An integrated energy system operation optimization method based on multi-strategy combination optimization

Abstract

The present invention provides an integrated energy system operation optimization method based on multiple strategy combination optimization, which relates to the field of integrated energy system control technology. The optimization method is: establishing an available power prediction model for the integrated energy system; obtaining the load demand of the integrated energy system to be optimized, and constructing an operation optimization model based on the purchased energy price and the load demand; taking the control strategy of each energy conversion device as the variable to be optimized, according to the gray wolf optimization algorithm that introduces the convergence factor nonlinear update strategy and the elite wolf rollback strategy and the available power prediction model, the operation optimization model is solved to obtain the optimal control strategy of each energy conversion device; according to the optimal control strategy of each energy conversion device, the corresponding energy conversion device is controlled to operate. The present invention takes into account the convergence speed and optimization accuracy of the gray wolf optimization algorithm by introducing the convergence factor nonlinear update strategy and the elite wolf rollback strategy, so as to ensure the accuracy of the subsequent integrated energy system operation optimization results.

Description

Technical Field

[0001] The present invention relates to the technical field of integrated energy system control, and in particular to an integrated energy system operation optimization method based on multiple strategy combination optimization. Background Art

[0002] Energy conversion equipment can realize the mutual conversion between different energy forms, and is the basis for the integrated energy system to achieve multi-energy complementarity and coordinated optimization. The control strategy of energy conversion equipment directly affects the operating efficiency and carbon emissions of the integrated energy system. In order to optimize the operation of the integrated energy system, iterative optimization algorithms are often used to optimize the control strategy of energy conversion equipment, so that the integrated energy system can reduce operating costs and carbon trading costs while meeting load requirements.

[0003] Among them, the Gray Wolf Optimization Algorithm is suitable for various scenarios that require complex decision-making and optimization, and can be used to optimize the control strategy of energy conversion equipment. However, due to the influence of nonlinear factors in various links of the actual energy conversion process, such as changes in operating efficiency, changes in climbing rate, and energy flow loss, if the traditional Gray Wolf Optimization Algorithm is directly used to iterate the optimization of the control strategy of energy conversion equipment, it is easy to fall into the local optimum.

[0004] Most existing technologies maintain the balance between global search and local search by changing the change strategy of the convergence factor to avoid falling into the local optimum. However, if the convergence factor changes too quickly, the problem of falling into the local optimum will occur quickly. If the convergence factor changes too slowly, the problem of slow convergence will occur. By changing the change strategy of the convergence factor, it is difficult to take into account the convergence speed and optimization accuracy of the gray wolf optimization algorithm. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art that the convergence speed and optimization accuracy of the gray wolf optimization algorithm are difficult to be taken into account in order to avoid falling into the local optimum, and the operation optimization effect of the integrated energy system cannot be guaranteed. A comprehensive energy system operation optimization method based on multiple strategy combination optimization is provided, and the control strategy of the energy conversion equipment in the integrated energy system is optimized and formulated by the gray wolf optimization algorithm. The convergence speed of the gray wolf optimization algorithm is improved by increasing the nonlinear update strategy of the convergence factor. At the same time, the elite wolf rollback strategy is further introduced to avoid the problem of falling into the local optimum caused by the excessively fast convergence in the process of updating the convergence factor. While ensuring the convergence speed, the optimization accuracy is guaranteed, and the operation optimization effect of the integrated energy system is improved.

[0006] The purpose of the present invention is achieved through the following technical solutions:

[0007] The integrated energy system operation optimization method based on multiple strategy combination optimization includes:

[0008] Step 1: Collect historical operation data of the integrated energy system to be optimized, and establish an available power prediction model of the integrated energy system based on a deep convolutional neural network;

[0009] Step 2: Obtain the load demand of the integrated energy system to be optimized, and build an operation optimization model based on the purchased energy price and load demand;

[0010] Step 3: Taking the control strategies of various energy conversion devices as the variables to be optimized, the operation optimization model is solved according to the gray wolf optimization algorithm that introduces the convergence factor nonlinear update strategy and the elite wolf rollback strategy and the available power prediction model to obtain the optimal control strategies of various energy conversion devices;

[0011] Step 4: Control the operation of the corresponding energy conversion equipment according to the optimal control strategy of each energy conversion equipment.

[0012] During the operation of the integrated energy system, the nonlinear factors in each link of the actual energy conversion process will affect the operation of the energy conversion equipment. Therefore, the nonlinear factors in each link of the energy input-conversion-output in the energy conversion process of the integrated energy system are deeply regressed and learned using a deep convolutional neural network to construct a power prediction model to ensure the accuracy of the optimization of the subsequent energy conversion equipment control strategy.

[0013] In order to achieve the low-carbon economic operation and optimization scheduling of the integrated energy system, the relevant cost is taken as the optimization target. Considering that the operation basis of the integrated energy system is to ensure that the corresponding load demand is met, the corresponding load demand is set as the corresponding constraint condition, and a specific operation optimization model is constructed. By adjusting the control strategy of the energy conversion equipment, it is ensured that the relevant cost is minimized while meeting the load demand.

