidate cell states; Combining the results of the forget gate and the input gate, the cell state at the previous moment is bitwise multiplied by the output of the forget gate, and then added to the candidate state generated by the input gate. If there is a sudden change in temperature and current, the abnormal situation is reflected by updating the cell state that stores long-term information. The calculation process is as follows: The output gate extracts and outputs the information related to the temperature and current mutation from the cell state as the hidden state h t , evaluate the working status of each device in the substation. If the temperature and current data of the device are normal, the output gate transfers the hidden state to the next time step. The calculation process is as follows: the t =σ(W o ·[h t-1 ,x t ]+b o ); h t =o t ⊙tanh(C t ); Among them, t is the output factor, and the Sigmoid activation function is used to determine whether each element is output. Tanh(C t ) makes the output value fall between -1 and 1, reflecting the change of the current cell state; For each time step t in each time window, the calculation process of the forget gate, input gate and output gate is repeated to gradually update the hidden state h t and cell state C t ,capture the dynamic changes in the time series; The output layer receives the hidden state h from the hidden layer output t , it is mapped to the target output dimension through the fully connected layer inside the output layer, and the probability of substation failure is output. By minimizing the binary cross entropy activation function, the difference between the predicted probability and the actual fault label is measured, and the LSTM model parameters are adjusted. The calculation process is as follows: P (故障) =σ(W y ·h(t)+b y ); Among them, P (故障) represents the probability of failure occurring at time t, W y represents the weight matrix connecting the hidden state to the output layer, b y is the bias term, N is the number of samples, y(t) is the actual fault label, if a fault occurs, y(t) = 1, if no fault occurs, y(t) = 0, L is the loss function value, which reflects the average difference between the predicted probability and the actual fault label.

7. The intelligent monitoring system for substations based on digital twin according to claim 6 is characterized in that: In the fault warning module, based on the LSTM model, the process of realizing the substation fault warning includes: If the calculated fault probability value reaches 0.8, it is considered that a fault is about to occur, and the operation and maintenance personnel are reminded to check and handle it in time. If the fault probability value is lower than 0.8, it is considered that the substation is in normal working condition. The simulation model is combined for correction. The temperature curve predicted by the transformer simulation model established based on the heat conduction equation is used to learn the transformer temperature change pattern. The current characteristics under standard operating conditions generated by the arc dynamic model are used to learn the circuit breaker current change pattern.

8. The intelligent monitoring system for substations based on digital twin according to claim 7 is characterized in that: In the data visualization module, the process of displaying the main performance indicators, trend charts and alarm information of the substation through the visualization interface includes: Build a visual interface to display the surface temperature and loop current values ​​of substation equipment in real time, and compare them with the prediction results of the physical simulation model; Observe the periodic and non-periodic components of temperature data, current data and their time and frequency domain characteristics through trend charts, evaluate the accuracy of the prediction results of the physical simulation model, monitor the load distribution over a long time span, and assist in optimizing the power grid operation strategy; When the failure probability reaches 0.8, the visualization interface displays an alarm prompt including the specific fault location, type and severity, and proposes countermeasures.

9. The intelligent monitoring system for substations based on digital twin according to claim 8 is characterized in that: In the data visualization module, the LSTM model is called to realize substation load prediction, and the mixed integer linear programming algorithm is used to minimize energy consumption and cost. The process of combining the reinforcement learning algorithm to adjust the substation load distribution in real time includes: The long short-term memory network model is called to capture long-term dependencies based on historical temperature and current data and physical simulation model prediction results to realize substation load forecasting, evaluate the equipment operating status, and use the mixed integer linear programming algorithm to define the objective function and constraints, which include load balance constraints, equipment capacity limitations, and operating rule constraints. The constraints minimize energy consumption and costs while meeting load demand, solve the optimal load distribution plan, introduce reinforcement learning algorithms for real-time adjustments, optimize scheduling decisions based on the current power grid status and historical experience, and improve scheduling strategies through cumulative reward mechanisms.

10. A method for intelligent monitoring of a substation based on digital twin, implemented based on the intelligent monitoring system for a substation based on digital twin as claimed in any one of claims 1 to 9, characterized in that: It consists of the following steps: S1. Collect temperature data on the surface of substation equipment through temperature sensors; S2, collecting current data in the substation circuit through current transformer; S3. Introduce a physical simulation model of the substation to simulate the behavior characteristics of each component in the substation, and combine historical data and real-time monitoring data to evaluate the health status of substation equipment; S4, build LSTM model to realize substation fault warning; S5. Display the main performance indicators, trend charts and alarm information of the substation through a visual interface; S6. Call the LSTM model to realize substation load forecasting. With the help of mixed integer linear programming algorithm, minimize energy consumption and cost while meeting load demand, and combine reinforcement learning algorithm to adjust substation load distribution in real time.

Images