New energy consumption software model construction method based on simulation platform

Abstract

The invention relates to the technical field of new energy power systems, and discloses a new energy consumption software model construction method based on a simulation platform, and the method comprises the steps: firstly dividing new energy consumption system units, determining the topological structure and operation parameters of the new energy consumption system units, and constructing a multi-physics field coupling simulation model; then, carrying out distributed parallel solution on the model, defining a global optimization target, constructing a joint optimization and network reconstruction model, respectively adopting a deep reinforcement learning algorithm and graph neural network optimization, and generating an optimization scheme through multi-agent collaboration; whether the optimization target converges is verified, if yes, control parameters are output, and if not, optimization variables are updated to continue optimization; and finally, the optimized parameters are imported into the digital twin platform calibration model integrated with the block chain technology, and hyper-parameters are adjusted. The method improves the accuracy of a new energy consumption system model, optimizes the system operation, enhances the cooperative stability, and promotes the efficient utilization of new energy and the sustainable development of a power system.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy power systems, and specifically to a method for constructing a software model for new energy consumption based on a simulation platform. Background Art

[0002] With the increasing global demand for clean energy, the proportion of new energy in the power system is constantly rising. New energy sources such as solar energy and wind energy have significant advantages of being clean and renewable, which is of great significance for alleviating the energy crisis and reducing environmental pollution. However, the large-scale integration of new energy has brought many severe challenges to the stable operation and efficient consumption of the power system.

[0003] From the perspective of power generation characteristics, new energy has strong intermittency and volatility. Taking photovoltaic power generation as an example, its output directly depends on the light intensity. It generates a large amount of electricity during the day when the light is sufficient, while completely stops generating electricity at night; and weather changes such as cloud cover will cause the photovoltaic power generation to fluctuate greatly instantaneously. Wind power generation is also affected by wind speed, and the unstable wind speed leads to the fluctuation of the power generation power of the wind turbines. This unstable power generation characteristic seriously conflicts with the relatively stable power supply and demand balance mode of the traditional power system, bringing great difficulties to the power balance and frequency stability control of the power system. If not effectively addressed, it will frequently cause grid frequency fluctuations, affect the power supply quality of the power system, and even threaten the safe and stable operation of the power grid.

[0004] Energy storage technology is crucial for new energy consumption, but there are many problems with the current energy storage systems. On the one hand, the cost of energy storage devices is extremely high. Whether it is lithium-ion batteries, flow batteries or supercapacitors, the initial investment cost is relatively high, which greatly restricts the large-scale application of energy storage systems and makes them unable to fully play the role of regulating the output fluctuations of new energy. On the other hand, the existing energy storage charge and discharge management strategies are not optimized enough and do not fully consider the life attenuation characteristics of energy storage devices. Unreasonable charge and discharge strategies will accelerate the aging of energy storage devices, shorten their service life, further increase the comprehensive cost of energy storage systems, and reduce their economic feasibility in the power system.

[0005] In terms of load management, the traditional power load management method is relatively extensive, and the enthusiasm of users to participate in demand response is not high. The response mechanism of power users to electricity price changes is not perfect, and they cannot flexibly adjust their electricity consumption behavior according to real-time electricity price signals. When new energy generation is excessive, it is difficult to effectively guide users to increase their electricity consumption load; while when new energy generation is insufficient, it is also impossible to promptly prompt users to reduce non-essential electricity consumption, resulting in the inefficient allocation of power resources and low new energy consumption efficiency.

[0006] The power grid transmission link also faces challenges. With the widespread distribution of new energy power generation sites, the topology of the power grid has become increasingly complex, making it difficult for traditional power grid planning and operation methods to adapt. The line impedance changes dynamically due to environmental factors such as temperature, resulting in increased network losses and reduced transmission efficiency. At the same time, existing power grid topology optimization methods often fail to respond to the complex and changing operating conditions of the power system in real time and accurately, and it is difficult to maximize the transmission efficiency while ensuring the safe and stable operation of the power grid.

[0007] There are many deficiencies in existing new energy consumption-related technologies and models. Some models do not fully consider various influencing factors of new energy power generation during construction, resulting in low prediction accuracy and unable to provide a reliable basis for the dispatching decision-making of the power system. Some optimization algorithms have low computational efficiency and are difficult to meet the requirements of real-time optimization of large-scale power systems. Moreover, there is no effective coupling and collaborative optimization mechanism between different unit models, and it is impossible to achieve the optimal operation of the new energy consumption system as a whole. Summary of the Invention

[0008] The purpose of the present invention is to provide a method for constructing a new energy consumption software model based on a simulation platform to solve the problems proposed in the above background technology.