[0014] The gray wolf optimization algorithm is further used to optimize the control strategy of the energy conversion equipment in the integrated energy system, and the convergence speed of the gray wolf optimization algorithm is improved by adding a nonlinear update strategy for the convergence factor. At the same time, the elite wolf rollback strategy is further introduced to avoid the problem of falling into the local optimum due to rapid convergence during the convergence factor update process. While ensuring the convergence speed, the optimization accuracy is guaranteed, thereby improving the operation optimization effect of the integrated energy system.

[0015] Finally, after determining the optimal control strategy for each energy conversion device, the corresponding energy conversion device is controlled according to the corresponding control strategy for operation control, thereby achieving low-carbon economic operation of the integrated energy system.

[0016] Furthermore, in step 3, according to the grey wolf optimization algorithm that introduces the convergence factor nonlinear update strategy and the elite wolf rollback strategy and the available power prediction model, the operation optimization model is solved to obtain the optimal control strategy for each energy conversion device, including:

[0017] Taking the control strategy sequence of energy conversion equipment as the variable to be optimized, the gray wolf population of the gray wolf optimization algorithm is generated, and the population and basic parameters are initialized;

[0018] The available power of each individual gray wolf is determined based on the available power prediction model, and the fitness of each individual gray wolf is calculated by combining the objective function of the running optimization model to divide the leader wolves and elite wolves.

[0019] Based on the convergence factor nonlinear update strategy, the positions of the remaining gray wolves are updated according to the positions of the leader wolf and the elite wolf;

[0020] Determine whether the iteration end condition is met;

[0021] If the iteration end condition is not met, the fitness value of each individual gray wolf is recalculated, the fitness and position of the leader wolf and the elite wolf are updated, and the historical position of the leader wolf is stored in the backup position set until the iteration end condition is met;

[0022] If the iteration end condition is met, the first alternative energy conversion equipment control strategy sequence is output, and a gray wolf individual is randomly selected from the backup position to replace the current elite wolf for rollback optimization. The first alternative energy conversion equipment control strategy sequence and the rollback optimization result are compared to determine the optimal energy conversion equipment control strategy sequence.

[0023] Furthermore, the gray wolf individuals are randomly selected from the backup position set to replace the current elite wolf for rollback optimization, and the first candidate energy conversion device control strategy sequence is compared with the rollback optimization result to determine the optimal energy conversion device control strategy sequence, including:

[0024] Randomly select a first replacement group and a second replacement group including different gray wolf individuals from the backup position set to replace the current elite wolves respectively;

[0025] Iterate again according to the current leading wolf and the elite wolf replaced by the first replacement group until the iteration end condition is met, and obtain the first rollback energy conversion device control strategy sequence;

[0026] At the same time, reiterate according to the current leading wolf and the elite wolf replaced by the second replacement group until the iteration end condition is met, and obtain the second rollback energy conversion device control strategy sequence;

[0027] The fitness values ​​of the first candidate energy conversion device control strategy sequence, the first rollback energy conversion device control strategy sequence and the second rollback energy conversion device control strategy sequence are compared, and the energy conversion device control strategy sequence with the smaller fitness value is selected as the optimal energy conversion device control strategy sequence.

[0028] Furthermore, the nonlinear update strategy based on the convergence factor updates the positions of the remaining gray wolf individuals according to the positions of the leader wolf and the elite wolf, including:

[0029] Update the convergence factor based on the current number of iterations and the maximum number of iterations combined with the update formula;

[0030] updating the coefficient vector based on the updated convergence factor;

[0031] According to the positions of the leader wolf and the elite wolf, the positions of the remaining gray wolf individuals are updated in combination with the updated coefficient vector.

[0032] Furthermore, the expression of the update formula is:

[0033] ;

[0034] in, is the updated convergence factor, is the initial value of the convergence factor, is the current iteration number, is the maximum number of iterations, is the fitness value of the leader wolf in the current iteration, is the fitness value of the leader wolf in the previous iteration, is the preset adjustment factor.

[0035] Furthermore, the load demand of the integrated energy system to be optimized is obtained in step 2, and an operation optimization model is constructed based on the purchased energy price and the load demand, including taking the minimization of the operation cost and carbon trading cost of the integrated energy system to be optimized as the optimization goal, and setting the constraints of the operation optimization model based on the load demand to construct the operation optimization model.

[0036] Furthermore, the constraint conditions for setting the operation optimization model based on load demand include:

[0037] A penalty function is introduced into the optimization objective, and load demand is used as the penalty function variable. Constraints that meet the load demand are set for running the optimization model.

[0038] Furthermore, the objective function of the running optimization model is expressed as:

[0039] ;

[0040] ;

[0041] ;

[0042] in, is the total cost, For operating costs, is the price of purchased natural gas power per unit, is the total amount of purchased natural gas, is the price of purchased unit electric power, is the amount of purchased electricity, is the carbon trading cost, is the carbon emission factor per unit of natural gas power purchased, is the carbon emission quota per unit of natural gas power purchased, is the carbon emission factor per unit of purchased electricity, is the carbon emission quota per unit of purchased electricity, To trade the price per kilogram of carbon dioxide, is the penalty function, and , is the total cooling load demand, To correspond to the cooling power available under the energy conversion equipment control strategy, is the total heat load demand, is the available heating power under the control strategy of the energy conversion equipment. is the total electrical load demand, It is the power available under the control strategy of the corresponding energy conversion equipment.

