Business expansion digital twin model establishment method and device based on hybrid drive

Abstract

The invention belongs to the technical field of electric power, and provides a business expansion digital twin model establishment method and device based on hybrid drive. Comprising the steps of collecting multivariate data, preprocessing the multivariate data based on an improved fractional order flexible space method, constructing a data driving model based on an improved TCN-BiGRU method, obtaining a hybrid driving model in combination with an improved semi-invariant method and the data driving model, constructing a dynamic model of the digital twin court, simulating and deducing a business expansion scheme and the like. According to the invention, a digital twinning technology is utilized to construct a power system model with real-time monitoring and optimization capabilities. The model can accurately simulate the power grid state of an active transformer area, analyze core parameter characteristics under different business expansion schemes, predict the influence of new energy access on power grid operation, and provide more accurate and flexible scheduling and decision basis for power grid management.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric power, and particularly relates to a method and device for establishing a digital twin model for business expansion and installation based on hybrid drive. Background Art

[0002] Traditional power grid load forecasting and power dispatching methods often fail to cope with the complexity of new energy access. Especially for large-scale new energy installation and business expansion requirements, there is a lack of effective comprehensive evaluation and optimization tools. In addition, the power system often relies on static models and planning methods based on historical data during actual operation, and cannot dynamically reflect the real-time changes in the grid state. Therefore, a dynamic and real-time updated business expansion and installation model constructed based on digital twin technology can effectively simulate the power system state of the substation area, predict the impact of new energy access, and provide more accurate data support for power grid dispatching and investment decisions. Summary of the Invention

[0003] To solve the above technical problems, the present invention provides a method for establishing a digital twin model for business expansion and installation based on hybrid drive to solve the problems in the prior art. The technical solution adopted by the present invention is as follows:

[0004] A method for establishing a digital twin model for business expansion and installation based on hybrid drive includes the following steps:

[0005] Step 1: Collect multivariate data, where the multivariate data includes: static data, dynamic data, and environmental data;

[0006] Step 2: Preprocess the multivariate data based on the improved fractional-order flexible space method, including: converting the multivariate data into an m*n matrix, where k is the type of data, to obtain the original high-dimensional multivariate data feature set m is the data dimension, mapping the high-dimensional multivariate data to low-dimensional multivariate data, establishing a multivariate local neighborhood graph and a multivariate local objective function, and optimizing to obtain the typical features of the multivariate data:

[0007] Define the multivariate local proximity weight matrix as:

[0008]

[0009] Where: is the multivariate proximity weight matrix, represents the neighborhood of length r of, t is the time series; is the jth group of data in the kth multivariate data;

[0010] Obtain the minimized multivariate local proximity objective function as:

[0011]

[0012] Wherein: X k is the input multi-source data matrix, is the local optimal projection, is the multi-source Laplacian matrix, is the diagonal matrix of, argmin() is a mathematical function used to find the parameter value that minimizes the function, and H is the projection matrix of the initial high-dimensional data onto the low-dimensional data;

[0013] Define the multi-source global proximity weight matrix as:

[0014]

[0015] Wherein: is the multi-source global proximity weight matrix, i represents a certain row in the matrix, and j represents a certain column in the matrix;

[0016] Maximize the multi-source global proximity objective function:

[0017]

[0018] Wherein: is the global optimal projection, is the multi-source Laplacian matrix, is the diagonal matrix of;

[0019] Establish an improved fractional-order flexible space learning model according to the minimization of the multi-source local proximity and the maximization of the multi-source global proximity function. After calculation by the Lagrange multiplier method, the optimal projection matrix is obtained and the optimal projection matrix is transformed into a low-dimensional feature subset according to the following formula:

[0020]

[0021] Wherein: Y k is the processed low-dimensional multi-source data matrix, is the optimal projection matrix;

[0022] Step 3: Construct a data-driven model based on the improved TCN-BiGRU method;

[0023] Step 4: Combine the improved semi-invariant method and the data-driven model in Step 3 to obtain a hybrid-driven model;

[0024] Step 5: Construct a digital twin substation area dynamic model and simulate and deduce the service expansion installation plan.

[0025] Furthermore, the static data includes the topology structure data of the power distribution area, the parameters of power equipment, and the service life of the equipment; the dynamic data includes the power operation data, the equipment status monitoring data, the new energy power generation data, and the load demand; the environmental data includes the meteorological data, the atmospheric pressure, and the precipitation data.

