Intelligent box-type substation monitoring method based on digital twinning

Abstract

The invention relates to the technical field of intelligent power grids, in particular to an intelligent box-type substation monitoring method based on digital twinning. According to the method, a digital twinborn scene of the box-type substation is built through a three-dimensional model obtained through three-dimensional modeling, a deep learning algorithm model is integrated, and real-time monitoring and future state prediction of operation data of the box-type substation are achieved. And through an improved iTransform model, predicting the operation state of each hardware in a future preset time, and performing real-time simulation display in a digital twinning scene. And automatically triggering a risk early warning mechanism when the risk of abnormal operation, fault or exceeding a preset safety threshold value is predicted. Besides, a remote monitoring center is established, digital twinborn scenes and real-time monitoring data of a plurality of box-type substations are managed in a centralized manner, mobile terminal access is supported, and the response speed and the decision-making efficiency are improved. According to the method, the monitoring efficiency and the fault prediction accuracy are remarkably improved, and a powerful guarantee is provided for safe and stable operation of an intelligent power grid.

Description

Technical Field

[0001] The present invention relates to the field of smart grid technology, and in particular to a smart box-type substation monitoring method based on digital twins. Background Art

[0002] As an important part of the power system, the operating status of box-type substations directly affects the stability and security of the power grid. Traditional box-type substation monitoring methods mostly rely on manual inspections and regular testing, which have problems such as low monitoring efficiency, slow response speed, and inaccurate fault prediction. With the development of the Internet of Things, big data, and artificial intelligence technologies, how to realize intelligent monitoring of box-type substations and improve monitoring efficiency and fault prediction accuracy has become a problem that needs to be solved urgently. Summary of the invention

[0003] In view of the above-mentioned technical deficiencies, the purpose of the present invention is to provide an intelligent box-type substation monitoring method based on digital twins to solve the problem of low intelligence of box-type substation monitoring in the prior art.

[0004] In order to solve the above technical problems, the present invention adopts the following technical solutions: First, the monitoring method of intelligent box-type substation based on digital twin includes: Step S100: performing three-dimensional modeling on the box-type substation to obtain a three-dimensional model of the box-type substation, and building a digital twin scene of the box-type substation based on the three-dimensional model of the box-type substation; Step S200: acquiring the operating data of the box-type substation in real time through sensors in the box-type substation; Step S300: inputting the operating data of the box-type substation into the algorithm model based on deep learning for analysis and processing, and predicting the operating status of each hardware of the box-type substation within a preset time in the future; Step S400: In the digital twin scenario, simulate and display the operating status of each hardware of the box-type substation in real time and the operating status of each hardware of the box-type substation within a preset time in the future; Step S500: When it is predicted that a hardware of the box-type substation may experience a risk of abnormal operation, failure, or exceeding a preset safety threshold within a preset time in the future, a risk warning mechanism is automatically triggered.

[0005] Preferably, in a possible implementation manner of the first aspect, the operating data of the box-type substation includes: partial discharge monitoring data, SF6 monitoring data, lightning arrester monitoring data, wireless temperature monitoring data, rainfall data, temperature and humidity data, smoke sensor data, water immersion data, air conditioning monitoring data, lighting monitoring data and fan monitoring data.

[0006] Preferably, in a possible implementation of the first aspect, the deep learning-based algorithm model adopts an improved iTransformer model and is trained using historical box-type substation operation data, wherein the historical box-type substation operation data includes: historical partial discharge monitoring data, historical SF6 monitoring data, historical lightning arrester monitoring data, historical wireless temperature monitoring data, historical rainfall data, historical temperature and humidity data, historical smoke sensor data, historical flooding data, historical air conditioning monitoring data, historical lighting monitoring data, historical fan monitoring data and historical box-type substation hardware operation data.

[0007] Preferably, in a possible implementation manner of the first aspect, it is characterized in that the improved iTransformer model structure includes: Embedding layer: used to learn variables Sequential feature identification of , independently aggregating the global features of each variable:

[0008] in represents the entire time series of the same variable, represents the hardware operation data of the box-type substation, D represents the time series, N represents the variable series, Indicates the time series changes of the corresponding variables in the past time; Self-attention layer: used to model the correlation of different variables, and analyze each position of the attention map using the correlation of different variables:

[0009] Among them, Q represents the query vector, K represents the key vector, is the dimension of the key vector, and The query vector and key vector corresponding to any two variables, and the query vector with high activity is retained; Feedforward neural network layer: Use feedforward network to encode word vectors and learn the amplitude, periodicity and frequency spectrum of any time series; Normalization layer: acts on the time series changes to make the features of all variables in a uniform distribution. The calculation formula of the normalization layer is:

[0010] in, represents the eigenvalue of the corresponding variable, represents the mean of the eigenvalues, represents the variance of the eigenvalue; Mapping layer: Integrate and map the learned feature vectors into the final prediction results.