[0043] Furthermore, the historical operation data in step one includes energy conversion equipment control strategy data and energy conversion equipment output data under all operating conditions.

[0044] The beneficial effects of the present invention are:

[0045] (1) During the operation of the integrated energy system, the nonlinear factors in each link of the actual energy conversion process will affect the operation of the energy conversion equipment. Therefore, the nonlinear factors in each link of the energy input-conversion-output in the energy conversion process of the integrated energy system are deeply regressed and learned using a deep convolutional neural network to construct a power prediction model to ensure the accuracy of the optimization of the subsequent energy conversion equipment control strategy.

[0046] In order to achieve the low-carbon economic operation and optimization scheduling of the integrated energy system, the relevant cost is taken as the optimization target. Considering that the operation basis of the integrated energy system is to ensure that the corresponding load demand is met, the corresponding load demand is set as the corresponding constraint condition, and a specific operation optimization model is constructed. By adjusting the control strategy of the energy conversion equipment, it is ensured that the relevant cost is minimized while meeting the load demand.

[0047] The Grey Wolf Optimization Algorithm is used to optimize the control strategy of the energy conversion equipment in the integrated energy system, and the convergence speed of the Grey Wolf Optimization Algorithm is improved by increasing the convergence factor nonlinear update strategy. At the same time, the elite wolf rollback strategy is further introduced to avoid falling into the local optimum due to too fast convergence. While ensuring the convergence speed, the optimization accuracy is guaranteed, thereby improving the operation optimization effect of the integrated energy system.

[0048] Finally, after determining the optimal control strategy for each energy conversion device, the corresponding energy conversion device is controlled according to the corresponding control strategy for operation control, thereby achieving low-carbon economic operation of the integrated energy system.

[0049] (2) When rolling back the elite wolves, the rollback is performed in multiple groups to increase the amount of data for verifying the optimal solution, so as to improve the accuracy of the optimal energy conversion equipment control strategy sequence determined based on the rollback results. The rollback is performed by replacing the elite wolves, and the replaced elite wolves are randomly selected from the historical positions of the leading wolves in the iterative process to improve the efficiency of the subsequent elite wolves in rolling back and optimizing.

[0050] (3) When setting the nonlinear update strategy of the convergence factor, the quality of the current optimal solution is selected as one of the parameters to dynamically adjust the value of the convergence factor to reflect the impact of the quality of the optimal solution on the search process, further improving the optimization accuracy of the gray wolf optimization algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a schematic diagram of a process of the present invention;

[0052] Figure 2 This is a schematic diagram of the structure of a park cooling, heating and power trigeneration integrated energy system according to an embodiment of the present invention;

[0053] Figure 3 It is a schematic diagram of the structure of an available power prediction model according to an embodiment of the present invention;

[0054] Figure 4 It is a schematic diagram of a convergence process curve of the optimization iteration of the Grey Wolf Optimization Algorithm in period A of an embodiment of the present invention;

[0055] Figure 5It is a schematic diagram of a convergence process curve of the optimization iteration of the Grey Wolf Optimization Algorithm in the B period of an embodiment of the present invention;

[0056] Figure 6 It is a schematic diagram of a convergence process curve of the optimization iteration of the Grey Wolf Optimization Algorithm in the C period of an embodiment of the present invention;

[0057] Figure 7 It is a schematic diagram of a convergence process curve of an optimization iteration of elite wolf rollback in period C of an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0059] Example: A comprehensive energy system operation optimization method based on multiple strategy combination optimization, such as Figure 1 As shown, including:

[0060] Step 1: Collect historical operation data of the integrated energy system to be optimized, and establish an available power prediction model of the integrated energy system based on a deep convolutional neural network;

[0061] Step 2: Obtain the load demand of the integrated energy system to be optimized, and build an operation optimization model based on the purchased energy price and load demand;

[0062] Step 3: Taking the control strategies of various energy conversion devices as the variables to be optimized, the operation optimization model is solved according to the gray wolf optimization algorithm that introduces the convergence factor nonlinear update strategy and the elite wolf rollback strategy and the available power prediction model to obtain the optimal control strategies of various energy conversion devices;

[0063] Step 4: Control the operation of the corresponding energy conversion equipment according to the optimal control strategy of each energy conversion equipment.

[0064] The historical operation data described in step one includes energy conversion equipment control strategy data and energy conversion equipment output data under all operating conditions.

[0065] The comprehensive energy system described in this embodiment is a park cooling, heating and power trigeneration comprehensive energy system, and its structure is as follows: Figure 2 shown.

[0066] Figure 2 Among them, AB is a gas boiler with a maximum heating power of 400kW; CERG is an electric refrigeration equipment with a maximum cooling power of 300kW; WARG is an absorption refrigeration equipment with a maximum cooling power of 300kW; CHP is a cogeneration equipment with a maximum power supply power of 120kW and a maximum heating power of 160kW.

