Source network load storage system operation optimization method and system based on digital twinning

Abstract

The invention belongs to the technical field of intelligent power grid construction, and relates to a source network load storage system operation optimization method and system based on digital twinning, and the method comprises the steps: a test entity which collects the information of a physical entity through various sensors, and inputs the information of the physical entity to a virtual entity; the virtual entity comprises a prediction model used for predicting the new energy power generation power according to the information of the physical entity; the multi-objective optimization model is used for optimizing and determining charging and discharging power of energy storage and tie line power of the source network load storage system and the power distribution network in real time; calculating a health index of the physical entity by a health index calculation model, and dividing health state grades according to the health index; and the digital twin model is used for mapping the physical entity according to the output result of the virtual entity. The prediction value can be compensated according to the similar day meteorological search algorithm, so that the final power prediction value of new energy power generation is determined, and the condition that prediction errors are gradually increased after long-term operation is avoided.

Description

Technical Field

[0001] This application belongs to the technical field of smart grid construction, and more specifically, it is a method and system for optimizing the operation of a source-grid-load-storage system based on digital twin. Background Art

[0002] The integration of source-grid-load-storage is an important part of smart grid construction. It refers to the comprehensive use of power sources, power grids, loads, and energy storage technologies to achieve efficient energy utilization and intelligent scheduling, in order to meet the increasingly complex power market demands and environmental protection requirements, which is of great significance for improving energy utilization efficiency, promoting new energy consumption, and facilitating energy transformation.

[0003] As the scale of the source-grid-load-storage system increases, the panoramic real-time information such as loads, meteorology, and equipment is constantly increasing, which greatly increases the dimension of the information matrix and slows down the data processing time, bringing a time lag to the information transmission. Therefore, traditional modeling and simulation technologies are difficult to meet the operation requirements of the source-grid-load-storage system. Traditional modeling methods are limited by problem scale, variable dimension, deterministic boundaries, and dependence on accurate physical models, and it is difficult to ensure computational reliability with computational efficiency as the premise when dealing with high-dimensional, time-varying, and nonlinear problems. Summary of the Invention

[0004] An embodiment of this application provides a system for optimizing the operation of a source-grid-load-storage system based on digital twin.

[0005] Another aspect of the embodiment of this application provides a method for optimizing the operation of a source-grid-load-storage system based on digital twin, which solves the lag problem of traditional methods in high-dimensional time-varying scenarios.

[0006] This application is implemented as follows. A system for optimizing the operation of a source-grid-load-storage system based on digital twin provided by an embodiment of this application, the system includes: A test entity, which collects information of the physical entity through various sensors and inputs the information of the physical entity into the virtual entity. The information of the physical entity includes meteorological data, historical new energy power generation, and equipment parameters in the physical entity. A virtual entity, which includes a prediction model for predicting new energy power generation according to the information of the physical entity; a multi-objective optimization model for real-time optimizing and determining the charge and discharge power of the energy storage and the connection line power between the source-grid-load-storage system and the distribution network; and a health index calculation model for calculating the health index of the physical entity and dividing the health status level according to the health index. A digital twin model, which is used to map the physical entity according to the output result of the virtual entity, and includes a wind power generation model, a photovoltaic power generation model, and an energy storage model.

[0007] Further, the physical entity is a source-network-load-storage system, which adopts an AC-DC hybrid architecture and includes a photovoltaic array, a wind turbine generator, an external power grid, an AC load, a DC load, a battery energy storage, and an inverter. Among them, the photovoltaic array and the battery energy storage are connected to the DC bus of the external power grid, the wind turbine generator and the AC load are connected to the AC bus of the external power grid, and the DC bus and the AC bus are coupled through an inverter.

[0008] Further, the prediction model is used to predict the new energy power generation according to the information of the physical entity, including: Decompose the new energy power generation into several intrinsic mode function components and a residual component; Reconstruct the intrinsic mode function components and the residual component with similar sample entropy values into multiple power components; Take the power components and the meteorological variable data related to the new energy power generation as inputs, and use a long short-term memory neural network to make a preliminary prediction of the new energy power generation to obtain a preliminary prediction value; Search the historical meteorological database for the meteorological day data most similar to the meteorological variable data related to the new energy power generation, obtain the actual value of the power generation output of the most similar meteorological day data, and use a long short-term memory neural network to predict the power generation output of the most similar meteorological day data to obtain a predicted value of the power generation output of the most similar meteorological day data; When it is monitored that the deviation between the preliminary prediction value of the new energy power generation at a certain moment and the actual value of the power generation output of the most similar meteorological day data exceeds the threshold, the compensation mechanism is automatically triggered, and the prediction error between the actual value of the power generation output of the most similar meteorological day data and the predicted value of the power generation output of the most similar meteorological day data is used to compensate the preliminary prediction value to obtain the final power prediction value; Further, the multi-objective optimization model performs multi-objective optimization on the objective function with constraint conditions to obtain the optimal Pareto solution set, and the optimal Pareto solution set determines the charge and discharge power of the energy storage and the connection line power between the source-network-load-storage system and the distribution network.

[0009] Further, the objective function with constraint conditions is to establish an objective function with the minimum operation cost of the source-network-load-storage system and the minimum environmental protection cost under the constraint conditions of power balance constraint, upper and lower limits of equipment processing constraint, power output constraint of the grid transformer, and energy storage equipment constraint. The operation cost of the source-network-load-storage system includes the new energy power generation cost calculated from the final power prediction value.