[0009] To achieve the above purpose, the present invention provides the following technical solution: A method for constructing a new energy consumption software model based on a simulation platform, the method includes: Step S1: Divide the new energy consumption system into a power generation unit, a energy storage unit, a load unit and a transmission unit, and determine the topology and operating parameters of each unit; Step S2: Based on real-time meteorological data and equipment operation data, construct a power generation unit output prediction model, an energy storage unit charge and discharge dynamic model, a load unit demand response model and a transmission unit network loss model respectively, and connect each model according to the actual energy flow relationship through a data interface to form a multi-physical field coupling simulation model of the new energy consumption system; Step S3: Perform distributed parallel solution on the multi-physical field coupling simulation model to obtain an initial simulation result; Step S4: Define the global optimization objective of the new energy consumption system, including at least one of minimizing the curtailment rate of wind and light and maximizing the power grid stability index; Step S5: Construct a joint optimization model of the power generation unit and the energy storage unit, use the power generation power dispatching and the energy storage charge and discharge rate as optimization variables, combine the initial simulation result of Step S3 as the feasible solution boundary, and use the deep reinforcement learning algorithm for policy iteration to generate the first optimization plan; Step S6: Determine whether the load unit introduces a dynamic electricity price response mechanism. If it does, correct the load demand curve based on the first optimization plan in step S5 and return to step S5; if not, execute step S7; Step S7: Construct a transmission unit network reconstruction model. With the node voltage deviation and line load rate as constraint conditions, use a graph neural network to dynamically optimize the power grid topology and generate a second optimization plan; Step S8: Integrate the optimization plans of step S5 and step S7, and coordinate the control instructions of each unit through a multi-agent cooperative optimization algorithm; Step S9: Verify whether the global optimization goal meets the convergence condition; if it does, output the optimization plan as the control parameters of the simulation platform; if not, return to step S5 to update the constraint range of the optimization variables.

[0010] Preferably, in step S1, the power generation unit includes one or more combinations of a photovoltaic array, a wind turbine, and a distributed gas turbine; the energy storage unit includes one or more combinations of a lithium-ion battery, a flow battery, and a supercapacitor; the transmission unit includes one or more combinations of an AC transmission line, a DC transmission line, and a flexible power electronic device; The topological structure includes the electrical connection relationship and geographical location distribution of each unit; the operating parameters include the rated capacity of the equipment, the efficiency curve, the aging coefficient, and environmental impact factors.

[0011] Preferably, the power generation unit output prediction model uses a temporal convolutional network to fuse meteorological satellite data and historical power sequences to predict the power generation fluctuation within the next 24 hours; the energy storage unit charge and discharge dynamic model is based on an electrochemical mechanism equation and an equivalent circuit model to simulate the life decay characteristics under different charge and discharge strategies; the load unit demand response model divides user types through cluster analysis and generates a load adjustment strategy in combination with the electricity price elasticity matrix; the transmission unit network loss model uses the adaptive finite element method to calculate the dynamic characteristics of the line impedance changing with temperature.

[0012] Preferably, in step S3, the distributed parallel solution uses a task scheduling mechanism based on the Kubernetes containerization architecture, splits the simulation model into independent computing nodes, realizes cross-node data synchronization through a message middleware, and uses a GPU to accelerate the solution of partial differential equations.

[0013] Preferably, in step S5, the deep reinforcement learning algorithm uses a Twin Delayed Deep Deterministic Policy Gradient (TD3) framework, designs the reward function as a weighted combination of curtailment of wind and light penalties, energy storage life loss, and power grid frequency regulation benefits, and dynamically updates the policy network parameters through an experience replay buffer.

[0014] Preferably, in step S7, the graph neural network constructs a heterogeneous graph structure with power grid nodes as vertices and lines as edges, dynamically allocates line weights using the graph attention mechanism, and optimizes the topological connection scheme through the gradient descent method, so that the improvement rate of the power transmission efficiency after network reconstruction is not lower than a preset threshold.

[0015] Preferably, in step S8, the multi-agent collaborative optimization algorithm designs a game strategy using the Nash equilibrium theory. Each unit agent exchanges local optimization information through a distributed consensus protocol and uses mixed-integer linear programming to solve the Pareto optimal solution.