[0026] Furthermore, the key information of the processed multi-source data is captured by a Temporal Convolutional Neural Network (TCN), and multi-dimensional convolution is adopted:

[0027]

[0028] where: d is the dilation coefficient, M is the size of the convolutional kernel, and g(i) is the i-th element in the convolutional kernel, is the input data, and s is the Laplace domain;

[0029] And the residual connection is performed by the following formula:

[0030] h(y) = y + g(y)

[0031] where: y is the data set, h(y) is the output of the residual connection, and g(y) is the non-linear transformation;

[0032] Two groups of unidirectional GRU neural network units are added to the TCN-BIGRU, and they are connected in opposite directions to form an improved TCN-BIGRU neural network model with a bidirectional network algorithm;

[0033] The influence of the input features on the application for new power connections is evaluated by the multi-head attention mechanism, and corresponding weights are assigned to each influencing factor. Each matrix is divided into multiple heads to obtain:

[0034]

[0035] A(Y) = Concat(A1, A2,..., A h )W o

[0036] where: A i is the attention output of the i-th head, d k is the dimension of the key vector, Q i is the i-th value query space, K i is the i-th key space, V i is the i-th value space, W o is the weight vector of the multi-source data, and A(Y) is the sum of the output attentions of h heads;

[0037] A data-driven model for the application for new power connections is constructed by a Temporal Convolutional Neural Network (TCN), a BiGRU model, and a multi-head attention mechanism to predict the power output fluctuations of the power distribution area load, photovoltaic power, and wind power.

[0038] Further, step 4 includes: constructing a power system calculation model based on the power topology model;

[0039] The power topology model is the IEEE 34-node system, where: Photovoltaic is connected to nodes 13 and 18, and wind energy is connected to nodes 19 and 22;

[0040] By combining the load normal probability model, the scenario joint probability distribution models of wind-wind and wind-photovoltaic, calculate the geometric distance between each pair of scenarios a and b in scenario N, and replace scenario a with the scenario g in scenario M that is closest to scenario a in geometric distance to form a new scenario set M`. Repeat this step to reduce the number of scenarios;

[0041] Express the node voltage U and branch power P as a linear sum of node injection power variables as follows:

[0042]

[0043] where: J0 and G0 are the partial derivatives of the node injection power imbalance and branch power with respect to the node voltage amplitude and phase angle, respectively, and ΔU (s) , ΔP (s) , ΔX (s) are the node voltage change, branch power change, and node injection power change after semi-invariant conversion;

[0044] Combined with the Gram-Charlier series, obtain the probabilistic power flow results, combine with the power topology model to obtain the physical driving model, use the physical driving model to provide grid operation constraints such as voltage stability, use the data-driven model in step 3 to compensate for the prediction ability of load and new energy fluctuations, and compare the real-time data with the predicted values of the model to adjust the model parameters to improve the accuracy, and obtain the hybrid driving model.

[0045] Further, step 5 includes: injecting the low-dimensional static data and historical data obtained in step 2 into the hybrid driving model to generate an initial twin model, and real-time accessing sensor data to dynamically update the substation area load, equipment status, and new energy access situation, and automatically adjust the key parameters of the twin model to ensure that the model is synchronized with the actual operating state, including the following steps:

[0046] Step 5.1: Input the new load access requirements of the user and the access locations and scales of photovoltaic and energy storage. Among them, the new load includes the load access point, load power, and load type;

[0047] Step 5.2: Simulate the impact of new energy access on the operation of the substation area in the twin model, and evaluate the impact of new energy access on grid loss and system stability by analyzing voltage fluctuations, current distribution, and line load rate changes;

[0048] Step 5.3: Use the maximum mutual trust coefficient method to analyze the relationship between the newly added load connection point and the voltage stability of the substation area, and the impact of new energy fluctuations on the line load rate and power distribution;

[0049] Step 5.4: Establish a feedback mechanism. By optimizing the business expansion and new energy access plans, adjust the connection points, power distribution, and load ratio, and output the optimal access plan.

[0050] The business expansion application digital twin model device based on hybrid drive includes: a data preprocessing module, a hybrid drive module, and a business expansion application plan analysis module; among them:

[0051] The data preprocessing module communicates with the power grid system to obtain multivariate data. By using the improved fractional-order flexible space method, the high-dimensional data is reduced in dimension, and finally, it is transmitted to the hybrid drive module;

[0052] The hybrid drive module uses the improved TCN-BIGRU method to perform deep learning on the low-dimensional data in Step 1 to form a data-driven module. Combining the IEEE 34-node system and the improved semi-invariant power flow calculation method to obtain a physical drive model, the data-driven model corrects the physical drive model, and transmits the result to the business expansion application analysis module;

[0053] The business expansion application plan analysis module communicates with the power grid, takes the user application data as input, analyzes the changes in node voltage, branch power, and node injection power of the user input, finds the plan with the least impact on the power grid stability, and obtains the optimal application plan.