[0010] The embodiment of the present application also provides a method for optimizing the operation of a source-network-load-storage system based on digital twin. The method includes: Collect the information of the physical entity; Predict the new energy power generation according to the information of physical entities; perform multi-objective optimization to determine the charging and discharging power of energy storage and the power of the connection line between the source-network-load-storage system and the distribution network; and calculate the health index of physical entities and divide the health status levels according to the health index. Map physical entities according to the output results of virtual entities.

[0011] Further, predicting the new energy power generation according to the information of physical entities includes: decomposing the new energy power generation into several intrinsic mode function components and residual components; Reconstructing the intrinsic mode function components and residual components with similar sample entropy values into multiple power components; Taking the power components and the meteorological variable data related to the new energy power generation as inputs, and using a long short-term memory neural network to make a preliminary prediction of the new energy power generation to obtain a preliminary prediction value; Searching for the meteorological day data most similar to the meteorological variable data related to the new energy power generation from the historical meteorological database, obtaining the actual value of the power generation output of the most similar meteorological day data, and using a long short-term memory neural network to predict the power generation output of the most similar meteorological day data to obtain the predicted value of the power generation output of the most similar meteorological day data; When it is monitored that the deviation between the preliminary prediction value of the new energy power generation at a certain moment and the actual value of the power generation output of the most similar meteorological day data exceeds the threshold, a compensation mechanism is automatically triggered, and the prediction error between the actual value of the power generation output of the most similar meteorological day data and the predicted value of the power generation output of the most similar meteorological day data is used to compensate the preliminary prediction value to obtain the final power prediction value.

[0012] Further, performing multi-objective optimization to determine the charging and discharging power of energy storage and the power of the connection line between the source-network-load-storage system and the distribution network includes: Taking power balance constraints, equipment processing upper and lower limit constraints, grid transformer output constraints, and energy storage equipment constraints as constraint conditions, establishing an objective function with the minimum operating cost of the source-network-load-storage system and the minimum environmental protection cost, where the operating cost of the source-network-load-storage system includes the new energy power generation cost calculated from the final power prediction value; Performing multi-objective optimization on the objective function with constraint conditions to obtain the optimal Pareto solution set, and the optimal Pareto solution set determines the charging and discharging power of energy storage and the power of the connection line between the source-network-load-storage system and the distribution network.

[0013] Further, the meteorological variable data related to the new energy power generation is to judge the strength of the linear relationship between meteorological variable data according to the size of the Pearson correlation coefficient of the meteorological variable data, and select the meteorological variable data more related to the new energy power generation, and the more related means complete positive linear correlation and complete negative linear correlation.

[0014] Further, search for the meteorological day data that is most similar to the meteorological variable data related to the new energy power generation power from the historical meteorological database, including: Establish a similar-day meteorological objective function, and the similar-day meteorological objective function is to minimize the sum of the Euclidean distances from each data point in the historical meteorological database to each meteorological variable data.

[0015] Further, the prediction error between the actual value of the power generation output of the most similar meteorological day data and the predicted value of the power generation output of the most similar meteorological day data is used to compensate the preliminary prediction value to obtain the final power prediction value, including: The prediction error is weighted and then superimposed on the preliminary prediction value to obtain the final power prediction value, and the weighting weight is the ratio of the similar-day meteorological objective function to the maximum allowable correction amount.

[0016] Further, perform multi-objective optimization on the objective function with constraint conditions, including: Randomly generate an initial population of size N. After non-dominated sorting, selection, crossover, and Gaussian mutation, generate an offspring population, and combine the two populations together to form a population of size 2N. The Gaussian mutation uses a random value generated by a Gaussian distribution to change the individual genes; Perform fast non-dominated sorting, and at the same time calculate the crowding degree of the individuals in each non-dominated layer. Select individuals according to the non-dominated relationship and the crowding degree of the individuals to form a new parent population; Generate a new offspring population through a genetic algorithm, combine the offspring population with the parent population, and use a jittered local search strategy for the non-dominated solutions of the current population. Repeat the above operations until the end condition is met; Use a jittered local search strategy for the non-dominated solutions of the current population, including: Set the maximum number of jitter attempts; Find the non-dominated solutions in the current population; For each non-dominated solution, calculate and record the hypervolume; Perturb the non-dominated solution with Gaussian mutation, calculate the hypervolume index. If the current hypervolume index is larger than before, eliminate the old solution and insert the new solution after Gaussian mutation perturbation into the population.

[0017] Compared with the prior art, each embodiment of the present application has at least the following beneficial effects: Through the digital twin technology, this application has successfully constructed a closed-loop optimization mechanism of "perception - prediction - optimization - feedback" for the source-grid-load-storage system, realizing the dynamic mapping of the entire life cycle of the physical system and the dynamic coupling between the virtual system and the physical system. The optimization method proposed in this application can fully exploit the temporal characteristics of temporal data for preliminary prediction, and when the prediction result error is too large due to meteorological fluctuations and other situations, it can compensate the predicted value according to the similar-day meteorological search algorithm, thereby determining the final power prediction value of new energy power generation and avoiding the situation where the prediction error gradually increases after long-term operation; based on the new energy power generation power prediction result, through the multi-objective optimization objective function, it can effectively reduce the system operation cost and environmental protection cost. Description of the Drawings

[0018] Figure 1 It is a schematic structural diagram of the digital twin control system of the source-grid-load-storage system provided by the embodiment of this application; Figure 2 It is a schematic flow chart of an operation optimization method for the source-grid-load-storage system based on digital twin provided by the embodiment of this application; Figure 3 It is a flow chart of a method for predicting the new energy power generation power according to the information of the physical entity provided by the embodiment of this application. Detailed Embodiments

[0019] In order to make the purpose, technical solutions and advantages of this application clearer, the following further details this application in conjunction with embodiments. It should be understood that the specific embodiments described here are only used to explain this application and are not used to limit this application.