[0016] Preferably, in step S9, the convergence conditions include that the residual of the objective function is less than the set tolerance, the number of iterations reaches the upper limit, or the fluctuation range of the optimization variables is stable within the neighborhood of the historical optimal solution.

[0017] Preferably, it further includes step S10: Import the optimized control parameters into the digital twin platform, calibrate the simulation model in real time through the virtual-real linkage mechanism, and dynamically adjust the model hyperparameters using the Bayesian optimization algorithm.

[0018] Preferably, in step S10, the digital twin platform integrates blockchain technology, stores the hash values of the simulation results and the measured data on the chain, verifies the data consistency through a smart contract, and triggers the model reconstruction process under abnormal working conditions.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention divides the new energy consumption system into power generation, energy storage, load, and transmission units in detail, accurately determines the topological structure and operating parameters of each unit, and the constructed multi-physical field coupling simulation model can highly restore the real operating state of the system. For example, the power output prediction model of the power generation unit integrates meteorological satellite data and historical power sequences, and uses a temporal convolutional network to predict the power generation fluctuation. Compared with traditional methods, the prediction accuracy is significantly improved, which can provide a more reliable basis for power dispatching in advance. The charge-discharge dynamic model of the energy storage unit is based on the electrochemical mechanism and the equivalent circuit model, and accurately simulates the life decay characteristics under different charge-discharge strategies, which helps to optimize the energy storage management strategy, extend the service life of the energy storage device, and reduce costs.

[0020] By defining the global optimization objective of the new energy consumption system, constructing a joint optimization model and a network reconstruction model, and combining a deep reinforcement learning algorithm and a graph neural network for optimization. In terms of the joint optimization of power generation and energy storage, a reward function is designed by comprehensively considering the penalties for wind and light curtailment, the loss of energy storage life, and the benefits of power grid frequency regulation. The Twin Delayed Deep Deterministic Policy Gradient (TD3) framework is used for policy iteration, effectively reducing the wind and light curtailment rate, improving the power grid frequency regulation benefits, and realizing the reasonable allocation of power generation and energy storage resources. The transmission unit network reconstruction model takes the node voltage deviation and line load rate as constraints, and uses the graph neural network for dynamic optimization, improving the transmission efficiency, reducing the network loss, and making the power grid operation more economical and efficient.

[0021] The multi-agent collaborative optimization algorithm is based on the Nash equilibrium theory, enabling each unit agent to exchange information through a distributed consensus protocol and using mixed-integer linear programming to solve the Pareto optimal solution, achieving collaborative control among the units. This process not only improves the overall performance of the system but also enhances the stability of the system under complex working conditions. For example, when there are significant fluctuations in new energy generation, each unit can quickly respond collaboratively. The energy storage unit rapidly adjusts the power, the load unit adjusts the electricity demand according to the dynamic electricity price response mechanism, and the transmission unit optimizes the network topology to ensure the stable power supply of the power system.

[0022] The application of the digital twin platform and blockchain technology provides strong support for the continuous optimization of the model and data management. The optimized control parameters are imported into the digital twin platform, and the simulation model is calibrated in real time through the virtual-real linkage mechanism. Combining with the Bayesian optimization algorithm, the model hyperparameters are dynamically adjusted, enabling the model to adapt to the dynamic changes of the power system. Blockchain technology uploads the hash values of the simulation results and measured data to the blockchain for storage, and uses smart contracts to verify the data consistency, ensuring the authenticity and reliability of the data. Under abnormal working conditions, it can trigger the model reconstruction process in a timely manner, further improving the reliability of the model and the safety of the power system operation.

[0023] By optimizing the operation of the new energy consumption system, the present invention effectively reduces the phenomena of wind and light curtailment, improves the utilization rate of new energy, reduces the dependence on traditional energy, and reduces energy waste and environmental pollution. At the same time, the reasonable energy storage management strategy prolongs the life of energy storage devices and reduces the cost of the energy storage system; the optimized power grid topology reduces the line loss and the power transmission cost. In the long run, these all contribute to promoting the sustainable development of the clean energy industry, promoting the transformation and upgrading of the energy structure, and making positive contributions to achieving the global low-carbon energy goal. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is the working principle diagram of the method for constructing a new energy consumption software model based on a simulation platform according to the present invention; Figure 2It is a diagram of the implementation steps for distributed parallel solution; Figure 3 It is a flowchart of the combined optimization algorithm for power generation and energy storage. Specific implementation manners