[0011] Preferably, in a possible implementation of the first aspect, the query vectors retaining high activity are specifically: selecting a preset number of key vectors in each sequence, calculating the inner product of each query vector and the preset number of key vectors, obtaining a preset number of scores for each query vector, selecting the highest score and the average to obtain a difference score, selecting a preset number of query vectors according to the difference score, and replacing the remaining query vectors with the average of the numerical vectors.

[0012] Preferably, in a possible implementation manner of the first aspect, step S400 specifically includes: Receive real-time box-type substation operation data and prediction data obtained based on deep learning algorithm models; Render the hardware of the three-dimensional model of the box-type substation in the digital twin scene and display the operating status of each hardware in real time; When the mouse clicks on a piece of hardware, the simulation will show the running status of the hardware within a preset time in the future.

[0013] Preferably, in a possible implementation manner of the first aspect, the step of displaying the operating status of each hardware in real time includes: reading the operating data of the real-time box-type substation according to a first preset time interval, establishing corresponding scripts for different hardware and mounting them on the hardware to display their operating status; The step of displaying the operating status of the hardware within a preset future time includes: reading the prediction data obtained by the algorithm model based on deep learning according to a second preset time interval, establishing corresponding scripts for different hardware and mounting them on the hardware, triggering an event when the mouse is clicked, and displaying the operating status of the hardware within a preset future time.

[0014] Preferably, in a possible implementation of the first aspect, the risk warning mechanism in step S500 includes generating a processing plan according to the prediction result, and notifying relevant personnel by SMS, email or system notification according to a preset priority.

[0015] Preferably, in a possible implementation manner of the first aspect, the method further includes establishing a remote monitoring center for the box-type substation, and centrally managing the digital twin scenes and real-time monitoring data of multiple box-type substations through a cloud platform.

[0016] Preferably, in a possible implementation manner of the first aspect, the remote monitoring center supports mobile terminal access, and management personnel view the operating status of the box-type substation, receive warning information and processing suggestions through terminal devices.

[0017] The beneficial effect of the present invention is that through the combination of three-dimensional modeling and deep learning algorithms, not only the intuitiveness and accuracy of intelligent box-type substation monitoring are improved, but also the future operating status can be accurately predicted, effectively avoiding the occurrence of faults. At the same time, the method supports user interaction, improves monitoring flexibility, and establishes a remote monitoring center to achieve centralized management and mobile access, further improving response speed and decision-making efficiency. In short, the method significantly improves monitoring efficiency and fault prediction accuracy, providing a strong guarantee for the safe and stable operation of the smart grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 A flow chart of a method for monitoring intelligent box-type substations based on digital twins is provided for this application. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0021] Embodiment 1: Figure 1 As shown, the present invention provides a monitoring method for an intelligent box-type substation based on digital twins, comprising: Step S100: Perform three-dimensional modeling on the box-type substation to obtain a three-dimensional model of the box-type substation, and build a digital twin scene of the box-type substation based on the three-dimensional model of the box-type substation.

[0022] Specifically, 3ds Max is used to perform three-dimensional modeling of the box-type substation to obtain the fbx format model of the box-type substation. The model is imported into Unity 3D to build a digital twin scene of the box-type substation.

[0023] Import the fbx model file exported from 3ds Max into the Unity 3D project; adjust the position, rotation, and scale of the model in the Inspector panel of Unity 3D to ensure its correct display in the scene. Add light sources to simulate natural light, set a skybox or background image to provide an environmental background for the scene, and add terrain or ground to simulate the placement environment of the box-type substation.

[0024] Use Unity 3D's scripting capabilities to add interactivity, including opening and closing doors, rotating equipment, and dynamically running equipment. Write logic to handle user input to enable interaction with models. Use LOD technology to reduce the details of distant models to improve rendering efficiency. Adjust the quality of lighting and shadows to reduce the computational burden while maintaining visual effects.