[0054] The present invention has the following beneficial effects:

[0055] Based on the digital twin technology, the present invention can obtain the real-time operation status of the substation area power grid and monitor the real-time data of transformers, feeders, loads, and new energy power generation. Secondly, it can be dynamically updated under the background of new energy access, reflect the changes in the power grid in real time, help power grid managers take countermeasures in time, and avoid equipment overload or power waste. Finally, it dynamically optimizes the characteristics of the cooperation of wind power, photovoltaics, and energy storage, improves the ability of the power grid to accommodate new energy, and avoids affecting the stability of the power grid operation due to excessive volatility of new energy power generation. Brief Description of the Drawings

[0056] Figure 1 It is the method flow chart of the present invention;

[0057] Figure 2 It is the schematic diagram of the TCN-BiGRU model structure;

[0058] Figure 3 It is the schematic diagram of the IEEE 34-node system. Detailed Embodiments

[0059] Next, in combination with the Figures 1 - 3 in the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. If not specifically specified, the technical means used in the embodiments are conventional means well known to those skilled in the art.

[0060] The present invention utilizes digital twin technology to construct a power system model with real-time monitoring and optimization capabilities. This model can accurately simulate the grid state of the active substation area, analyze the characteristics of core parameters under different business expansion plans, predict the impact of new energy access on grid operation, and provide more accurate and flexible scheduling and decision-making basis for grid management. First, obtain various types of data such as substation area, equipment, and environment, and preprocess the multivariate data according to the improved fractional-order flexible space method. Secondly, construct a data-driven model based on the improved TCN-BiGRU method. Then, combine the improved semi-invariant method and the data-driven model to obtain a hybrid-driven model. Finally, inject static data and historical data into the hybrid-driven model to generate an initial twin model, and simulate and deduce the business expansion application plan.

[0061] A method for establishing a digital twin model for business expansion application based on hybrid drive includes the following steps:

[0062] Step 1: Collect multivariate data, where the multivariate data includes: static data, dynamic data, and environmental data;

[0063] As Figure 2 , the collection of multivariate data provides rich information support for constructing a digital twin substation area model, specifically including static data, dynamic data, and environmental data. Among them, the static data includes data such as the substation area topological structure data, power equipment parameters, and the usage cycle of equipment. The dynamic data is the data that changes in real time under the operation state of the substation area, collected by sensors, SCADA systems, etc., including power operation data, equipment status monitoring data, new energy generation data, and load demand data, etc. The environmental data is the external factor data under the substation area equipment and new energy access, specifically including meteorological data, atmospheric pressure, precipitation data, etc.

[0064] Step 2: Preprocess the multivariate data based on the improved fractional-order flexible space method;

[0065] Convert the multivariate data obtained in step 1 into an m*n matrix, where k is the type of data, to obtain the original high-dimensional multivariate data feature set m is the data dimension. By mapping the high-dimensional multivariate data to low-dimensional multivariate data and retaining the similarity between the data. By establishing a multivariate local neighborhood graph and a multivariate local objective function, and optimizing to obtain the typical features of the multivariate data.

[0066] Define the multi - variable local proximity weight matrix as:

[0067]

[0068] Where: is the multi - variable proximity weight matrix, denotes the neighborhood of length r of, and t is the time series. is the j - th group of data in the k - th variable data.

[0069] The minimized multi - variable local proximity objective function is:

[0070]

[0071]

[0072] Where: is the local optimal projection, is the multi - variable Laplacian matrix, is the diagonal matrix of. argmin() is a mathematical function used to find the parameter value that minimizes the function, and H is the projection matrix from the initial high - dimensional data to the low - dimensional data.

[0073] Define the multi - variable global proximity weight matrix as:

[0074]

[0075] Where: is the multi - variable global proximity weight matrix; i represents a certain row in the matrix, and j represents a certain column in the matrix.

[0076] Maximize the multi - variable global proximity objective function:

[0077]

[0078] Where: is the global optimal projection, is the multi - variable Laplacian matrix, is the diagonal matrix of.