[0025] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0026] Please refer to Figures 1-3 , the present invention provides a method for constructing a new energy consumption software model based on a simulation platform, aiming to efficiently solve complex problems in the process of new energy consumption and improve the utilization efficiency and stability of new energy in the power system. The overall implementation process is as follows: Step S1: The new energy consumption system is carefully divided into a power generation unit, an energy storage unit, a load unit, and a transmission unit. For each unit, its topological structure and operating parameters are accurately determined. The topological structure covers the electrical connection relationships and geographical location distributions of each unit, and the operating parameters include equipment rated capacity, efficiency curve, aging coefficient, and environmental impact factors, etc. These parameters provide key basic data for subsequent model construction.

[0027] Step S2: According to real-time meteorological data and equipment operation data, a power generation unit output prediction model, an energy storage unit charge and discharge dynamic model, a load unit demand response model, and a transmission unit network loss model are respectively constructed. Through a data interface, these models are connected according to the actual energy flow relationship to form a multi-physical field coupling simulation model of the new energy consumption system, so as to truly simulate the energy interaction and physical processes between each unit in the system.

[0028] Step S3: The distributed parallel solution technology is used to solve the multi-physical field coupling simulation model to obtain the initial simulation results. Using the task scheduling mechanism based on the Kubernetes containerization architecture, the simulation model is split into independent computing nodes, cross-node data synchronization is realized with the help of a message middleware, and the GPU is used to accelerate the solution of partial differential equations to improve the solution efficiency.

[0029] Step S4: Clearly define the global optimization objectives of the new energy consumption system, such as minimizing the wind and light curtailment rate and maximizing the power grid stability index, etc., to provide a clear direction and measurement standard for subsequent optimization work.

[0030] Step S5: Construct a joint optimization model for the power generation unit and the energy storage unit. Take the power generation power dispatch and the energy storage charge and discharge rate as optimization variables, and determine the feasible solution boundary in combination with the initial simulation results obtained in Step S3. Use the deep reinforcement learning algorithm for policy iteration to generate the first optimization plan and achieve the coordinated optimization of power generation and energy storage.

[0031] Step S6: Determine whether the load unit introduces a dynamic electricity price response mechanism. If it is introduced, correct the load demand curve based on the first optimization plan in Step S5 and return to Step S5 for re-optimization; if it is not introduced, execute Step S7.

[0032] Step S7: Construct a transmission unit network reconfiguration model. With the node voltage deviation and line load rate as constraint conditions, use the graph neural network to dynamically optimize the power grid topology, generate the second optimization plan, and optimize the power grid transmission performance.

[0033] Step S8: Integrate the optimization plans of Step S5 and Step S7, and coordinate the control instructions of each unit through the multi-agent collaborative optimization algorithm to achieve the overall optimization control of the system.

[0034] Step S9: Verify whether the global optimization goal meets the convergence conditions. The convergence conditions include that the residual of the objective function is less than the set tolerance, the number of iterations reaches the upper limit, or the fluctuation range of the optimization variables stabilizes within the neighborhood of the historical optimal solution. If it is satisfied, output the optimization plan as the control parameters of the simulation platform; if it is not satisfied, return to Step S5 to update the constraint range of the optimization variables and continue the optimization.

[0035] Step S10: Import the optimized control parameters into the digital twin platform, calibrate the simulation model in real time through the virtual-real linkage mechanism, and use the Bayesian optimization algorithm to dynamically adjust the model hyperparameters. The digital twin platform integrates blockchain technology, uploads the hash values of the simulation results and the measured data to the blockchain for storage, verifies the data consistency through smart contracts, and triggers the model reconstruction process under abnormal conditions to ensure the accuracy and reliability of the model.

[0036] The following further describes the implementation of the present invention in combination with Embodiments 1 to 4.

[0037] Embodiment 1: The power generation unit output prediction model uses a temporal convolutional network (TCN) to fuse meteorological satellite data with historical power series to predict power generation fluctuations within the next 24 hours. Taking the photovoltaic array power generation prediction as an example, meteorological satellite data contains information such as solar radiation intensity and cloud cover, and the historical power series records the actual power generation of the photovoltaic array over a period of time in the past. After preprocessing, these data are input into the temporal convolutional network. The structure of the temporal convolutional network consists of multiple convolutional layers and pooling layers, and the size and step size of its convolution kernel are optimized according to the characteristics of the data. The convolutional layer is used to extract features from the data to capture the time series features in the data, such as the relationship between the daily variation of solar radiation intensity and the power generation of the photovoltaic array. The output of the prediction model is the predicted value of photovoltaic power generation per hour in the next 24 hours. , , the prediction formula can be expressed as: ,in is the temporal convolutional network model, for Weather satellite data at the moment, for The historical power series data corresponding to the time.