[0025] Step S200: acquiring the operating data of the box-type substation in real time through sensors in the box-type substation.

[0026] Specifically, the sensors in the box-type substation include: Partial discharge sensor: used to monitor partial discharge phenomena in equipment inside box-type substations and detect potential insulation faults in a timely manner.

[0027] SF6 gas concentration sensor: used to monitor the concentration of SF6 gas and ensure the normal operation of SF6 gas insulation equipment.

[0028] Arrester leakage current sensor: used to monitor the arrester leakage current and evaluate the arrester performance and service life.

[0029] Wireless temperature sensor: used to monitor the temperature of internal equipment in box-type substations, such as transformers, cables, etc., to detect overheating in a timely manner.

[0030] Rainfall sensor: used to monitor the rainfall in the area where the box-type substation is located and evaluate the impact of rainfall on the operation of the substation.

[0031] Temperature and humidity sensor: used to monitor the temperature and humidity inside the box-type substation to ensure that the equipment operates in a suitable environment.

[0032] Smoke sensor: used to monitor the smoke concentration inside the box-type substation and detect fire hazards in time.

[0033] Water immersion sensor: used to monitor whether there is water immersion inside the box-type substation to prevent equipment damage and safety accidents caused by water immersion.

[0034] Air conditioning status sensor: used to monitor the operating status and parameters of air conditioning equipment, such as temperature, humidity, etc., to ensure the normal operation of the air conditioning system.

[0035] Light intensity sensor or lighting equipment status sensor: used to monitor the light intensity and lighting equipment status inside the box-type substation to ensure the normal operation of the lighting system.

[0036] Fan status sensor: used to monitor the operating status and parameters of the fan equipment, such as speed, vibration, temperature, etc., to ensure the normal operation of the fan system.

[0037] The operating data obtained through the sensors in the box-type substation include: partial discharge monitoring data, SF6 monitoring data, lightning arrester monitoring data, wireless temperature monitoring data, rainfall data, temperature and humidity data, smoke sensor data, water immersion data, air conditioning monitoring data, lighting monitoring data and fan monitoring data.

[0038] In this embodiment, a communication management machine is used as a data acquisition and transmission device in the box-type substation. The sensor transmits the operating data to the communication management machine through the RS-485 interface, and configures the corresponding communication parameters (including baud rate, data bit, stop bit, and check bit, etc.) on the communication management machine to ensure smooth communication with the sensor. In addition, the communication management machine integrates an optical fiber communication module, which is used to transmit the operating data of the box-type substation to the remote control center through an optical fiber dedicated data channel. When configuring the communication parameters, the communication management machine is assigned a primary and backup IP address. At the same time, it is also necessary to configure the IP address of the remote control center to establish a stable network connection. During the data transmission process, the TCP protocol is used to ensure the integrity and sequentiality of data transmission to ensure that the remote control center can accurately receive the operating data of the box-type substation.

[0039] Step S300: Input the operating data of the box-type substation into the algorithm model based on deep learning for analysis and processing, and predict the operating status of each hardware of the box-type substation within a preset time in the future.

[0040] Specifically, the algorithm model based on deep learning adopts the improved iTransformer model and is trained using historical box-type substation operation data. The historical box-type substation operation data include: historical partial discharge monitoring data, historical SF6 monitoring data, historical lightning arrester monitoring data, historical wireless temperature monitoring data, historical rainfall data, historical temperature and humidity data, historical smoke sensor data, historical flooding data, historical air conditioning monitoring data, historical lighting monitoring data, historical fan monitoring data and historical box-type substation hardware operation data.

[0041] The improved iTransformer model structure includes: Embedding layer: used to learn variables Sequential feature identification of , independently aggregating the global features of each variable:

[0042] in, represents the entire time series of the same variable, represents the hardware operation data of the box-type substation, D represents the time series, N represents the variable series, Indicates the time series changes of the corresponding variable in the past.

[0043] Self-attention layer: used to model the correlation of different variables, and analyze each position of the attention map using the correlation of different variables:

[0044] Among them, Q represents the query vector, K represents the key vector, is the dimension of the key vector, and The query vector and key vector corresponding to any two variables, and the query vector with high activity is retained.

[0045] Feedforward neural network layer: Use feedforward networks to encode word vectors and learn the amplitude, periodicity, and frequency spectrum of any time series.

