Park distributed equipment group intelligent linkage method and system based on load characteristics

Abstract

The invention provides a park distributed equipment group intelligent linkage method and system based on load characteristics, and relates to the technical field of energy scheduling, and the method comprises the steps: obtaining real-time load data of a plurality of energy consumption units in a park, and inputting a deep neural network model to generate a load characteristic curve; a multi-objective optimization model is constructed based on the load characteristic curve, and an optimal operation scheme is solved by using an improved particle swarm algorithm; and finally, controlling each distributed energy device according to the optimal operation scheme, and triggering a dynamic optimization correction mechanism in case of abnormality or sudden load change. Through load characteristic analysis and intelligent optimization control, the system operation cost can be minimized, the energy utilization efficiency can be maximized, and coordinated operation and load dynamic balance of the distributed energy equipment group in the park can be realized.

Description

Technical Field

[0001] The present invention relates to energy dispatching technology, and in particular to a method and system for intelligent linkage of a park distributed equipment group based on load characteristics. Background Art

[0002] Park energy management is an important means to improve energy efficiency and reduce operating costs. With the rapid development of distributed energy technology, more and more parks are beginning to use distributed energy equipment such as photovoltaics, energy storage, and electric refrigeration to meet their own energy needs. How to effectively coordinate and control these devices to make them work together and dynamically adjust the operation strategy according to load changes has become a key challenge facing park energy management.

[0003] Traditional park energy management methods usually use rule-based control strategies or simple optimization algorithms, which are difficult to fully consider the impact of complex factors such as load characteristics, equipment characteristics, and environmental factors, resulting in low energy efficiency and high operating costs. For example, some methods only control the start and stop of equipment according to a pre-set time plan, which cannot adapt to the dynamic changes of load; other methods use single-objective optimization strategies, only consider economic benefits or environmental benefits, and it is difficult to achieve multi-objective collaborative optimization.

[0004] The prior art has the following defects and deficiencies:

[0005] 1. Lack of in-depth exploration of load characteristics. Traditional energy management methods are usually based on simple load forecasting models, which are difficult to accurately capture complex characteristics such as load peak and valley distribution, change trends, and correlations between different energy-consuming units, making it difficult for control strategies to adapt to dynamic changes in loads.

[0006] Second, the coordinated control capabilities of distributed energy equipment are insufficient. Traditional control methods usually control each device independently, lack consideration of the interaction between devices, and are difficult to achieve coordinated and optimized operation of distributed energy equipment, thus restricting the improvement of overall energy utilization efficiency.

[0007] 3. Lack of adaptability to dynamic environmental factors. In actual operation, factors such as ambient temperature and light intensity will have a significant impact on the operation of loads and distributed energy equipment. Traditional energy management methods usually lack the ability to effectively perceive and adapt to these dynamic environmental factors, making it difficult to achieve optimal control effects. Summary of the invention

[0008] The embodiments of the present invention provide a method and system for intelligent linkage of a distributed device group in a park based on load characteristics, which can solve the problems in the prior art.

[0009] A first aspect of an embodiment of the present invention provides a method for intelligent linkage of a distributed device group in a park based on load characteristics, comprising:

[0010] Acquire real-time load data of multiple energy-consuming units in the park, wherein the real-time load data includes power load data, heating load data, and cooling load data; input the real-time load data into a pre-trained deep neural network model, wherein the deep neural network model is trained based on historical load data and environmental parameter data to generate load characteristic curves of the multiple energy-consuming units, wherein the load characteristic curves include peak-valley distribution characteristics, load change trend characteristics, and load correlation characteristics;