[0020] This application is used for the source-grid-load-storage system and optimizes the operation of the source-grid-load-storage system to solve the problem that traditional modeling methods are limited by the problem scale, variable dimension, deterministic boundary, and dependence on accurate physical models, and it is difficult to ensure the computational reliability on the premise of computational efficiency when dealing with high-dimensional, time-varying, and non-linear problems.

[0021] See Figure 1 As shown, it is a schematic structural diagram of the digital twin control system of the source-grid-load-storage system provided by the embodiment of this application, and a schematic diagram of an operation optimization system for the source-grid-load-storage system based on digital twin, including: A test entity, which collects the information of the physical entity through various sensors and inputs the information of the physical entity into the virtual entity. The information of the physical entity includes meteorological data, historical new energy power generation power, and equipment parameters in the physical entity; Among them, the physical entity, that is, the physical entity of the source-network-load-storage system, which adopts an AC-DC hybrid architecture and includes a photovoltaic array, a wind turbine, an external power grid, an AC load, a DC load, a battery energy storage, and an inverter. Among them, the photovoltaic array and the battery energy storage are connected to the DC bus of the external power grid, the wind turbine and the AC load are connected to the AC bus of the external power grid, and the DC bus and the AC bus are coupled through an inverter. This physical entity significantly improves the utilization rate of renewable energy by coordinating the combined output of photovoltaic and wind power. At the same time, relying on the stable power support of the external power grid and configuring a battery energy storage device with an appropriate capacity, the charge and discharge strategy can be flexibly adjusted according to the load demand and the output of wind and solar power generation; The virtual entity includes a prediction model for predicting the new energy power generation according to the information of the physical entity; a multi-objective optimization model for optimizing and determining the charge and discharge power of the energy storage and the connection line power between the source-network-load-storage system and the distribution network in real time; and a health index calculation model for calculating the health index of the physical entity, dividing the health state level according to the health index, and conducting a health state assessment; The digital twin model, that is, the digital twin modeling model of the source-network-load-storage system, is a high-fidelity mapping of the physical entity of the source-network-load-storage system, including a wind power generation model, a photovoltaic power generation model, a battery energy storage model, an external power grid model, an AC load model, and a DC load model. That is, it is a model mapping corresponding to the physical entity. According to the predicted new energy power generation power, the charge and discharge power of the energy storage, the connection line power between the source-network-load-storage system and the distribution network, and the health state level of the virtual entity, the physical entity can be visually displayed after mapping.

[0022] The working principle of an operation optimization system of a source-network-load-storage system based on digital twin provided by an embodiment of the present application is as follows: On the one hand, the test entity collects meteorological data such as light intensity similar to that of the physical entity through various sensors, and inputs it together with historical new energy power generation power, various equipment parameters, etc. into the virtual entity. The virtual entity performs real-time calculation and analysis based on its own algorithm, and obtains the predicted value of new energy power generation power through decomposition-reconstruction-prediction-correction power prediction; on the other hand, according to the predicted power result and equipment parameters, combined with the multi-objective optimization algorithm, the operation state of the source-network-load-storage system is simulated and calculated to obtain the system operation optimization result, and the optimization result and the health state assessment result are input into the digital twin model together. Since it is a high-fidelity mapping of the physical entity, it can provide references for the regulation and operation, comprehensive evaluation, intelligent operation and maintenance, and fault recovery of the physical entity of the source-network-load-storage system. At the same time, the operation state of the actual equipment can also be fed back to improve and optimize the digital twin model.

[0023] The prediction model is used to predict the new energy power generation according to the information of physical entities, including decomposing the new energy power generation into several intrinsic mode function components and residual components; It is used to reconstruct the intrinsic mode function components and residual components with similar sample entropy values into multiple power components; It is used to take the power components and meteorological variable data related to the new energy power generation as inputs, and use a long short-term memory neural network to make a preliminary prediction of the new energy power generation to obtain a preliminary prediction value; It is used to search the historical meteorological database for the meteorological day data most similar to the meteorological variable data related to the new energy power generation, obtain the actual value of the power generation output of the most similar meteorological day data, and use a long short-term memory neural network to predict the power generation output of the most similar meteorological day data to obtain a predicted value of the power generation output of the most similar meteorological day data; When it is monitored that the deviation between the preliminary prediction value of the new energy power generation at a certain moment and the actual value of the power generation output of the meteorological day data of the most similar day exceeds the threshold, it automatically triggers a compensation mechanism, that is, the prediction error between the actual value of the power generation output of the most similar meteorological day data and the predicted value of the power generation output of the most similar meteorological day data is used to compensate the preliminary prediction value to obtain the final power prediction value; The multi-objective optimization algorithm performs multi-objective optimization using the objective function with constraints to obtain the optimal Pareto solution set, and the optimal Pareto solution set determines the charge and discharge power of the energy storage and the connection line power between the source-network-load-storage system and the distribution network; The objective function with constraints is to establish an objective function with the minimum operating cost of the source-network-load-storage system and the minimum environmental protection cost under the constraints of power balance constraint, upper and lower limits of equipment processing, transformer output constraint of the power grid, and energy storage equipment constraint, where the operating cost of the source-network-load-storage system includes the new energy power generation cost calculated from the final power prediction value; The health index calculation model quantifies the real-time health status of equipment and predicts potential failure risks through multi-dimensional data analysis and model calculation.