[0046] Normalization layer: acts on the time series changes to make the features of all variables in a uniform distribution. The calculation formula of the normalization layer is:

[0047] in, represents the eigenvalue of the corresponding variable, represents the mean of the eigenvalues, Represents the variance of the eigenvalue.

[0048] Mapping layer: Integrate and map the learned feature vectors into the final prediction results.

[0049] In this embodiment, missing values ​​or outliers are first removed from the historical box-type substation operation data, and the data is standardized or normalized to eliminate the impact of dimensions between different variables; the data is then divided into a training set, a validation set, and a test set for model training, validation, and testing; and the data of each variable is arranged in time series to form time series data.

[0050] The time series data of each variable is embedded to learn the sequence feature identifier of the variable. The embedded vector contains the time series change information of the variable in the past.

[0051] The attention mechanism is used to model the correlation between different variables. For any two variables, the inner product of their query vector and key vector is calculated to obtain the attention score. 25 key vectors are selected in each sequence, and the inner product of each query vector and key vector is calculated to obtain 25 scores for each query vector. The highest score and the average of each query vector are selected to obtain the difference score. The top 25 query vectors with the largest difference are selected according to the difference score, and the remaining query vectors are replaced by the average of the numerical vectors, thereby enhancing the model's attention to key information and reducing the amount of calculation.

[0052] The embedded vectors are further encoded using a feedforward neural network, which not only learns the amplitude, periodicity, and frequency spectrum of any time series, but also captures the complex features in the time series, including trend changes and abnormal fluctuations, thereby improving the ability to predict the operating status of various hardware in the box-type substation within a preset time in the future.

[0053] Next, the output of the feedforward neural network layer is normalized to ensure that the features of all variables are in a relatively uniform distribution, reducing the deviation between features and improving the stability of model training and the accuracy of prediction results.

[0054] Finally, the learned feature vectors are integrated and mapped to the final prediction results. The feature vectors are mapped to the predicted hardware operating status using a fully connected layer, thereby achieving accurate prediction of the future operating status of each hardware in the box-type substation. In this embodiment, the preset time is two hours, that is, the operating status of each hardware in the box-type substation is predicted within the next two hours.

[0055] Step S400: In the digital twin scenario, simulate and display the operating status of each hardware of the box-type substation in real time and the operating status of each hardware of the box-type substation within a preset time in the future.

[0056] Receive real-time box-type substation operation data and prediction data obtained based on deep learning algorithm models Specifically, the backend service first connects to the database and executes SQL queries to obtain data predicted by the deep learning algorithm model. The data is formatted into a unified structure and provided to Unity3D through the RESTful API interface. Through WebSocket technology, a long connection is established between Unity 3D and the backend service to push updated data in real time. Unity 3D requests this data regularly or on demand through asynchronous programming mode, and uses a data caching mechanism to reduce the number of network requests and reduce network latency.

[0057] The hardware of the three-dimensional model of the box-type substation in the digital twin scene is rendered, and the operating status of each hardware is displayed in real time; when the mouse clicks on a certain hardware, the operating status of the hardware within the preset time in the future is simulated and displayed.

[0058] Specifically, the real-time box-type substation operation data is read at a first preset time interval, and corresponding scripts are established for different hardware and mounted on the hardware to display its operation status; the step of displaying the operation status of the hardware within a preset time in the future includes: reading the prediction data obtained by the algorithm model based on deep learning at a second preset time interval, and corresponding scripts are established for different hardware and mounted on the hardware, and an event is triggered when the mouse is clicked to display the operation status of the hardware within a preset time in the future. In this embodiment, the first preset time interval is 0.5 seconds, and the second preset time interval is 10 minutes.

[0059] In Unity 3D, the material and shader systems are used to give visual effects to each hardware model in the digital twin, and the complexity and performance of the scene are balanced by optimizing rendering settings and using LOD technology. At the same time, colliders are added to each hardware model, and C# scripts are written to listen for mouse click events. When a piece of hardware is clicked, the script identifies the clicked hardware and sends a request to the backend server to obtain the predicted operating status data of the hardware in the next two hours.

[0060] Based on the received prediction data, the Unity 3D application dynamically updates the state of the hardware model, including visual effects such as color, brightness, and texture, by modifying the model's Shader properties or applying animations. In order to simulate the operating state of the hardware in the future, animation clips are created according to the type and function of the hardware, and these animation clips are dynamically switched or mixed using the Animator Controller.