[0011] Based on the load characteristic curve, a multi-objective optimization model of the distributed energy equipment group in the park is constructed, wherein the optimization objectives of the multi-objective optimization model include minimizing the system operation cost and maximizing the energy utilization efficiency, and the constraints of the multi-objective optimization model include equipment operation parameter constraints, energy balance constraints and environmental constraints; the multi-objective optimization model is solved by using an improved particle swarm algorithm to obtain the optimal operation plan of the distributed energy equipment group in the park, wherein the optimal operation plan includes the power generation power of photovoltaic power generation equipment, the charging and discharging power of energy storage equipment and the cooling power of electric refrigeration equipment, wherein the improved particle swarm algorithm adopts an adaptive local search strategy and an elite solution dynamic maintenance mechanism on the basis of the particle swarm algorithm;

[0012] According to the optimal operation plan, control instructions are sent to each distributed energy device through the park energy management system, and the control instructions include the equipment start and stop status, operating power and operating parameters; the operating status of each distributed energy device is monitored in real time, and when abnormal equipment operation or sudden load changes are detected, a dynamic optimization correction mechanism is triggered. The dynamic optimization correction mechanism dynamically updates the optimal operation plan based on a rolling time domain to ensure the coordinated operation of the park's distributed energy equipment group and the dynamic balance of the load.

[0013] Inputting the real-time load data into a pre-trained deep neural network model, wherein the deep neural network model is trained based on historical load data and environmental parameter data, and generating load characteristic curves of the plurality of energy consumption units includes:

[0014] Calculate the peak-to-valley difference index of the maximum load, the minimum load and the average load within 24 hours corresponding to the real-time load data according to the deep neural network model, and obtain the change rate of the peak-to-valley difference in the time dimension to obtain the intra-day peak-to-valley change rate; obtain the peak duration index by weighted calculation of the duration of each peak interval, and the weight coefficient of each duration interval in the peak duration index is determined based on the load characteristics;

[0015] Perform multi-scale wavelet decomposition on the real-time load data to obtain multi-layer detail components and approximate components; calculate the trend change rate and acceleration characteristics based on the approximate components; perform energy density analysis on the detail components, calculate the energy value of detail components of each scale and their proportion in the total energy, and obtain the energy proportion, where the total energy includes the sum of the detail component energy and the approximate component energy;

[0016] The dynamic correlation coefficient between the loads of different energy-consuming units is calculated by using a sliding window method, and the dynamic correlation coefficient is calculated based on the covariance and standard deviation of the load data in the window; the correlation fluctuation index is calculated based on the dynamic correlation coefficient, and the correlation fluctuation index represents the stability of the correlation coefficient within the observation period;

[0017] The peak-to-valley difference index, intraday peak-to-valley change rate, peak duration index, trend change rate, acceleration characteristics, energy proportion, dynamic correlation coefficient and correlation fluctuation index are combined to construct a load characteristic vector; the feature importance is calculated based on the variance of each load characteristic vector, and the feature importance is determined by the ratio of the variance of the load characteristic vector to the sum of the variances of all load characteristic vectors, thereby generating load characteristic curves of the multiple energy-consuming units.

[0018] Based on the load characteristic curve, a multi-objective optimization model of the park distributed energy equipment group is constructed. The optimization objectives of the multi-objective optimization model include minimizing the system operation cost and maximizing the energy utilization efficiency, including:

[0019] The equipment fuel cost is calculated based on the output power, equipment efficiency and unit fuel price of each equipment in the park distributed energy equipment group at each moment, and the equipment fuel cost, the pre-acquired equipment operation and maintenance cost and the pre-acquired equipment start-up and shutdown cost are added together to obtain the single equipment operation cost. The operation costs of all equipment are accumulated in the time dimension and the equipment dimension to obtain the total system operation cost;

[0020] Obtain the output energy and input energy of each device, calculate the ratio of the output energy to the input energy to obtain the system comprehensive efficiency; obtain the available energy and total energy in various energy forms, calculate the weighted sum of the ratios of the available energy and total energy in various energy forms based on a preset weight coefficient to obtain an energy quality coefficient; multiply the system comprehensive efficiency by the energy quality coefficient to obtain the system energy utilization efficiency;

[0021] A multi-objective optimization model is constructed, with the total operating cost of the system as the first optimization objective and the inverse of the system energy efficiency as the second optimization objective; the constraints of the multi-objective optimization model include equipment operating parameter constraints, energy balance constraints and environmental constraints.