[0024] See Figure 2 , the flow chart of the method for optimizing the operation of the source-network-load-storage system based on digital twin in the embodiment of the present application. A method for optimizing the operation of the source-network-load-storage system based on digital twin includes: S201 Collect information of physical entities; S202 Predict the new energy power generation according to the information of physical entities; perform multi-objective optimization to determine the charge and discharge power of the energy storage and the connection line power between the source-network-load-storage system and the distribution network; and calculate the health index of the physical entity, and divide the health status level according to the health index; S203 Map the physical entity according to the output result of the virtual entity.

[0025] Among them, referring to Figure 3 the flowchart of the method for predicting new energy power generation according to the information of physical entities shown in Figure 3 , which includes: S301 Decompose the new energy power generation into several intrinsic mode function components and a residual component; Collect meteorological variable data related to new energy power generation through various sensors. The meteorological variable data related to new energy power generation includes solar irradiance, temperature, humidity, air pressure, etc. A large amount of data is included in the collected meteorological variable data related to new energy power generation, and the calculation involved in this data is a huge amount of calculation. In order to reduce the subsequent calculation amount and improve the calculation speed, in one embodiment, the Pearson analysis method is used to select the meteorological variable data more relevant to new energy power generation from the collected meteorological variable data related to new energy power generation for decomposition to obtain several intrinsic mode function components and a residual component.

[0026] In the Pearson analysis method, the Pearson correlation coefficient is a statistic used to measure the linear relationship between two random variables, including perfect positive linear correlation, that is, when one variable increases, the other variable also increases proportionally; perfect negative linear correlation, that is, when one variable increases, the other variable decreases proportionally; no linear correlation, that is, there is no linear correlation between the two variables.

[0027] According to the magnitude of the Pearson correlation coefficient through the Pearson analysis method, the strength of the linear relationship between variables can be judged, and the meteorological variable data more relevant to new energy power generation can be selected. The so-called more relevant here refers to perfect positive linear correlation and perfect negative linear correlation.

[0028] The decomposition method can adopt the complete ensemble empirical mode decomposition with adaptive noise. The complete ensemble empirical mode decomposition with adaptive noise introduces adaptive noise and multiple iterations. In each iteration, white noise is added to the intrinsic mode function (IMF) after EMD decomposition. This way of gradually adding noise can effectively reduce the influence of white noise on the decomposition result and improve the accuracy and stability of the decomposition.

[0029] The complete ensemble empirical mode decomposition with adaptive noise introduces the white noise for adaptive auxiliary decomposition in each decomposition stage and performs ensemble averaging on the decomposition results of each stage. This method overcomes the problems of low decomposition efficiency and easy white noise residue in the ensemble empirical mode decomposition. At the same time, it further suppresses the occurrence of mode mixing. The uncertainty and randomness of meteorological variables lead to the uncertainty and randomness of new energy power generation. By decomposing the new energy power generation into several intrinsic mode function components and a residual component through the complete ensemble empirical mode decomposition with adaptive noise, the complexity of the original new energy power generation data can be reduced.

[0030] S302 reconstructs the intrinsic mode function components and residual components with similar sample entropy values into multiple power components; By comparing the complexity of the intrinsic mode function components and residual components with power time series obtained by adaptive noise complete ensemble empirical mode decomposition through sample entropy values, and reconstructing the components with similar sample entropy values into new components, the number of components can be reduced, thereby improving the overall new energy power prediction efficiency. The calculation steps for reconstructing the intrinsic mode function components and residual components with similar sample entropy values into multiple power components are as follows: Step 1 Reconstruct the time series with a time length of into an -dimensional vector represented by as follows: , where , is the selected spatial dimension vector, that is, the window length, representing the continuous points starting from the -th point. The time series is the intrinsic mode function component and residual component with a power time series.

[0031] Step 2 Calculate the Euclidean distance between and , and define the maximum distance between each component as the maximum contribution component distance as shown in the following formula: , where and are two arbitrary components, , .

[0032] Step 3 Using the given metric value , count the number corresponding to each and calculate the probability of matching all using the following formula: : , Step 4 Solve the , , …, , -corresponding average value using the following formula: , Step 5 reconstructs the time series into dimensions, and repeats Steps 2 to 4 to solve for , and the sample entropy value is obtained using the following formula: , In practical applications, generally take 1, 2; Take 0.10 - 0.25 times the standard deviation of the time series.

[0033] S303 uses the power component and meteorological variable data related to the new - energy power generation as inputs, and uses a long short - term memory neural network to perform a preliminary prediction on the new - energy power generation to obtain a preliminary prediction value; Based on the preliminary prediction of the new - energy power generation by the long short - term memory neural network to obtain the preliminary prediction value, the input data is the power component and meteorological variable data related to the new - energy power generation, and the output is the power predicted for different power components. Finally, the powers predicted for different power components output are superimposed and summed to obtain the preliminary prediction result of the new - energy power generation.