[0061] Throughout the entire process, the asynchronous programming model is used to avoid blocking the main thread and ensure the smooth operation of the application. At the same time, data compression technology is used to reduce the data size and reduce network transmission time. In addition, a network monitoring function is implemented to detect network delays and connection status, and adjust data request strategies as needed.

[0062] Step S500: When it is predicted that a hardware of a box-type substation may have a risk of abnormal operation, failure, or exceeding a preset safety threshold in the future, a risk warning mechanism is automatically triggered.

[0063] Set operating parameter thresholds for each hardware of the box-type substation. In this embodiment, the following thresholds are included: the threshold for partial discharge monitoring data is 5000PC; the threshold for SF6 monitoring data is 1000ppm; the threshold for lightning arrester monitoring data is 150mA; the threshold for wireless temperature monitoring data is 85℃ for busbar temperature and 75℃ for connection point temperature; the threshold for smoke sensor data is 2.5~5μg / m3; When the prediction results of the deep learning-based algorithm model show that the operating parameters of a certain hardware may exceed the set threshold, it is judged that there is a risk of abnormal operation or failure. Sound alarms, visual alarms and information push are used to ensure that relevant personnel can receive early warning information in a timely manner.

[0064] According to the severity of the risk, warnings are divided into different levels, such as minor, moderate and severe. Different levels of warnings correspond to different warning signals and processing procedures.

[0065] Minor warning: When the prediction results show that the operating parameters of a certain hardware are close to but have not exceeded the threshold, a minor warning is triggered. At this time, the system only issues an audible or visual alarm to remind relevant personnel to pay attention and check the status of the equipment.

[0066] Moderate warning: When the prediction results show that the operating parameters of a certain hardware are about to exceed the threshold, or have slightly exceeded but have not yet caused serious consequences, a moderate warning is triggered. At this time, in addition to sound and visual alarms, the system will also send warning information to the mobile devices of relevant personnel through information push so that they can understand and take measures in time.

[0067] Serious warning: When the prediction results show that the operating parameters of a certain hardware have seriously exceeded the threshold, or may cause serious consequences such as equipment failure or shutdown, a serious warning is triggered. At this time, the system not only emits a strong sound and visual alarm, but also sends warning information multiple times through information push, and automatically triggers emergency processing procedures, including equipment shutdown, power cut-off, etc., to prevent the fault from spreading further.

[0068] Each time an early warning is triggered, information such as the time, level, triggering cause, and processing results of the early warning are recorded for subsequent analysis and improvement.

[0069] In one embodiment, the method further includes establishing a remote monitoring center for the box-type substation, and using cloud platform technology to centrally manage the digital twin scene and its real-time monitoring data. As a unified data center, the cloud platform receives sensor data from each box-type substation, and processes, stores and analyzes it. Through the cloud platform, the operating status of each substation can be viewed in real time, including partial discharge monitoring data, SF6 monitoring data, lightning arrester monitoring data, wireless temperature monitoring data, rainfall data, temperature and humidity data, smoke sensor data, flooding data, air conditioning monitoring data, lighting monitoring data and fan monitoring data.

[0070] In addition, the remote monitoring center has a mobile access function. Managers can access the remote monitoring center anytime and anywhere through mobile phones, tablets and other terminal devices to check the operating status of the box-type substation. Once an abnormal situation occurs in the substation, such as equipment failure, environmental monitoring data exceeding the standard, etc., the cloud platform will immediately send an early warning message to the manager and give corresponding treatment suggestions. Managers can take prompt measures based on the early warning information to ensure the safe operation of the substation.

[0071] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. The intelligent box-type substation monitoring method based on digital twin is characterized by: include: Step S100: performing three-dimensional modeling on the box-type substation to obtain a three-dimensional model of the box-type substation, and building a digital twin scene of the box-type substation based on the three-dimensional model of the box-type substation; Step S200: acquiring the operating data of the box-type substation in real time through sensors in the box-type substation; Step S300: inputting the operating data of the box-type substation into the algorithm model based on deep learning for analysis and processing, and predicting the operating status of each hardware of the box-type substation within a preset time in the future; Step S400: In the digital twin scenario, simulate and display the operating status of each hardware of the box-type substation in real time and the operating status of each hardware of the box-type substation within a preset time in the future; Step S500: When it is predicted that a hardware of the box-type substation may experience a risk of abnormal operation, failure, or exceeding a preset safety threshold within a preset time in the future, a risk warning mechanism is automatically triggered.