[0022] The improved particle swarm algorithm is used to solve the multi-objective optimization model, and the optimal operation plan of the park's distributed energy equipment group is obtained, including:

[0023] Constructing particle position vectors and velocity vectors that characterize the operating status of the device in different time periods, limiting the velocity vector within a preset velocity range, and initializing the particle swarm using a time-based adaptive inertia weight, wherein the adaptive inertia weight decreases nonlinearly with the increase in the number of iterations, to obtain an initialized particle swarm;

[0024] For each particle in the initialized particle swarm, a dual-objective fitness function is constructed based on system operation cost and energy utilization efficiency for evaluation, a dominance level of each particle is determined by non-dominated sorting, and a crowding distance between particles is calculated as a diversity index; based on the dominance level and the crowding distance, a dynamic learning factor is used to update the particle position vector and velocity vector;

[0025] For the updated particle position vector, a mixed Gaussian perturbation containing standard deviations of different scales is constructed for local search. The weights of perturbations of different scales are adjusted by adaptive mixing coefficients. A solution space density function is constructed based on the congestion level of the current solution. The local search direction is dynamically adjusted using the solution space density function to obtain a local optimal solution.

[0026] Establishing an elite solution evaluation system including dominance level, distance diversity and target space distribution for the local optimization solution, and calculating the comprehensive score of the solution based on the elite solution evaluation system; maintaining the elite solution set by using a dynamic capacity strategy that is adaptively adjusted according to the population size of the current particle swarm;

[0027] The convergence index is calculated based on the continuous change of the particle position of the elite solution set; the density analysis and uniformity evaluation are performed on the converged non-dominated solution set, and finally an optimal Pareto solution set with uniform distribution is output, and the optimal Pareto solution set corresponds to the optimal operation plan of the distributed energy equipment group.

[0028] The method further comprises:

[0029] A dynamic search step is constructed based on the number of iterations, and the dynamic search step decreases as the number of iterations increases in a power function; a mixed Gaussian distribution with different scale standard deviations is constructed, and the mixed Gaussian distributions with different scale standard deviations are combined using a mixing coefficient that is adaptively adjusted according to the iteration process to obtain a mixed Gaussian perturbation vector; the distance distribution between the solutions in the current solution set is calculated, and a density function that reflects the degree of congestion in the solution space is established, and the dynamic search step, the mixed Gaussian perturbation vector and the density function are multiplied to construct an enhanced local search operator;

[0030] For the solution set optimized by the enhanced local search operator, the dominance level is obtained by calculating the number of solutions dominated by other solutions, the distance diversity index is calculated based on the minimum Euclidean distance between the solutions, and the target space distribution index is obtained by calculating the deviation of each solution in the target space relative to the average value; the dominance level, distance diversity index and target space distribution index are weightedly combined by weight coefficients to construct a comprehensive scoring function for elite solutions;

[0031] According to the population size of the current particle swarm, the upper limit of the dynamic capacity of the elite solution set is linearly determined within the preset minimum capacity and maximum capacity range; the solutions with the best dominance level and whose number does not exceed the upper limit of the dynamic capacity are selected to form the updated elite solution set;

[0032] For the updated elite solution set, the position change of the solution between adjacent iterations is calculated, and a population convergence index is constructed based on the statistical characteristics of the position change; non-dominated solutions that meet the population convergence index and are evenly distributed are selected from the updated elite solution set, and finally an optimal Pareto solution set with even distribution is formed.