[0034] The long short - term memory neural network adopted includes a forget gate, an input gate, and an output gate.

[0035] S304 searches the historical meteorological database for the meteorological day data most similar to the meteorological variable data related to the new - energy power generation, obtains the actual value of the power generation output of the most similar meteorological day data, and uses a long short - term memory neural network to predict the most similar meteorological day data to obtain the predicted value of the power generation output of the most similar meteorological day data.

[0036] The historical meteorological database covers meteorological day data related to the new - energy power generation. According to the meteorological variable data, the meteorological day data most similar to the meteorological variable data related to the new - energy power generation is searched in the historical meteorological database. The meteorological day data includes solar irradiance, temperature, humidity, and air pressure, etc. There is the new - energy power generation corresponding to the meteorological day data in the historical meteorological database. These are all obtained by statistical analysis of historical data. Generally speaking, due to the regional nature of meteorology, the historical meteorological database adopted is the historical meteorological database of the area where the source - grid - load - storage system operates optimally. The actual value of the power generation output of the most similar meteorological day data is also obtained from the corresponding actual value of the power generation output of the most similar meteorological day data in the historical meteorological database. Here, the actual value of the power generation output is the new - energy power generation.

[0037] Among them, the long short - term memory neural network is used to predict the most similar meteorological day data to obtain the predicted value of the power generation output of the most similar meteorological day data; The long short-term memory neural network here is the same as that in S303. There may be an error between the predicted power generation output value of the most similar meteorological day data and the actual power generation output value of the most similar meteorological day data. Through the processing of the most similar meteorological day data, it is used to correct or compensate the preliminary prediction value obtained from the meteorological variable data related to new energy power generation in the next step; S305 When it is monitored that the deviation between the preliminary prediction value of the new energy power generation power at a certain moment and the actual power generation output value of the most similar meteorological day data exceeds the threshold, the compensation mechanism is automatically triggered, and the prediction error between the actual power generation output value of the most similar meteorological day data and the predicted power generation output value of the most similar meteorological day data is used to compensate the preliminary prediction value to obtain the final power prediction value; In one embodiment, multi-objective optimization is performed to determine the charge and discharge power of the energy storage and the connection line power between the source-network-load-storage system and the distribution network, including: taking power balance constraints, upper and lower limits of equipment processing constraints, grid transformer output constraints, and energy storage equipment constraints as constraint conditions, and establishing an objective function with the minimum operation cost of the source-network-load-storage system and the minimum environmental protection cost, where the operation cost of the source-network-load-storage system includes the new energy power generation cost calculated from the final power prediction value; In one embodiment, multi-objective optimization is performed on the objective function with constraint conditions to obtain the optimal Pareto solution set, and the optimal Pareto solution set determines the charge and discharge power of the energy storage and the connection line power between the source-network-load-storage system and the distribution network; In one embodiment, key parameters affecting the lifespan are selected according to the equipment types in the source-network-load-storage system, and the data is normalized to obtain the standardized values of each parameter; based on the defined health index calculation model, the health index of the equipment is calculated, the health status level is divided according to the health index, and it is visually displayed in the digital twin virtual model. When the health index is lower than the preset threshold, a maintenance warning or an adaptive control instruction is triggered. Generally, the health status assessment of each device in the source-network-load-storage system includes: selecting the key parameters of each device; normalizing the processing; calculating the health index; determining the health status level of each device and taking corresponding countermeasures.

[0038] In one embodiment, searching for the most similar meteorological day data of the meteorological variable data related to new energy power generation from the historical meteorological database includes: Establishing a similar-day meteorological objective function, and the similar-day meteorological objective function is to minimize the sum of the Euclidean distances from each data point in the historical meteorological database to each meteorological variable data.

[0039] Similar-day meteorological objective function: , The objective function needs to meet the following conditions: , In the formula: is the meteorological objective function of similar days, is the meteorological data set; is the data including solar irradiance, temperature, humidity and air pressure; is the Pearson correlation coefficient, ; is the Euclidean distance from each data point to each ; is the number of meteorological data types; is the number of data points; is the maximum allowable correction amount, and corrections outside this range are invalid.

[0040] Search for the most similar meteorological day data in the historical meteorological database and the meteorological variable data related to new energy power generation collected currently by solving the meteorological objective function of similar days.

[0041] In an embodiment, when it is monitored that the deviation between the predicted value of new energy power generation at a certain moment and the actual power generation output value of the meteorological day data of the most similar day exceeds the threshold, the compensation mechanism is automatically triggered, that is, the predicted value at this moment is compensated by using the meteorological search algorithm of similar days. Among them, the threshold function is defined as: , is the actual power generation output value of the most similar meteorological day data at this moment, is the preliminary predicted value at this moment. When , the correction is triggered.

[0042] In an embodiment, the preliminary predicted value of the prediction error is weighted and corrected, and the weight size is determined according to the meteorological similarity degree, so as to obtain the final power prediction value of the new energy output , and its calculation formula is: , Among them, is the preliminary predicted value, is the actual power generation output value of the most similar meteorological day data, is the predicted value of the power generation output of the most similar meteorological day data, The final power prediction value of new energy power generation is a part of the operation cost of the source-network-load-storage system. The uncertain new energy power generation is calculated as the final power prediction value of new energy power generation through prediction for the new energy power generation cost.