[0038] The dynamic model of energy storage unit charging and discharging is based on the electrochemical mechanism equation and the equivalent circuit model to simulate the life attenuation characteristics under different charging and discharging strategies. Taking lithium-ion batteries as an example, according to the electrochemical mechanism, the charging and discharging process of the battery involves chemical reactions, and its capacity change is related to the ion diffusion, charge transfer and other processes inside the battery. The equivalent circuit model equates the battery to a combination circuit of a resistor, capacitor and voltage source. By combining the electrochemical mechanism equation and the equivalent circuit model equation, the battery life at different charging and discharging currents is obtained. Voltage changes under and capacity changes Relationship: , ,in is the open circuit voltage of the battery, is the internal resistance of the battery, is a constant related to battery characteristics, is the number of reaction electrons, is the Faraday constant. By simulating the changes in these parameters under different charge and discharge strategies, the battery life decay can be evaluated, such as calculating the battery Remaining capacity after charge and discharge cycles .

[0039] The load unit demand response model divides user types through cluster analysis and generates a load adjustment strategy in combination with the electricity price elasticity matrix. First, collect a large amount of user electricity consumption data, including information such as electricity consumption time and electricity consumption volume. Use a cluster analysis algorithm, such as the K-means clustering algorithm, to divide users into different types, such as industrial users, commercial users, and residential users. For each user type, construct an electricity price elasticity matrix based on historical electricity consumption data and electricity price changes. The elements of the electricity price elasticity matrix represent the electricity price elasticity coefficient of the th type of user in the time period, and its calculation formula is: , where is the load change amount caused by the electricity price change of the th type of user in the time period, is the initial load of the th type of user in the time period, and is the initial electricity price of the time period. When the electricity price changes, calculate the load adjustment amount of each type of user according to the electricity price elasticity matrix, so as to generate a load adjustment strategy and achieve demand response.

[0040] The transmission unit network loss model calculates the dynamic characteristics of the line impedance changing with temperature using the adaptive finite element method. Divide the transmission line into multiple finite element units, and the impedance of each unit is related to the temperature . According to transmission line theory, the line impedance includes resistance and reactance , and changes with temperature, and its relationship is as described in Embodiment 1. Through the adaptive finite element method, automatically adjust the density of the finite element mesh according to the change of the line temperature distribution to improve the calculation accuracy. When calculating the network loss, first calculate the current of each line, calculate the loss of each line according to the power loss formula , and then sum the losses of all lines to obtain the loss of the entire transmission network, where represents the line number.

[0041] Through the above accurate model construction method, each unit model can more accurately reflect the actual operation situation, providing a powerful tool for the simulation and optimization of the new energy consumption system.

[0042] Embodiment 2: This embodiment details the application of distributed parallel solution technology in the solution of the multi-physical field coupling simulation model of the new energy consumption system.

[0043] When performing distributed parallel solution for the multi-physical field coupling simulation model, a task scheduling mechanism based on the Kubernetes containerization architecture is adopted. Kubernetes is an open-source container orchestration platform that can automatically deploy, scale, and manage containerized applications. First, the simulation model is split into multiple independent computing nodes, with each node corresponding to a container. For example, the power generation unit model, energy storage unit model, load unit model, and transmission unit model are encapsulated in different containers respectively. These containers run in the Kubernetes cluster, and Kubernetes dynamically allocates tasks to each container according to the resource conditions of each node (such as CPU, memory, GPU, etc.), achieving reasonable scheduling of tasks.

[0044] Cross-node data synchronization is achieved through a message middleware. The message middleware selected is Kafka, which is a high-throughput distributed message system. During the simulation process, data needs to be exchanged between each unit model. For example, the output data of the power generation unit needs to be transmitted to the load unit and energy storage unit models. As a message queue, Kafka receives the data sent from one unit model and accurately delivers it to the other unit model containers that need this data, ensuring reliable transmission of data between different nodes.