[0079] Establish an improved fractional - order flexible space learning model according to the minimized multi - variable local proximity and maximized multi - variable global proximity functions. After calculating by the Lagrange multiplier method, the optimal projection matrix is obtained and the optimal projection matrix is transformed into a low - dimensional feature subset according to the following formula.

[0080]

[0081] Among them: Y k is the processed low-dimensional multi-source data matrix, is the optimal projection matrix, and X k is the input multi-source data matrix.

[0082] Step 3: Construct a data-driven model based on the improved TCN-BiGRU method;

[0083] Capture the key information of the processed multi-source data through a temporal convolutional neural network, and use multi-dimensional convolution:

[0084]

[0085] Among them: d is the dilation coefficient, M is the size of the convolution kernel, and g(i) is the i-th element in the convolution kernel, is the input data, and s is the Laplace domain.

[0086] And the residual is connected by the following formula:

[0087] h(y) = y + g(y)

[0088] Among them: y is the data set, h(y) is the output of the residual connection, and g(y) is the non-linear transformation.

[0089] Add 2 groups of unidirectional GRU neural network units through TCN-BIGRU, and connect them in the opposite direction to form an improved TCN-BIGRU neural network model with a bidirectional network algorithm, analyzing the states at the previous and subsequent moments while expanding the attention scope;

[0090] Evaluate the influence of the input features on the application for new electricity connection and installation by the multi-head attention mechanism, and assign corresponding weights to each influencing factor. Divide each matrix into multiple heads to obtain:

[0091]

[0092] A(Y) = Concat(A1, A2,..., A h )W o

[0093] Among them: A i is the attention output of the i-th head, d k is the dimension of the key vector, Q i the i-th value query space, K i is the i-th key space, V i is the i-th value space, W o is the weight vector of the multi-source data, and A(Y) is the sum of the output attentions of h heads.

[0094] A data-driven model for business expansion application is constructed by a temporal convolutional neural network, a BiGRU model, and a multi-head attention mechanism to predict the power output fluctuations of the substation area load, photovoltaic, and wind power. Figure 2 The schematic diagram of the TCN-BiGRU model structure is given.

[0095] Step 4: Combine the improved semi-invariant method and the data-driven model in Step 3 to obtain a hybrid-driven model;

[0096] Based on the power topology model, a power system calculation model is constructed. The power topology model is the IEEE 34-node system as shown in Figure (3), where: Photovoltaic is connected at nodes 13 and 18, and wind energy is connected at nodes 19 and 22; On the basis of this power topology model, the power flow calculation method of the improved semi-invariant method is used to calculate the node voltage, current distribution, and power flow direction.

[0097] By combining the load normal probability model, the wind-wind and wind-photovoltaic scenario joint probability distribution models, calculate the geometric distance between each pair of scenarios a and b in scenario N, and replace scenario a with scenario g that has the closest geometric distance to scenario a in scenario M to form a new scenario set M`. Repeat this step to reduce the number of scenarios.

[0098] Express the node voltage U and the branch power P as the linear sum of the node injection power variables as follows:

[0099]

[0100] Where: J0 and G0 are the node injection power imbalance and the partial derivatives of the branch power with respect to the node voltage amplitude and phase angle respectively, and ΔU (s) , ΔP (s) , ΔX (s) Are the node voltage change, branch power change, and node injection power change after semi-invariant conversion;

[0101] Combine the Gram-Charlier series to obtain the probabilistic power flow result. Combine the power topology model to obtain a physical-driven model. The physical-driven model provides grid operation constraints such as voltage stability. The data-driven model in Step 3 compensates for the prediction ability of load and new energy fluctuations, and compares the real-time data with the predicted values of the model, and adjusts the model parameters to improve the accuracy to obtain a hybrid-driven model.

[0102] Step 5: Construct a digital twin substation area dynamic model and simulate and deduce the business expansion application plan;

[0103] Inject the low-dimensional static data and historical data obtained in step 2 into the hybrid-driven model to generate an initial twin model, and connect to sensor data in real time to dynamically update the load of the power distribution area, the status of equipment, and the access situation of new energy, and automatically adjust the key parameters of the twin model (such as line impedance, load distribution) to ensure that the model is synchronized with the actual operating status.

[0104] Step 5.1: Input the access requirements of newly added loads by users and the access locations and scales of photovoltaic and energy storage. Among them, the newly added loads include load access points, load powers, and load types (residential, industrial, commercial).