[0043] With power balance constraints, equipment processing upper and lower limit constraints, grid transformer output constraints and energy storage equipment constraints as constraints, an objective function is established to minimize the operating cost and environmental protection cost of the source-grid-load-storage system.

[0044] The lowest operating cost of the source-grid-load-storage system refers to the cost of using grid electricity, the operating costs of other distributed power equipment, the operating and maintenance costs of power generation equipment, and the operating and maintenance costs of energy storage equipment during the operation of the source-grid-load-storage system. The operating cost of the source-grid-load-storage system is: , in, yes Distributed power generation at all times Operation and maintenance cost coefficient; yes Distributed power generation at all times The output power, is the number of distributed generation sources; yes Purchase electricity at any time; yes The electricity price at the time; is the amount of electricity purchased by the source-grid-load-storage system, The cost of renewable energy generation is the product of the predicted final power output of renewable energy generation and the operation and maintenance cost coefficient.

[0045] The environmental protection cost is the main pollutants in the operation of the source-grid-load-storage system, including 、 and The processing cost is as follows: , In the formula, is the number of pollutant types ( 、 or ), is a distributed power source The pollutant type is Unit processing cost per hour; is the emission factor.

[0046] The power balance constraint refers to maintaining a dynamic balance between the real-time load power of all energy-consuming devices in the source-grid-load-storage system, the real-time power generation power of the power supply equipment, the real-time power of the grid input system, and the real-time charge and discharge power of the energy storage equipment. The sum of the power of the four types of equipment is always zero. Only under this premise can the entire source-grid-load-storage system be kept in a constant voltage state, allowing the source-grid-load-storage system to operate stably. The formula is as follows: , Among them, represents the operating power of the th load device at moment, represents the real-time power of the power grid at moment, represents the real-time power of the th power supply device at moment, is the real-time power of the energy storage device at moment.

[0047] The upper and lower limits of the device output constraint mean that all adjustable devices have their own output range during the control process. For example, the charging power of charging pile devices, the output power of air conditioning devices, etc. When designing the constraint conditions, the power ranges of these devices should be fully considered to ensure that after the optimization calculation, the regulation parameters of the corresponding devices are within the upper and lower limits of the device output, so as to ensure the normal operation of the devices. The formula is as follows: , Among them, represents the output power of the th device (including power supply devices, power grid devices, and energy storage devices) at moment, represents the minimum output power of the th device, represents the maximum output power of the th device.

[0048] The power output constraint of the power grid transformer means that for the park, the energy it can draw from the power grid and the energy it can send to the power grid are both limited. The device that restricts the received and sent energy is the distribution transformer in this area. Therefore, during the optimization calculation process, the power output capacity of the power grid transformer should be considered and used as a constraint condition to ensure the stable operation of the system. The formula is as follows: , Among them, is the output power of the transformer at moment, is the minimum output power of the transformer, is the maximum output power of the transformer.

[0049] The energy storage device constraint refers to the relevant constraints on the SOC value of the energy storage device. Through research on the products of various energy storage device manufacturers, it is found that when the SOC value of the battery is less than 20%, the energy storage device will no longer discharge, and when the battery SOC value is greater than 95%, the energy storage device will no longer charge. The formula is as follows: , wherein, is the percentage of the remaining real-time battery power of the energy storage device at moment, is the minimum remaining battery power of the energy storage device, is the maximum remaining battery power of the energy storage device. is the real-time operating power of the energy storage device at moment, is the minimum output power of the energy storage device, is the maximum output power of the energy storage device.

[0050] Select an optimization algorithm based on the objective function and the characteristics of related constraints, construct an improved multi-objective non-dominated sorting genetic algorithm, perform multi-objective optimization decision-making, obtain the optimal Pareto solution set, and determine the charge and discharge power of the energy storage, the tie-line power between the system and the distribution network, etc.

[0051] The characteristics of the objective function and related constraints mean that since the designed objective function includes two objectives, namely the operation cost of the source-network-load-storage system and the environmental protection cost, multi-objective optimization is selected. Since the source-network-load-storage system requires a relatively large number of types and parameters of regulating equipment, the optimization ability of the optimization algorithm is required to be relatively high.

[0052] In one embodiment, multi-objective optimization is performed on the objective function with constraint conditions, including: Randomly generate an initial population with a size of A. After non-dominated sorting, selection, crossover, and Gaussian mutation, generate an offspring population, and combine the two populations together to form a population with a size of 2A; Perform fast non-dominated sorting, and at the same time calculate the crowding degree of individuals in each non-dominated layer. Select appropriate individuals according to the non-dominated relationship and the crowding degree of individuals to form a new parent population; Generate a new offspring population through the basic operations of the genetic algorithm. Combine the offspring population with the parent population. Use a jittered local search strategy for the non-dominated solutions of the current population, and repeat the above operations until the end condition is met. The basic operations here refer to initializing the population, fitness evaluation, selection operation, crossover operation, and mutation operation, and finally generating a new offspring population.

[0053] Select a suitable algorithm based on the characteristics of the proposed optimization objective function and related constraints. The algorithm is required to be a multi-objective optimization algorithm with high optimization ability, as stable as possible, and with as little computational effort as possible. In the embodiments of the present application, Gaussian mutation is used instead of polynomial mutation to breed the next generation of individuals. A jittered local search strategy is adopted, and Gaussian mutation is used to perturb the non-dominated solutions. Calculate the hypervolume index. If the hypervolume index is larger than before, it proves that there is still a certain room for improvement in the current population. Then, the old solutions are eliminated, and the perturbed new solutions are inserted into the population, which helps the algorithm to jump out of the local optimum.