[0045] GPU is utilized to accelerate the solution of partial differential equations. In the new energy consumption system model, there are some partial differential equations describing physical processes, such as the Maxwell equations describing the power transmission process. The GPU (Graphics Processing Unit) has powerful parallel computing capabilities and can significantly improve the solution speed of partial differential equations compared to traditional CPUs. Taking the solution of partial differential equations by the finite element method as an example, the finite element calculation tasks are assigned to the GPU for execution. On the GPU, by parallel computing the equation solutions of multiple finite element units, the solution time is greatly shortened. Suppose the original CPU solution time is , after using GPU acceleration, the solution time is , and the acceleration ratio . Through actual tests, the acceleration ratio in this system can reach more than [X], effectively improving the solution efficiency of the simulation model.

[0046] Through the collaborative application of task scheduling based on the Kubernetes containerization architecture, data synchronization of the message middleware, and GPU acceleration technology, efficient distributed parallel solution of the multi-physical field coupling simulation model is achieved, providing a guarantee for quickly obtaining the initial simulation results.

[0047] Example 3: This example focuses on introducing the application of the deep reinforcement learning algorithm in the joint optimization model of the power generation unit and energy storage unit, as well as the application of the graph neural network in the transmission unit network reconstruction model. Through these optimization algorithms, the optimized operation of each part of the system is realized, improving the new energy consumption efficiency and grid performance.

[0048] When constructing the joint optimization model of power generation units and energy storage units, the double-delayed deep deterministic policy gradient (TD3) framework is used for optimization. Taking a regional new energy power generation and energy storage joint system as an example, the optimization variable is power generation scheduling and energy storage charge and discharge rate , , Design reward function to optimize time interval. It is a weighted combination of wind and solar power abandonment penalties, energy storage life loss, and grid frequency regulation benefits. The formula is: ,in , , is the weight coefficient, and . Penalty for abandoning wind and light The amount of wind and solar power abandoned Related, , is the penalty coefficient; energy storage life loss The life attenuation is determined based on the calculation of the energy storage unit charging and discharging dynamic model, such as , is the coefficient related to life loss, is the energy storage capacity attenuation; grid frequency regulation benefit Frequency deviation from the grid and FM power related, , is the profit coefficient.

[0049] Dynamically update the policy network parameters through the experience replay buffer. During the simulation, the state of each step ,action ,award and the next state Stored in the experience playback buffer. When the buffer stores a certain amount of data, a batch of data is randomly sampled from the buffer for training the policy network. The policy network updates parameters based on the sampled data, so that the policy network can learn a better strategy, thereby generating the first optimization plan, achieving coordinated optimization of the power generation unit and the energy storage unit, reducing the wind and solar power abandonment rate, extending the energy storage life, and improving the grid frequency regulation benefits.

[0050] When constructing the transmission unit network reconstruction model, the graph neural network is used to dynamically optimize the power grid topology. Taking a city distribution network as an example, the graph neural network constructs a heterogeneous graph structure with grid nodes as vertices and lines as edges. In the graph neural network, the graph attention mechanism is used to dynamically allocate line weights. and the nodes connected to it , by calculating the attention coefficient to determine the edge weight, and the calculation formula is: where is the weight matrix, , , are the feature vectors of the nodes, is the node set of neighbor nodes. Optimize the topological connection scheme by the gradient descent method, using the node voltage deviation and the line load rate as the constraint conditions, and the goal is to make the improvement rate of the transmission efficiency after network reconstruction not lower than the preset threshold . The calculation formula of the improvement rate of transmission efficiency is: where is the transmission power after network reconstruction, is the transmission power before network reconstruction. By continuously optimizing the topological connection, generate the second optimization scheme, improve the transmission efficiency of the power grid, reduce the line loss, and enhance the operation stability of the power grid.

[0051] Through the effective application of these two optimization algorithms, the collaborative optimization of power generation and energy storage in the new energy consumption system and the improvement of the power grid transmission performance are realized, providing key technical support for the overall optimized operation of the system.