[0033] According to the optimal operation plan, the control instructions sent to each distributed energy device through the park energy management system include:

[0034] Constructing a device start-stop instruction set, a power control instruction set and a parameter adjustment instruction set, wherein the device start-stop instruction set includes a start-stop status identifier of each device, the power control instruction set includes a target power value of each device, and the parameter adjustment instruction set includes an operating parameter vector of each device;

[0035] Based on the energy efficiency coefficient, response speed coefficient and flexibility coefficient of each device, the energy efficiency coefficient, response speed coefficient and flexibility coefficient are weighted and combined by a weight coefficient to obtain an importance index of each device, and the control instruction execution priority of each device is determined according to the importance index;

[0036] Establishing a device startup timing matrix, wherein the matrix elements in the device startup timing matrix represent the startup response delay between the corresponding two devices, dividing the instruction execution time window based on the device startup timing matrix, and determining the key execution time nodes in the time window in sequence according to the constraint relationship of the device startup timing matrix;

[0037] According to the importance index of each device and the startup timing matrix, the start and stop state identifiers in the device start and stop instruction set are decomposed in order of priority to obtain a device start and stop instruction sequence, and the target power value in the power control instruction set is decomposed in order of priority to obtain a power control instruction sequence;

[0038] The device start and stop instruction sequence and the power control instruction sequence are combined according to the divided time windows and key execution time nodes to form a phased control instruction containing an execution sequence, wherein the phased control instruction contains the start and stop status identification, target power value and operating parameter vector of each device at each key time node.

[0039] A second aspect of an embodiment of the present invention provides a park distributed device group intelligent linkage system based on load characteristics, including:

[0040] The first unit is used to obtain real-time load data of multiple energy-consuming units in the park, wherein the real-time load data includes power load data, heating load data, and cooling load data; the real-time load data is input into a pre-trained deep neural network model, wherein the deep neural network model is trained based on historical load data and environmental parameter data to generate load characteristic curves of the multiple energy-consuming units, wherein the load characteristic curves include peak-valley distribution characteristics, load change trend characteristics, and load correlation characteristics;

[0041] The second unit is used to construct a multi-objective optimization model of the distributed energy equipment group in the park based on the load characteristic curve, the optimization objectives of the multi-objective optimization model include minimizing the system operation cost and maximizing the energy utilization efficiency, and the constraints of the multi-objective optimization model include equipment operation parameter constraints, energy balance constraints and environmental constraints; using an improved particle swarm algorithm to solve the multi-objective optimization model to obtain the optimal operation plan of the distributed energy equipment group in the park, the optimal operation plan includes the power generation power of photovoltaic power generation equipment, the charging and discharging power of energy storage equipment and the cooling power of electric refrigeration equipment, wherein the improved particle swarm algorithm adopts an adaptive local search strategy and an elite solution dynamic maintenance mechanism on the basis of the particle swarm algorithm;

[0042] The third unit is used to send control instructions to each distributed energy device through the park energy management system according to the optimal operation plan, and the control instructions include the equipment start and stop status, operating power and operating parameters; monitor the operating status of each distributed energy device in real time, and when abnormal equipment operation or sudden load change is detected, trigger the dynamic optimization correction mechanism, which dynamically updates the optimal operation plan based on the rolling time domain to ensure the coordinated operation of the park's distributed energy equipment group and the dynamic balance of the load.

[0043] A third aspect of the embodiments of the present invention

[0044] An electronic device is provided, comprising:

[0045] processor;

[0046] a memory for storing processor-executable instructions;

[0047] The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0048] According to a fourth aspect of the embodiments of the present invention,

[0049] A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0050] The beneficial effects of this application are as follows:

[0051] 1. Reduce system operating costs and improve energy efficiency: Through the multi-objective optimization model, the system operating cost minimization and energy efficiency maximization are comprehensively considered, and the improved particle swarm algorithm is used to solve the optimal operation plan, which can effectively reduce the operating cost of the park energy system and improve energy efficiency.

[0052] 2. Improve load forecasting accuracy and equipment operation stability: Based on the deep neural network model, historical load data and environmental parameter data are trained to accurately predict the load characteristic curve of the energy-consuming unit, including peak and valley distribution characteristics, load change trend characteristics and load correlation characteristics, thereby improving load forecasting accuracy and ensuring the stability of equipment operation.