[0105] Step 5.2: Simulate the impact of new energy access on the operation of the power distribution area in the twin model, and evaluate the impact of new energy access on power grid loss and system stability by analyzing voltage fluctuations, current distribution, and changes in line load rates.

[0106] Step 5.3: Use the maximum mutual trust coefficient method to analyze the relationship between the newly added load access point and the voltage stability of the power distribution area, and the impact of new energy fluctuations on line load rates and power distribution.

[0107] Step 5.4: Establish a feedback mechanism, adjust the access point, power distribution, and load ratio by optimizing the business expansion and new energy access plans, and output the optimal access plan.

[0108] The business expansion digital twin model device based on hybrid drive includes: a data preprocessing module, a hybrid drive module, and a business expansion plan analysis module; among them:

[0109] The data preprocessing module communicates with the power grid system to obtain multivariate data, reduces the dimension of the high-dimensional data by the improved fractional-order flexible space method, and finally transmits it to the hybrid drive module;

[0110] The hybrid drive module uses the improved TCN-BIGRU method to perform deep learning on the low-dimensional data in step 1 to form a data-driven module, combines the IEEE 34-node system and the improved semi-invariant power flow calculation method to obtain a physical drive model, corrects the physical drive model with the data-driven model, and transmits the result to the business expansion analysis module;

[0111] The business expansion plan analysis module communicates with the power grid, takes the user's application data as input, analyzes the changes in node voltages, branch powers, and node injection powers input by the user, finds the plan with the least impact on power grid stability, and obtains the optimal application plan.

[0112] The embodiments described above are only descriptions of the preferred embodiments of the present invention, and do not limit the scope of the present invention. Without departing from the spirit of the design of the present invention, various deformations, variations, modifications, and substitutions made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for establishing a digital twin model for business expansion and installation based on hybrid drive, characterized in that: The following steps are involved: Step 1: Collect multivariate data, which includes static data, dynamic data and environmental data; Step 2: Preprocess the multivariate data based on the improved fractional-order flexible space method, including: converting the multivariate data into an m*n matrix, where k is the data type, and obtaining the original high-dimensional multivariate data feature set m is the data dimension. High-dimensional multivariate data is mapped to low-dimensional multivariate data. By establishing a multivariate local neighborhood graph and a multivariate local objective function, the typical characteristics of multivariate data are obtained through optimization: Define the multivariate local proximity weight matrix as: in: is the multivariate proximity weight matrix, express The length of a neighborhood is r, and t is a time series; is the jth group of data in the kth metadata; The objective function of minimizing multivariate local proximity is obtained as: Where: X k is the input multivariate data matrix, is the local optimal projection, is the multivariate Laplace matrix, for The diagonal matrix of , argmin() is a mathematical function used to find the parameter value that minimizes the function, and H is the projection matrix of the initial high-dimensional data to the low-dimensional data; Define the multivariate global proximity weight matrix as: in: is the multivariate global proximity weight matrix, i represents a row in the matrix, and j represents a column in the matrix; Maximize the multivariate global proximity objective function: in: is the global optimal projection, is the multivariate Laplace matrix, for The diagonal matrix of ; An improved fractional-order flexible space learning model is established by minimizing the multivariate local proximity and maximizing the multivariate global proximity function, and the optimal projection matrix is ​​calculated by the Lagrange multiplier method. And the optimal projection matrix is ​​converted into a low-dimensional feature subset according to the following formula: Where: Y k is the processed low-dimensional multivariate data matrix, is the optimal projection matrix; Step 3: Build a data-driven model based on the improved TCN-BiGRU method; Step 4: Combine the improved semi-quantitative method and the data-driven model in step 3 to obtain a hybrid driven model; Step 5: Build a dynamic model of the digital twin substation, and simulate and deduce the business expansion plan.

2. The method for establishing a digital twin model for business expansion and installation based on hybrid drive according to claim 1 is characterized in that: In step 1, static data includes substation topology data, power equipment parameters and equipment life cycle; dynamic data includes power operation data, equipment status monitoring data, new energy power generation data and load demand; environmental data includes meteorological data, atmospheric pressure and precipitation data.