[0067] The energy conversion equipment control strategy data includes the electric power supplied to the electric refrigeration equipment, the natural gas power supplied to the gas boiler, the thermal power supplied to the absorption refrigeration equipment, the natural gas power supplied to the cogeneration equipment, and the park's purchased electric power data.

[0068] The energy conversion equipment output data includes the total power available for supply in the park, the total heating power available in the park, the total cooling power available in the park, etc.

[0069] Because the operating efficiency of each energy conversion equipment will change when it operates under variable conditions, and the climbing rate will also change during the actual loading and unloading process. There are also many irreversible losses in the energy flow transmission process of the energy conversion equipment. These nonlinear factors in each link of the actual energy conversion process will affect the operation of the energy conversion equipment. Therefore, a deep regression model of the control strategy of each energy conversion equipment in the integrated energy system and the overall available power, heat and cooling power of the integrated energy system is established through a deep convolutional neural network, that is, a power supply prediction model, so that the nonlinear factors of the energy input-conversion-output links in the energy conversion process of the integrated energy system are deeply regressed and learned using a deep convolutional neural network.

[0070] Through the controllable power prediction model, the predicted values ​​of the overall available electrical, thermal and cooling power of the park's combined cooling, heating and power energy system can be calculated in real time when different energy conversion equipment control strategies are running.

[0071] Among them, when building a deep convolutional neural network model, the collected historical operation data is first preprocessed, and the corresponding data set is constructed through data cleaning, normalization and feature extraction operations.

[0072] In combination with the energy conversion equipment included in the park's combined cooling, heating and power energy system, and the mutual influence under the operation of the corresponding control strategy, a neural network structure is set. The structure of the convolutional neural network model set in this embodiment is shown in Table 1:

[0073] Table 1 Schematic diagram of deep convolutional neural network model structure

[0074]

[0075] The input layer parameter n in Table 1 represents the number of controllable parameters of the energy conversion equipment control strategy in the park's combined cooling, heating and power energy system to be optimized.

[0076] Initialize the neural network and divide the data set into corresponding training set, validation set and test set. Use the training set to train the constructed deep convolutional neural network model, and use the validation set data to verify the trained model. Adjust the model structure and parameter settings according to the verification results to improve the model performance. Finally, use the test set data to evaluate the performance of the final model, calculate the model's prediction accuracy, generalization ability and other key indicators, and finally obtain the trained available power prediction model.

[0077] When performing subsequent operation optimization, the corresponding energy conversion equipment control strategy is input into the available power prediction model to obtain the overall available electrical, thermal and cooling power of the integrated energy system.

[0078] The available power prediction model in this embodiment is specifically as follows: Figure 3 As shown, Figure 3 The input data of the power prediction model is the energy conversion equipment control strategy The output data is the total power, heating and cooling power of the park. ,in, To supply electrical power to electric refrigeration equipment, To supply heat power to absorption refrigeration equipment, To supply natural gas power to gas boilers, To supply natural gas power to the cogeneration plant, Purchase electric power for the park.

[0079] In order to achieve the low-carbon economic operation and optimal scheduling of the integrated energy system, the optimization goal is to minimize the operating cost and carbon trading cost. The operation basis of the integrated energy system is to meet the load demand of users in the park where it is located. Therefore, the load demand is further used as a constraint condition of the operation optimization model to construct a specific operation optimization model.

[0080] Among them, in this embodiment, by clarifying the load type in the park where the integrated energy system is located and retrieving the corresponding historical load data, the load characteristics and user energy consumption behavior analysis can be carried out, and then the load demand forecast can be carried out according to the historical load data and analysis results through machine learning such as support vector regression and gradient boosting decision tree to obtain the load demand of the integrated energy system to be optimized.

[0081] The load demand of the integrated energy system to be optimized is obtained as described in step 2, and an operation optimization model is constructed based on the purchased energy price and the load demand, including taking the minimization of the operation cost and carbon trading cost of the integrated energy system to be optimized as the optimization goal, and setting the constraint conditions of the operation optimization model based on the load demand to construct the operation optimization model.

[0082] The constraint conditions for setting the operation optimization model based on load demand include:

[0083] A penalty function is introduced into the optimization objective, and load demand is used as the penalty function variable. Constraints that meet the load demand are set for running the optimization model.

[0084] By introducing a penalty function into the optimization objective, the constraints on the overall electricity, heating and cooling loads required by the park users can be realized, thereby intuitively reflecting the impact of load demand on the optimization objective.

[0085] The expression of the objective function of the running optimization model is:

[0086] ;

[0087] ;

[0088] ;

[0089] in, is the total cost, is the operating cost, which is the current price of purchased energy multiplied by the energy consumed under the current energy conversion equipment control strategy. is the price of purchased natural gas per unit power, is the total amount of purchased natural gas, is the price of purchased unit electric power, is the amount of purchased electricity, is the carbon trading cost, is the carbon emission factor per unit of natural gas power purchased, is the carbon emission quota per unit of natural gas power purchased, is the carbon emission factor per unit of purchased electricity, is the carbon emission quota per unit of purchased electricity, To trade the price per kilogram of carbon dioxide, is the penalty function, and , is the total cooling load demand, To correspond to the cooling power available under the energy conversion equipment control strategy, is the total heat load demand, is the available heating power under the control strategy of the energy conversion equipment. is the total electrical load demand, It is the power available under the control strategy of the corresponding energy conversion equipment.