[0052] Example 4: When integrating the optimization schemes of steps S5 and S7, a multi-agent collaborative optimization algorithm is used to coordinate the control instructions of each unit. Taking the power generation unit, energy storage unit, and transmission unit in the new energy consumption system as examples, each unit is regarded as an agent. The Nash equilibrium theory is used to design the game strategy, and each unit agent exchanges local optimization information through the distributed consensus protocol. In the distributed consensus protocol, each agent broadcasts its own optimization scheme (such as the power generation power scheduling plan of the power generation unit, the charge and discharge strategy of the energy storage unit, and the network reconstruction scheme of the transmission unit) to other agents. Each agent adjusts its own optimization scheme according to the information received from other agents, combined with its own goals and constraint conditions. The mixed integer linear programming is used to solve the Pareto optimal solution to achieve the collaborative optimization among units. For example, the power generation unit hopes to maximize the power generation revenue, the energy storage unit hopes to minimize its own loss, and the transmission unit hopes to reduce the network loss and ensure the power supply reliability. Through the mixed integer linear programming, an optimal set of control instructions that balances the interests of each unit is found, so that the overall performance of the system is improved.

[0053] In step S9, verify whether the global optimization objective meets the convergence conditions. The convergence conditions include that the residual of the objective function is less than the set tolerance, the number of iterations reaches the upper limit, or the fluctuation range of the optimization variables stabilizes within the neighborhood of the historical optimal solution. Assume that the objective function is the comprehensive cost function of the new energy consumption system , which includes multiple parts such as power generation cost, energy storage cost, and curtailment cost of wind and solar power, that is . In each iteration process, calculate the objective function value of the current iteration and the objective function value of the previous iteration residual . When is less than the set tolerance , it is considered that the residual of the objective function meets the convergence conditions. At the same time, set the maximum number of iterations . If the number of iterations reaches , stop the iteration regardless of whether the objective function converges. In addition, monitor the fluctuation range of the optimization variables (such as power generation power dispatch, energy storage charge and discharge rate, etc.). When the change amount of the optimization variables in multiple iterations is less than a certain threshold and stabilizes within the neighborhood of the historical optimal solution, it is also determined that the convergence conditions are met. If the convergence conditions are met, output the optimization scheme as the control parameters of the simulation platform; if not, return to step S5 to update the constraint range of the optimization variables and continue the optimization.

[0054] In step S10, import the optimized control parameters into the digital twin platform. The digital twin platform integrates blockchain technology and stores the hash values of the simulation results and measured data on the blockchain. Taking a certain power system as an example, in the digital twin platform, collect the measured data such as the actual power generation of the power generation unit, the charge and discharge status of the energy storage unit, the actual power consumption of the load unit, and the line loss of the transmission unit. At the same time, run the simulation model to obtain the corresponding simulation results. Use a hash function (such as the SHA-256 algorithm) to calculate the hash values of the measured data and simulation results respectively, and obtain the hash values and . Store these hash values on the blockchain. The decentralized and immutable characteristics of the blockchain ensure the security and credibility of the data. Verify the data consistency through a smart contract. The verification rules are preset in the smart contract. When and When they are equal, the data is considered consistent; if they are not equal, the model reconstruction process under abnormal conditions is triggered. In the model reconstruction process, analyze the reasons for the generation of abnormal data, such as equipment failures, parameter changes, etc., adjust and retrain the model of the new energy consumption system, and use the Bayesian optimization algorithm to dynamically adjust the model hyperparameters. The Bayesian optimization algorithm constructs a probability model of the objective function and predicts the value of the next optimal hyperparameter based on the existing data, improving the training efficiency of the model while ensuring the accuracy of the model, so that the simulation model can better reflect the operation of the actual system.

[0055] Through the above embodiments, the specific implementation methods of the new energy consumption software model construction method in the present invention in each key link are elaborated in detail. From the determination of unit parameters, model construction, solution technology, application of optimization algorithms to system optimization verification and digital twin application, each part cooperates closely to achieve the efficient simulation, optimization and operation management of the new energy consumption system, providing strong technical support for the wide application of new energy in the power system.