[0054] Gaussian mutation uses random values generated by a Gaussian distribution to change the individual genes. The details of Gaussian mutation are as follows: Let be an element of a single vector with real number encoding in the

[0055] -th generation. In the formula, obeys the Gaussian distribution , is called the scale parameter.

[0056] Adopting a jittered local search strategy and using Gaussian mutation to perturb the non-dominated solutions can be decomposed into the following steps: Set the maximum number of jitter attempts; Find the non-dominated solutions in the current population; For each non-dominated solution, calculate and record the hypervolume; Use Gaussian mutation to perturb the non-dominated solutions, calculate the hypervolume index. If the current hypervolume index is larger than before, eliminate the old solutions and insert the new solutions perturbed by Gaussian mutation into the population.

[0057] Perform multi-objective optimization on the objective function with constraint conditions to obtain the optimal Pareto solution set. The optimal Pareto solution set determines the charge and discharge power of the energy storage and the power of the connection line between the source-network-load-storage system and the distribution network.

[0058] In one embodiment, key parameters affecting the lifespan are selected according to the device types in the source-network-load-storage system. The key parameters include: Battery energy storage system: state of charge, state of health, internal resistance, temperature; Photovoltaic inverter: output efficiency, harmonic distortion rate, radiator temperature; Wind turbine bearing: vibration amplitude, spectral characteristics, lubricant viscosity.

[0059] In one embodiment, based on a defined health index calculation model, the health index of a device is calculated, the health status level is divided according to the health index, and the result is sent back to the digital twin model. When the health index is lower than a preset threshold, a maintenance warning instruction is triggered. The health index calculation model is as follows: , In the formula, is the health index, is the normalized value of the th key parameter; taking the health index calculation of a battery energy storage system as an example, there are four core parameters, so is 4; is the weight, , and the weight determination method can adopt a statistical regression method based on failure history.

[0060] The division of the health status level according to the health index can adopt the following grading criteria: When is in the range of 0.6 - 1.0, the status level is healthy, and the countermeasure is normal operation; When is in the range of 0.3 - 0.6, the status level is early warning, and the countermeasure is planned maintenance; When is in the range of 0 - 0.3, the status level is dangerous, and the countermeasure is immediate shutdown for maintenance.

[0061] By evaluating the health status of each device and sending back the evaluation results, the transformation from "post - fault handling" to "pre - fault prevention" can be achieved, reducing unplanned outages.

[0062] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. An operation optimization system for a source-network-load-storage system based on digital twin, characterized in that, The system includes: A test entity that collects information of a physical entity through various sensors and inputs the information of the physical entity into a virtual entity. The information of the physical entity includes meteorological data, historical new energy power generation and equipment parameters in the physical entity; A virtual entity that includes a prediction model for predicting new energy power generation according to the information of the physical entity; a multi-objective optimization model for optimizing and determining the charge and discharge power of energy storage and the connection line power between the source-network-load-storage system and the distribution network in real time; and a health index calculation model for calculating the health index of the physical entity and dividing the health status level according to the health index; A digital twin model for mapping the physical entity according to the output result of the virtual entity, including a wind power generation model, a photovoltaic power generation model, a battery energy storage model, an external power grid model, an AC load model and a DC load model.

2. The operation optimization system of the source-grid-load-storage system based on digital twin according to claim 1, wherein, The physical entity is a source-network-load-storage system, which adopts an AC-DC hybrid architecture and includes a photovoltaic array, a wind turbine, an external power grid, an AC load, a DC load, battery energy storage and an inverter. Among them, the photovoltaic array and the battery energy storage are connected to the DC bus of the external power grid, the wind turbine and the AC load are connected to the AC bus of the external power grid, and the DC bus and the AC bus are coupled through an inverter.

3. The source-network-load-storage system operation optimization system based on digital twin according to claim 1, wherein The prediction model is used to predict the new energy power generation according to the information of the physical entity, including: Decomposing the new energy power generation into a number of intrinsic mode function components and a residual component; Reconstructing the intrinsic mode function components and the residual component with similar sample entropy values into multiple power components; Using the power components and the meteorological variable data related to the new energy power generation as inputs, and using a long short-term memory neural network to make a preliminary prediction of the new energy power generation to obtain a preliminary prediction value; Searching the historical meteorological database for the meteorological day data most similar to the meteorological variable data related to the new energy power generation, obtaining the actual power generation output value of the most similar meteorological day data, and using a long short-term memory neural network to predict the power generation output value of the most similar meteorological day data; When it is monitored that the deviation between the preliminary prediction value of the new energy power generation at a certain moment and the actual power generation output value of the most similar meteorological day data exceeds the threshold, the compensation mechanism is automatically triggered, and the prediction error between the actual power generation output value of the most similar meteorological day data and the predicted power generation output value of the most similar meteorological day data is used to compensate the preliminary prediction value to obtain the final power prediction value.

4. The operation optimization system of the source-network-load-storage system based on digital twin according to claim 1, characterized in that, The multi-objective optimization model performs multi-objective optimization on the objective function with constraints to obtain the optimal Pareto solution set, and the optimal Pareto solution set determines the charge and discharge power of energy storage and the connection line power between the source-network-load-storage system and the distribution network.