[0090] In the above penalty function, a and b are used to refer to the variables in the penalty function respectively, which have no actual meaning and guide the subsequent penalty function calculation for load demand and available power.

[0091] In this embodiment, the carbon emission factor of the purchased natural gas power unit is Set to 0.234kg / kWh, the carbon emission quota per unit of natural gas power purchased Set to 0.1925kg / kWh, the carbon emission factor per unit of purchased electricity Set to 0.986kg / kWh, the carbon emission quota per unit of purchased electricity Set to 0.728kg / kWh.

[0092] After the operation optimization model is constructed, in order to quickly achieve the optimization goal of minimizing the operating cost and carbon trading cost, the operation optimization model is solved by the Grey Wolf Optimization Algorithm to determine the optimal energy conversion equipment control strategy sequence. The optimal energy conversion equipment control strategy sequence includes the control strategies of each energy conversion equipment.

[0093] The Gray Wolf Optimization Algorithm is a meta-heuristic algorithm that simulates the hunting behavior of gray wolf groups in nature. The algorithm is based on the social hierarchy structure and hunting behavior of gray wolves, and finds the optimal solution by simulating the hierarchy and hunting behavior of gray wolves. In the algorithm, gray wolves are divided into four types: , , and . represents the optimal solution, and represents the second-best solution and the third-best solution, while is responsible for executing decisions and exploring the search space. To lead the wolf, and The elite wolf and the control strategy of the energy conversion equipment are in the position of the gray wolf, and the operation optimization model is solved.

[0094] Among them, in step 3, according to the grey wolf optimization algorithm that introduces the convergence factor nonlinear update strategy and the elite wolf rollback strategy and the available power prediction model, the operation optimization model is solved to obtain the optimal control strategy of each energy conversion device, including:

[0095] Taking the control strategy sequence of energy conversion equipment as the variable to be optimized, the gray wolf population of the gray wolf optimization algorithm is generated, and the population and basic parameters are initialized;

[0096] The available power of each individual gray wolf is determined based on the available power prediction model, and the fitness of each individual gray wolf is calculated by combining the objective function of the running optimization model to divide the leader wolves and elite wolves.

[0097] Based on the convergence factor nonlinear update strategy, the positions of the remaining gray wolves are updated according to the positions of the leader wolf and the elite wolf;

[0098] Determine whether the iteration end condition is met;

[0099] If the iteration end condition is not met, the fitness value of each individual gray wolf is recalculated, the fitness and position of the leader wolf and the elite wolf are updated, and the historical position of the leader wolf is stored in the backup position set until the iteration end condition is met;

[0100] If the iteration end condition is met, the first alternative energy conversion equipment control strategy sequence is output, and a gray wolf individual is randomly selected from the backup position to replace the current elite wolf for rollback optimization. The first alternative energy conversion equipment control strategy sequence and the rollback optimization result are compared to determine the optimal energy conversion equipment control strategy sequence.

[0101] The above basic parameters include population size, maximum number of iterations, search dimension, search range, initial value and termination value of convergence factor, etc. The basic parameters can be set according to requirements. In this embodiment, the population size is 500 and the maximum number of iterations is 200.

[0102] When initializing the population, the initial positions of the corresponding number of gray wolf individuals are randomly generated in the search space according to the set population plan.

[0103] The total cost mentioned in the objective function of the running optimization goal is taken as the fitness, and the fitness is calculated by combining the corresponding total cost formula with the energy conversion equipment control strategy sequence corresponding to the initial position.

[0104] Among them, since the total cost calculation formula needs to take into account the available power of each energy conversion equipment, the available power corresponding to each gray wolf individual is further determined based on the available power prediction model to assist in the calculation of fitness.

[0105] The nonlinear update strategy based on the convergence factor updates the positions of the remaining gray wolf individuals according to the positions of the leader wolf and the elite wolf, including:

[0106] Update the convergence factor based on the current number of iterations and the maximum number of iterations combined with the update formula;

[0107] updating the coefficient vector based on the updated convergence factor;

[0108] According to the positions of the leader wolf and the elite wolf, the positions of the remaining gray wolf individuals are updated in combination with the updated coefficient vector.

[0109] The expression of the update formula is:

[0110] ;

[0111] in, is the updated convergence factor, is the initial value of the convergence factor, is the current iteration number, is the maximum number of iterations, is the fitness value of the leader wolf in the current iteration, is the fitness value of the leader wolf in the previous iteration, is the preset adjustment factor.