[0056] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0057] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for constructing a new energy consumption software model based on a simulation platform, characterized in that The method includes: Step S1: Divide the new - energy consumption system into a power - generation unit, an energy - storage unit, a load unit, and a transmission unit, and determine the topological structure and operating parameters of each unit; Step S2: Based on real - time meteorological data and equipment operating data, construct a power - generation unit output prediction model, an energy - storage unit charge - discharge dynamic model, a load unit demand - response model, and a transmission unit network - loss model respectively. Connect each model according to the actual energy - flow relationship through a data interface to form a multi - physical - field coupling simulation model of the new - energy consumption system; Step S3: Perform distributed parallel solution on the multi - physical - field coupling simulation model to obtain an initial simulation result; Step S4: Define the global optimization objective of the new - energy consumption system, including at least one of minimizing the curtailment rate of wind and light and maximizing the power - grid stability index; Step S5: Construct a joint optimization model of the power - generation unit and the energy - storage unit, with power - generation power scheduling and energy - storage charge - discharge rate as optimization variables. Combine the initial simulation result of Step S3 as the feasible - solution boundary, and use the deep reinforcement learning algorithm for policy iteration to generate a first optimization plan; Step S6: Determine whether the load unit introduces a dynamic electricity - price response mechanism. If it is introduced, correct the load - demand curve based on the first optimization plan in Step S5 and return to Step S5; if it is not introduced, execute Step S7; Step S7: Construct a transmission - unit network - reconstruction model, with node - voltage deviation and line - load rate as constraint conditions, and use a graph neural network to dynamically optimize the power - grid topology to generate a second optimization plan; Step S8: Integrate the optimization plans of Step S5 and Step S7, and coordinate the control instructions of each unit through a multi - agent collaborative optimization algorithm; Step S9: Verify whether the global optimization objective meets the convergence condition; if it meets, output the optimization plan as the control parameter of the simulation platform; if it does not meet, return to Step S5 to update the constraint range of the optimization variables.

2. The method according to claim 1, wherein In Step S1, the power - generation unit includes one or more combinations of a photovoltaic array, a wind turbine, and a distributed gas turbine; the energy - storage unit includes one or more combinations of a lithium - ion battery, a flow battery, and a supercapacitor; the transmission unit includes one or more combinations of an AC transmission line, a DC transmission line, and a flexible power - electronics device; The topological structure includes the electrical connection relationship and geographical location distribution of each unit; the operating parameters include equipment rated capacity, efficiency curve, aging coefficient, and environmental impact factors.

3. The method according to claim 1, characterized in that, In Step S2, the power - generation unit output prediction model uses a time - convolutional network to fuse meteorological satellite data and historical power sequences to predict the power - generation fluctuation within the next 24 hours; the energy - storage unit charge - discharge dynamic model is based on an electrochemical mechanism equation and an equivalent - circuit model to simulate the life - decay characteristics under different charge - discharge strategies; the load - unit demand - response model divides user types through clustering analysis and generates a load - adjustment strategy in combination with an electricity - price elasticity matrix; the transmission - unit network - loss model uses an adaptive finite - element method to calculate the dynamic characteristics of line impedance changing with temperature.

4. The method according to claim 1, wherein In step S3, the distributed parallel solution adopts a task scheduling mechanism based on the Kubernetes containerization architecture, splits the simulation model into independent computing nodes, realizes cross-node data synchronization through a message middleware, and uses GPU to accelerate the solution of partial differential equations.

5. The method according to claim 1, wherein In step S5, the deep reinforcement learning algorithm adopts the Twin Delayed Deep Deterministic Policy Gradient (TD3) framework, designs the reward function as a weighted combination of curtailment penalties, energy storage life losses, and grid frequency regulation benefits, and dynamically updates the policy network parameters through an experience replay buffer.

6. The method according to claim 1, wherein In step S7, the graph neural network constructs a heterogeneous graph structure with power grid nodes as vertices and lines as edges, dynamically allocates line weights using the graph attention mechanism, and optimizes the topological connection scheme through the gradient descent method, so that the improvement rate of the transmission efficiency after network reconstruction is not lower than the preset threshold.

7. The method according to claim 1, wherein In step S8, the multi-agent collaborative optimization algorithm designs the game strategy using the Nash equilibrium theory. Each unit agent exchanges local optimization information through a distributed consensus protocol and uses mixed-integer linear programming to solve the Pareto optimal solution.

8. The method according to claim 1, characterized in that In step S9, the convergence conditions include that the residual of the objective function is less than the set tolerance, the number of iterations reaches the upper limit, or the fluctuation range of the optimization variables stabilizes within the neighborhood of the historical optimal solution.

9. The method according to claim 1, wherein It also includes step S10: Import the optimized control parameters into the digital twin platform, calibrate the simulation model in real time through the virtual-real linkage mechanism, and dynamically adjust the model hyperparameters using the Bayesian optimization algorithm.

10. The method according to claim 9, characterized in that, In step S10, the digital twin platform integrates blockchain technology, stores the hash values of the simulation results and measured data on the chain, verifies the data consistency through a smart contract, and triggers the model reconstruction process under abnormal working conditions.

Images