5. The operation optimization system of the source-grid-load-storage system based on digital twin according to claim 4, wherein The objective function with constraint conditions is to establish an objective function with the minimum operation cost of the source-grid-load-storage system and the minimum environmental protection cost under the constraint conditions of power balance constraint, upper and lower limits of equipment processing constraint, power output constraint of the grid transformer, and energy storage equipment constraint. The operation cost of the source-grid-load-storage system includes the new energy generation cost calculated based on the final power prediction value.

6. A method for optimizing the operation of a source-network-load-storage system based on digital twin, characterized in that, This method includes: Collecting information of physical entities; Predicting the new energy generation power according to the information of physical entities; performing multi-objective optimization to determine the charge and discharge power of the energy storage and the connection line power between the source-grid-load-storage system and the distribution network; and calculating the health index of the physical entities and dividing the health status levels according to the health index. Mapping the physical entities according to the output results of the virtual entities.

7. The operation optimization method of the source-network-load-storage system based on digital twin according to claim 6, characterized in that Predicting the new energy generation power according to the information of physical entities, including: decomposing the new energy generation power into several intrinsic mode function components and residual components; Reconstructing the intrinsic mode function components and residual components with similar sample entropy values into multiple power components; Taking the power components and the meteorological variable data related to the new energy generation power as inputs, and using a long short-term memory neural network to preliminarily predict the new energy generation power to obtain a preliminary prediction value; Searching the historical meteorological database for the meteorological day data most similar to the meteorological variable data related to the new energy generation power, obtaining the actual power generation output value of the most similar meteorological day data, and using a long short-term memory neural network to predict the power generation output value of the most similar meteorological day data; When it is monitored that the deviation between the preliminary prediction value of the new energy generation power at a certain moment and the actual power generation output value of the most similar meteorological day data exceeds the threshold, the compensation mechanism is automatically triggered, and the prediction error between the actual power generation output value of the most similar meteorological day data and the predicted power generation output value of the most similar meteorological day data is used to compensate the preliminary prediction value to obtain the final power prediction value.

8. The operation optimization method of the source-network-load-storage system based on digital twin according to claim 7, wherein, Performing multi-objective optimization to determine the charge and discharge power of the energy storage and the connection line power between the source-grid-load-storage system and the distribution network, including: Establishing an objective function with the minimum operation cost of the source-grid-load-storage system and the minimum environmental protection cost under the constraint conditions of power balance constraint, upper and lower limits of equipment processing constraint, power output constraint of the grid transformer, and energy storage equipment constraint. The operation cost of the source-grid-load-storage system includes the new energy generation cost calculated based on the final power prediction value; Performing multi-objective optimization on the objective function with constraint conditions to obtain the optimal Pareto solution set, and the optimal Pareto solution set determines the charge and discharge power of the energy storage and the connection line power between the source-grid-load-storage system and the distribution network.

9. The operation optimization method of the source-network-load-storage system based on digital twin according to claim 7, characterized in that, The meteorological variable data related to the new energy generation power is to judge the strength of the linear relationship between meteorological variable data according to the size of the Pearson correlation coefficient of the meteorological variable data, and select the meteorological variable data more related to the new energy generation power. The so-called more related means complete positive linear correlation and complete negative linear correlation.

10. The operation optimization method of the source-network-load-storage system based on digital twin according to claim 7, characterized in that, Searching the historical meteorological database for the meteorological day data most similar to the meteorological variable data related to the new energy generation power, including: A similar-day meteorological objective function is established, and the similar-day meteorological objective function minimizes the sum of the Euclidean distances from each data point in the historical meteorological database to each meteorological variable data.

11. The operation optimization method of the source-network-load-storage system based on digital twin according to claim 7, characterized in that, The prediction error between the actual power generation output value of the most similar meteorological day data and the predicted power generation output value of the most similar meteorological day data is used to compensate the preliminary prediction value to obtain the final power prediction value, including: The prediction error is weighted and then superimposed on the preliminary prediction value to obtain the final power prediction value, and the weight of the weighting is the ratio of the similar-day meteorological objective function to the maximum allowable correction amount.

12. The operation optimization method of the source-grid-load-storage system based on digital twin according to claim 6, characterized in that, Multi-objective optimization is performed on the objective function with constraint conditions, including: An initial population of size N is randomly generated. Through non-dominated sorting, selection, crossover, and Gaussian mutation, an offspring population is generated, and the two populations are combined to form a population of size 2N. The Gaussian mutation uses random values generated by a Gaussian distribution to change the individual genes. Fast non-dominated sorting is performed, and at the same time, the crowding degree of the individuals in each non-dominated layer is calculated. According to the non-dominated relationship and the crowding degree of the individuals, individuals are selected to form a new parent population. A new offspring population is generated through a genetic algorithm. The offspring population and the parent population are combined. A jittered local search strategy is used for the non-dominated solutions of the current population. The above operations are repeated until the end condition is met. A jittered local search strategy is used for the non-dominated solutions of the current population, including: Setting the maximum number of jitter attempts. Finding the non-dominated solutions in the current population. For each non-dominated solution, calculating and recording the hypervolume. Perturbing the non-dominated solution with Gaussian mutation, calculating the hypervolume index. If the current hypervolume index is larger than before, the old solution is eliminated, and the new solution after Gaussian mutation perturbation is inserted into the population.

Images