[0112] In the Gray Wolf Optimization Algorithm, the convergence factor is used to balance the global search capability and local search capability of the algorithm. As the iteration proceeds, the value of the convergence factor will gradually decrease, thereby guiding the algorithm to gradually focus on the optimal solution from a wide search space. In the early stage, a larger convergence factor means that the Gray Wolf has a larger exploration range in the search space, thereby assisting the algorithm to jump out of the local optimal solution and conduct a global search, but it may sacrifice some convergence speed. In the middle and late stages, as the iteration proceeds, the convergence factor gradually decreases, and the Gray Wolf's search range also shrinks, which can assist the algorithm to conduct a detailed search near the optimal solution and improve the convergence speed.

[0113] In order to optimize the convergence speed, a convergence factor nonlinear update strategy is adopted to update the convergence factor of each iteration.

[0114] Among them, in the update formula of the convergence factor, The first part ensures that the convergence factor decreases with the increase of the number of iterations, while the exponential function part also involves the quality of the solution in the update process of the convergence factor, thereby associating the convergence speed with the quality of the solution of each iteration, and further optimizing the convergence factor for the balance between global search capability and local search capability.

[0115] in, is a preset adjustment factor that can affect the dynamic adjustment range of the convergence factor. In this embodiment, , the specific value can be obtained through specific experiments, and the specific or Experiments are then conducted and adjustments are made based on the experimental results to determine the specific values ​​of the adjustment factors.

[0116] In the position update algorithm of the gray wolf optimization algorithm, the position update is mainly associated with the coefficient vector, and the coefficient vector is associated with the convergence coefficient. Therefore, after updating the convergence factor, the coefficient vector is updated according to the convergence factor, thereby updating the corresponding position update formula, and then according to the positions of the leader wolf and the elite wolf, the positions of the remaining gray wolf individuals are updated based on the updated position update formula.

[0117] The change of the convergence factor presents an exponential trend. In order to avoid the problem of falling into the local optimal solution due to the convergence of the convergence factor too quickly in the later stage, after completing the optimization iteration and outputting the optimal solution, the elite wolf corresponding to the optimal solution is replaced. The gray wolf optimization algorithm is re-executed based on the replaced elite wolf and leader wolf to realize the rollback of the elite wolf. The quality of the solution of this optimization iteration is verified according to the rollback result.

[0118] If the fitness of the optimal solution after re-execution is greater than or equal to the fitness of the optimal solution of this optimization iteration, it proves that it has not fallen into a local optimal solution and the optimal solution of this optimization iteration can be used.

[0119] If the fitness of the optimal solution after re-execution is less than the fitness of the optimal solution of this optimization iteration, it proves that this optimization iteration has not achieved the optimal effect and may fall into a local optimal solution. The optimal solution after re-execution is used as the optimal solution for the final output.

[0120] In order to further improve the accuracy of the optimal solution, when rolling back the elite wolves, two replacement groups including different gray wolf individuals are used for rolling back.

[0121] Specifically, the method randomly selects a gray wolf individual from the backup position set, replaces the current elite wolf to perform rollback optimization, compares the first candidate energy conversion device control strategy sequence with the rollback optimization result, and determines the optimal energy conversion device control strategy sequence, including:

[0122] A first replacement group and a second replacement group including different gray wolf individuals are randomly selected from the backup position set to replace the current elite wolves respectively;

[0123] Iterate again according to the current leading wolf and the elite wolf replaced by the first replacement group until the iteration end condition is met, and obtain the first rollback energy conversion device control strategy sequence;

[0124] At the same time, reiterate according to the current leading wolf and the elite wolf replaced by the second replacement group until the iteration end condition is met, and obtain the second rollback energy conversion device control strategy sequence;

[0125] The fitness values ​​of the first candidate energy conversion device control strategy sequence, the first rollback energy conversion device control strategy sequence and the second rollback energy conversion device control strategy sequence are compared, and the energy conversion device control strategy sequence with the smaller fitness value is selected as the optimal energy conversion device control strategy sequence.

[0126] Taking the optimization of the energy conversion equipment control strategy of the park's combined cooling, heating and power energy system in periods A, B and C as an example, optimization is performed using the comprehensive energy system operation optimization method provided in this embodiment.

[0127] The control strategy of energy conversion equipment in the park's combined cooling, heating and power energy system during period A is optimized. The effect comparison before and after optimization is shown in Table 2:

[0128] Table 2 Comparison of the effects of the park's combined cooling, heating and power energy system before and after optimization during period A

[0129]

[0130] In the optimization process of the gray wolf optimization algorithm, the reliability of the optimal solution output by its optimization iteration is verified by the elite wolf rollback strategy. Therefore, the optimal solution output by it is used as the optimal energy conversion equipment control strategy sequence. The convergence process curve of its optimization iteration is as follows: Figure 4 shown.

[0131] According to Table 2 and Figure 4 It can be seen that in period A, the price of energy purchased from the park and the total electricity load required by users in the park are , heat load , cooling load Under the constraint of , the sum of the current one-hour operation cost of the park's comprehensive energy system and the carbon trading cost gradually decreased from 171.9 yuan to 113.0 yuan with the optimization iteration of Gray Wolf. The total energy supply of the optimized park, including the total power supply, heat and cooling power of the park, achieved a balance between supply and demand with the overall power, heat and cooling load required by the park users, and made it possible to meet the overall power load required by the park users. , heat load , cooling load Under the constraint of , the sum of the current integrated energy system operating cost and the carbon trading cost is minimized, that is, the optimization objective function is minimized.