2. The intelligent box-type substation monitoring method according to claim 1, characterized in that: The operating data of the box-type substation includes: partial discharge monitoring data, SF6 monitoring data, lightning arrester monitoring data, wireless temperature monitoring data, rainfall data, temperature and humidity data, smoke sensor data, flooding data, air conditioning monitoring data, lighting monitoring data and fan monitoring data.

3. The intelligent box-type substation monitoring method according to claim 1, characterized in that: The deep learning-based algorithm model adopts an improved iTransformer model and is trained using historical box-type substation operation data. The historical box-type substation operation data includes: historical partial discharge monitoring data, historical SF6 monitoring data, historical lightning arrester monitoring data, historical wireless temperature monitoring data, historical rainfall data, historical temperature and humidity data, historical smoke sensor data, historical flooding data, historical air conditioning monitoring data, historical lighting monitoring data, historical fan monitoring data and historical box-type substation hardware operation data.

4. The intelligent box-type substation monitoring method according to claim 3, characterized in that: The improved iTransformer model structure includes: Embedding layer, used to learn variables Sequential feature identification of , independently aggregating the global features of each variable: in, represents the entire time series of the same variable, represents the hardware operation data of the box-type substation, D represents the time series, N represents the variable series, Indicates the time series changes of the corresponding variables in the past time; The self-attention layer is used to model the correlation of different variables and analyze each position of the attention map using the correlation of different variables: Among them, Q represents the query vector, K represents the key vector, is the dimension of the key vector, and The query vector and key vector corresponding to any two variables, and the query vector with high activity is retained; Feedforward neural network layer, which uses feedforward network to encode word vectors and learn the amplitude, periodicity and frequency spectrum of any time series; The normalization layer acts on the time series changes to make the characteristics of all variables in a uniform distribution. The calculation formula of the normalization layer is: in, represents the eigenvalue of the corresponding variable, represents the mean of the eigenvalues, represents the variance of the eigenvalue; The mapping layer integrates and maps the learned feature vectors into the final prediction results.

5. The intelligent box-type substation monitoring method according to claim 4, characterized in that: The method of retaining highly active query vectors is as follows: a preset number of key vectors are selected in each sequence, the inner product of each query vector and the preset number of key vectors is calculated to obtain a preset number of scores for each query vector, the highest score and the average score are selected to obtain a difference score, a preset number of query vectors are selected according to the difference score, and the remaining query vectors are replaced by the average of the numerical vectors.

6. The intelligent box-type substation monitoring method according to claim 1, characterized in that: The step S400 specifically includes: Receive real-time box-type substation operation data and prediction data obtained based on deep learning algorithm models; Render the hardware of the three-dimensional model of the box-type substation in the digital twin scene and display the operating status of each hardware in real time; When the mouse clicks on a piece of hardware, the simulation will show the running status of the hardware within a preset time in the future.

7. The intelligent box-type substation monitoring method according to claim 6, characterized in that: The step of displaying the operating status of each hardware in real time includes: reading the operating data of the real-time box-type substation according to a first preset time interval, establishing corresponding scripts for different hardware and mounting them on the hardware to display their operating status; The step of displaying the operating status of the hardware within a preset future time includes: reading the prediction data obtained by the algorithm model based on deep learning according to a second preset time interval, establishing corresponding scripts for different hardware and mounting them on the hardware, triggering an event when the mouse is clicked, and displaying the operating status of the hardware within a preset future time.

8. The intelligent box-type substation monitoring method according to claim 1, characterized in that: The risk warning mechanism in step S500 includes generating a processing plan based on the prediction results, and notifying relevant personnel through SMS, email or system notification according to the preset priority.

9. The intelligent box-type substation monitoring method according to claim 1, characterized in that: The method also includes establishing a remote monitoring center for the box-type substation, and centrally managing the digital twin scenes and real-time monitoring data of multiple box-type substations through a cloud platform.

10. The intelligent box-type substation monitoring method according to claim 9, characterized in that: The remote monitoring center supports mobile terminal access, and managers can view the operating status of the box-type substation, receive warning information and processing suggestions through terminal devices.

Images