3. The method for establishing a digital twin model for business expansion and installation based on hybrid drive according to claim 1 is characterized in that: Step 3 includes: capturing key information of the processed multivariate data through a temporal convolutional neural network, using multidimensional convolution: Where: d is the expansion coefficient, M is the size of the convolution kernel, g(i) is the i-th element in the convolution kernel, is the input data, s is the pull domain; And the residuals are connected by the following formula: h(y)=y+g(y) Where: y is the data set, h(y) is the residual connection output, and g(y) is the nonlinear transformation; Two groups of unidirectional GRU neural network units are added through TCN-BIGRU, and they are connected in opposite directions to form an improved TCN-BIGRU neural network model with a bidirectional network algorithm; The multi-head attention mechanism evaluates the influence of input features on business expansion registration, assigns corresponding weights to each influencing factor, and divides each matrix into multiple heads to obtain: A(Y)=Concat(A1,A2,...,A h )W o Among them: A i is the attention output of the i-th head, d k is the key vector dimension, Q i The i-th value query space, K i is the i-th key space, V i is the i-th value space, W o is the weight vector of multivariate data, and A(Y) is the sum of the output attentions of h heads; A data-driven model for business expansion reporting is constructed through a time-series convolutional neural network, a BiGRU model, and a multi-head attention mechanism to predict the power output fluctuations of the station load, photovoltaic power, and wind power.

4. The method for establishing a digital twin model for business expansion and installation based on hybrid drive according to claim 1 is characterized in that: Step 4 includes: constructing a power system calculation model based on the power topology model; The power topology model is an IEEE 34-node system, where photovoltaic power is connected at nodes 13 and 18, and wind power is connected at nodes 19 and 22; By combining the load normal probability model, wind-wind and wind-solar scene joint probability distribution models, the geometric distance between each pair of scenes a and b in scene N is calculated, and the scene g with the closest geometric distance to scene a in scene M replaces scene a to form a new scene set M`. This step is repeated to reduce the number of scenes. The linear sum of node voltage U and branch power P representing node injection power variables is expressed as follows: Where: J0, G0 are the partial derivatives of node injection power imbalance and branch power to node voltage amplitude and phase angle, ΔU (s) , ΔP (s) , ΔX (s) is the node voltage change, branch power change and node injection power change after semi-quantity conversion; Combined with the Gram-Charlier series, the probabilistic power flow results are obtained, and combined with the power topology model, the physical driving model is obtained. The physical driving model provides grid operation constraints such as voltage stability. The data-driven model in step 3 compensates for the prediction ability of load and renewable energy fluctuations, compares the real-time data with the predicted values ​​of the model, adjusts the model parameters to improve the accuracy, and obtains a hybrid driving model.

5. The method for establishing a digital twin model for business expansion and installation based on hybrid drive according to claim 1 is characterized in that: Step 5 includes: injecting the low-dimensional static data and historical data obtained in step 2 into the hybrid drive model to generate an initial twin model, and accessing sensor data in real time, dynamically updating the load of the substation, equipment status and new energy access status, and automatically adjusting the key parameters of the twin model to ensure that the model is synchronized with the actual operating status, including the following steps: Step 5.1: Input the user's newly added load access requirements and the access locations and access scales of photovoltaic and energy storage. The newly added loads include load access points, load power and load types. Step 5.2: Simulate the impact of renewable energy access on substation operation in the twin model, and evaluate the impact of renewable energy access on grid loss and system stability by analyzing voltage fluctuations, current distribution, and line load rate changes; Step 5.3: Use the maximum mutual trust coefficient method to analyze the relationship between the newly added load access points and the voltage stability of the substation area, and the impact of new energy fluctuations on line load rate and power distribution; Step 5.4: Establish a feedback mechanism to optimize business expansion and new energy access solutions, adjust access points, power distribution and load ratio, and output the optimal access solution.

6. A digital twin model device for business expansion based on hybrid drive, characterized in that: The method for establishing a digital twin model of business expansion and installation based on hybrid drive according to any one of claims 1 to 5 comprises: a data preprocessing module, a hybrid drive module, and a business expansion and installation solution analysis module; wherein: The data preprocessing module communicates with the power grid system to obtain multivariate data, and reduces the dimension of high-dimensional data by improving the fractional-order flexible space method. Finally, it is transmitted to the hybrid drive module; The hybrid drive module will improve the TCN-BIGRU method to conduct deep learning on the low-dimensional data in step 1 to form a data-driven module, combine the IEEE 34-node system and the improved semi-invariant power flow calculation method to obtain the physical drive model, use the data-driven model to correct the physical drive model, and pass the result to the business expansion reporting analysis module; The business expansion application plan analysis module is interconnected with the power grid, takes the user application data as input, analyzes the user input node voltage change, branch power change and node injection power change, finds the plan with the least impact on the stability of the power grid, and obtains the optimal application plan.

Images