[0132] Before and after optimization, the sum of 1-hour operating cost and carbon trading cost decreased by 34.3%, 1-hour operating cost decreased by 32.5%, and 1-hour carbon trading cost decreased by 52.2%.

[0133] The control strategy of energy conversion equipment in the park's combined cooling, heating and power energy system during period B is optimized. The effect comparison before and after optimization is shown in Table 3:

[0134] Table 3 Comparison of the effects of the park's combined cooling, heating and power energy system before and after optimization during period B

[0135]

[0136] In the optimization process of the gray wolf optimization algorithm, the reliability of the optimal solution output by its optimization iteration is verified by the elite wolf rollback strategy. Therefore, the optimal solution output by it is used as the optimal energy conversion equipment control strategy sequence. The convergence process curve of its optimization iteration is as follows: Figure 5 shown.

[0137] Depend on Figure 5 From Table 3, we can see that in period B, the price of energy purchased from the park and the total electricity load required by users in the park , heat load , cooling load Under the constraint of , the sum of the current one-hour operation cost of the park's comprehensive energy system and the carbon trading cost gradually decreased from 267.5 yuan to 197.3 yuan with the optimization iteration of Gray Wolf. The total energy supply of the optimized park, including the total power supply, heat and cooling power of the park, achieved a balance between supply and demand with the overall power, heat and cooling load required by the park users, and made it possible to meet the overall power load required by the park users. , heat load , cooling load Under the constraint of , the sum of the current integrated energy system operating cost and the carbon trading cost is minimized, that is, the optimization objective function is minimized.

[0138] Before and after optimization, the sum of 1-hour operating cost and carbon trading cost decreased by 26.2%, 1-hour operating cost decreased by 26.3%, and 1-hour carbon trading cost decreased by 26.0%.

[0139] The control strategy of energy conversion equipment in the park's combined cooling, heating and power energy system during period C is optimized. The effect comparison before and after optimization is shown in Table 4:

[0140] Table 4 Comparison of the effects of the park's CCHP integrated energy system before and after optimization during period C

[0141]

[0142] During the optimization process of the Grey Wolf optimization algorithm, it was found that the fitness of the optimal solution of the rollback result was lower than the fitness of the optimal solution output by its optimization iteration. Therefore, it was judged that there might be a problem of falling into a local optimal solution. The one with lower fitness was selected from the rollback result and the optimal solution output by the optimization iteration as the optimal energy conversion equipment control strategy sequence.

[0143] Among them, the convergence process curve of the optimization iteration and the convergence process curve of the optimization iteration of the selected elite wolf rollback are as follows: Figure 6 and Figure 7 shown.

[0144] Depend on Figure 7 From Table 4, we can see that in period C, the price of energy purchased from the park and the total electricity load required by users in the park are , heat load , cooling load Under the constraint of , the sum of the current one-hour operation cost of the park's comprehensive energy system and the carbon trading cost gradually decreased from 384.4 yuan to 272.3 yuan with the optimization iteration of Gray Wolf. The total energy supply of the optimized park, including the total power supply, heat and cooling power of the park, achieved a supply and demand balance with the overall power, heat and cooling load required by the park users, and made it possible to meet the overall power load required by the park users. , heat load , cooling load Under the constraint of , the sum of the current integrated energy system operating cost and the carbon trading cost is minimized, that is, the optimization objective function is minimized.

[0145] Before and after optimization, the sum of 1-hour operating cost and carbon trading cost decreased by 29.2%, 1-hour operating cost decreased by 28.7%, and 1-hour carbon trading cost decreased by 34.9%.

[0146] Combined with the comparison results before and after optimization of the above-mentioned periods A, B and C, the effectiveness of the method shown in this embodiment in optimizing the low-carbon economic operation of the existing park's trigeneration integrated energy system is directly verified, and it is possible to achieve accurate low-carbon economic operation optimization scheduling of the actual existing park integrated energy system.

[0147] The above-described embodiment is only a preferred solution of the present invention and does not limit the present invention in any form. There are other variations and modifications without exceeding the technical solution described in the claims.

Claims

1. An integrated energy system operation optimization method based on multiple strategy combination optimization, characterized in that: include: Step 1: Collect historical operation data of the integrated energy system to be optimized, and establish an available power prediction model of the integrated energy system based on a deep convolutional neural network; Step 2: Obtain the load demand of the integrated energy system to be optimized, and build an operation optimization model based on the purchased energy price and load demand; Step 3: Taking the control strategies of various energy conversion devices as the variables to be optimized, the operation optimization model is solved according to the gray wolf optimization algorithm that introduces the convergence factor nonlinear update strategy and the elite wolf rollback strategy and the available power prediction model to obtain the optimal control strategies of various energy conversion devices; Step 4: Control the operation of the corresponding energy conversion equipment according to the optimal control strategy of each energy conversion equipment; Step three includes: Taking the control strategy sequence of energy conversion equipment as the variable to be optimized, the gray wolf population of the gray wolf optimization algorithm is generated, and the population and basic parameters are initialized; The available power of each individual gray wolf is determined based on the available power prediction model, and the fitness of each individual gray wolf is calculated by combining the objective function of the running optimization model to divide the leader wolves and elite wolves. Based on the convergence factor nonlinear update strategy, the positions of the remaining gray wolves are updated according to the positions of the leader wolf and the elite wolf; The nonlinear update strategy based on the convergence factor updates the positions of the remaining gray wolf individuals according to the positions of the leader wolf and the elite wolf, including: According to the current number of iterations and the maximum number of iterations, the convergence factor is updated according to the update formula; updating the coefficient vector based on the updated convergence factor; According to the positions of the leader wolf and the elite wolf, the positions of the remaining gray wolf individuals are updated in combination with the updated coefficient vector; The expression of the update formula is: ； in, is the updated convergence factor, is the initial value of the convergence factor, is the current iteration number, is the maximum number of iterations, is the fitness value of the leader wolf in the current iteration, is the fitness value of the leader wolf in the previous iteration, is the preset adjustment factor.

2. The integrated energy system operation optimization method based on multiple strategy combination optimization according to claim 1 is characterized in that: Step 3, according to the grey wolf optimization algorithm that introduces the convergence factor nonlinear update strategy and the elite wolf rollback strategy and the available power prediction model, solves the operation optimization model to obtain the optimal control strategy for each energy conversion device, and also includes: Determine whether the iteration end condition is met; If the iteration end condition is not met, the fitness value of each individual gray wolf is recalculated, the fitness and position of the leader wolf and the elite wolf are updated, and the historical position of the leader wolf is stored in the backup position set until the iteration end condition is met; If the iteration end condition is met, the first alternative energy conversion equipment control strategy sequence is output, and a gray wolf individual is randomly selected from the backup position to replace the current elite wolf for rollback optimization. The first alternative energy conversion equipment control strategy sequence and the rollback optimization result are compared to determine the optimal energy conversion equipment control strategy sequence.

3. The integrated energy system operation optimization method based on multiple strategy combination optimization according to claim 2 is characterized in that: The method of randomly selecting a gray wolf individual from the backup position set to replace the current elite wolf for rollback optimization, comparing the first candidate energy conversion device control strategy sequence with the rollback optimization result, and determining the optimal energy conversion device control strategy sequence includes: Randomly select a first replacement group and a second replacement group including different gray wolf individuals from the backup position set to replace the current elite wolves respectively; Iterate again according to the current leading wolf and the elite wolf replaced by the first replacement group until the iteration end condition is met, and obtain the first rollback energy conversion device control strategy sequence; At the same time, reiterate according to the current leading wolf and the elite wolf replaced by the second replacement group until the iteration end condition is met, and obtain the second rollback energy conversion device control strategy sequence; The fitness values ​​of the first candidate energy conversion device control strategy sequence, the first rollback energy conversion device control strategy sequence and the second rollback energy conversion device control strategy sequence are compared, and the energy conversion device control strategy sequence with the smaller fitness value is selected as the optimal energy conversion device control strategy sequence.

4. The integrated energy system operation optimization method based on multiple strategy combination optimization according to claim 1 is characterized in that: The load demand of the integrated energy system to be optimized is obtained as described in step 2, and an operation optimization model is constructed based on the purchased energy price and the load demand, including taking the minimization of the operation cost and carbon trading cost of the integrated energy system to be optimized as the optimization goal, and setting the constraint conditions of the operation optimization model based on the load demand to construct the operation optimization model.

5. The integrated energy system operation optimization method based on multiple strategy combination optimization according to claim 4 is characterized in that: The constraint conditions for setting the operation optimization model based on load demand include: A penalty function is introduced into the optimization objective, and load demand is used as the penalty function variable. Constraints that meet the load demand are set for running the optimization model.

6. The integrated energy system operation optimization method based on multiple strategy combination optimization according to claim 5 is characterized in that: The expression of the objective function of the running optimization model is: ； ； ； in, is the total cost, For running costs, is the price of purchased natural gas power per unit, is the total amount of purchased natural gas, is the price of purchased unit electric power, is the amount of purchased electricity, is the carbon trading cost, is the carbon emission factor per unit of natural gas power purchased, is the carbon emission quota per unit of natural gas power purchased, is the carbon emission factor per unit of purchased electricity, is the carbon emission quota per unit of purchased electricity, To trade the price per kilogram of carbon dioxide, is the penalty function, and , is the total cooling load demand, To correspond to the cooling power available under the energy conversion equipment control strategy, is the total heat load demand, is the available heating power under the control strategy of the energy conversion equipment. is the total electricity load demand, It is the power available under the control strategy of the corresponding energy conversion equipment.

7. The integrated energy system operation optimization method based on multiple strategy combination optimization according to claim 1 is characterized in that: The historical operation data described in step one includes energy conversion equipment control strategy data and energy conversion equipment output data under all operating conditions.

Images