[0053] 3. Achieve dynamic load balance and enhance system adaptability: The dynamic optimization and correction mechanism can monitor the equipment operating status and load changes in real time, and dynamically update the optimal operation plan based on the rolling time domain when the equipment operates abnormally or the load changes suddenly, to ensure the coordinated operation of the distributed energy equipment group in the park and the dynamic balance of the load, and enhance the system's adaptability to emergencies. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A schematic diagram of a process of a method for intelligent linkage of a distributed device group in a park based on load characteristics according to an embodiment of the present invention;

[0055] Figure 2 The present invention is a schematic diagram of the structure of a park distributed equipment group intelligent linkage system based on load characteristics according to an embodiment of the present invention. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0058] Figure 1 FIG. 1 is a flow chart of a method for intelligent linkage of a distributed equipment group in a park based on load characteristics according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0059] S101. Obtain real-time load data of multiple energy-consuming units in the park, wherein the real-time load data includes power load data, heating load data, and cooling load data; input the real-time load data into a pre-trained deep neural network model, wherein the deep neural network model is trained based on historical load data and environmental parameter data to generate load characteristic curves of the multiple energy-consuming units, wherein the load characteristic curves include peak-valley distribution characteristics, load change trend characteristics, and load correlation characteristics;

[0060] S102. Based on the load characteristic curve, a multi-objective optimization model of the distributed energy equipment group in the park is constructed, wherein the optimization objectives of the multi-objective optimization model include minimizing the system operation cost and maximizing the energy utilization efficiency, and the constraints of the multi-objective optimization model include equipment operation parameter constraints, energy balance constraints, and environmental constraints; the multi-objective optimization model is solved by using an improved particle swarm algorithm to obtain an optimal operation plan of the distributed energy equipment group in the park, wherein the optimal operation plan includes the power generation power of the photovoltaic power generation equipment, the charging and discharging power of the energy storage equipment, and the cooling power of the electric refrigeration equipment, wherein the improved particle swarm algorithm adopts an adaptive local search strategy and an elite solution dynamic maintenance mechanism on the basis of the particle swarm algorithm;

[0061] S103. According to the optimal operation plan, control instructions are sent to each distributed energy device through the park energy management system, and the control instructions include the equipment start and stop status, operating power and operating parameters; the operating status of each distributed energy device is monitored in real time, and when abnormal equipment operation or sudden load change is detected, a dynamic optimization correction mechanism is triggered. The dynamic optimization correction mechanism dynamically updates the optimal operation plan based on a rolling time domain to ensure the coordinated operation of the park's distributed energy equipment group and the dynamic balance of the load.

[0062] In an optional implementation, the real-time load data is input into a pre-trained deep neural network model, the deep neural network model is trained based on historical load data and environmental parameter data, and generating load characteristic curves of the multiple energy consumption units includes:

[0063] Calculate the peak-to-valley difference index of the maximum load, the minimum load and the average load within 24 hours corresponding to the real-time load data according to the deep neural network model, and obtain the change rate of the peak-to-valley difference in the time dimension to obtain the intra-day peak-to-valley change rate; obtain the peak duration index by weighted calculation of the duration of each peak interval, and the weight coefficient of each duration interval in the peak duration index is determined based on the load characteristics;

[0064] Perform multi-scale wavelet decomposition on the real-time load data to obtain multi-layer detail components and approximate components; calculate the trend change rate and acceleration characteristics based on the approximate components; perform energy density analysis on the detail components, calculate the energy value of detail components of each scale and their proportion in the total energy, and obtain the energy proportion, where the total energy includes the sum of the detail component energy and the approximate component energy;

[0065] The dynamic correlation coefficient between the loads of different energy-consuming units is calculated by using a sliding window method, and the dynamic correlation coefficient is calculated based on the covariance and standard deviation of the load data in the window; the correlation fluctuation index is calculated based on the dynamic correlation coefficient, and the correlation fluctuation index represents the stability of the correlation coefficient within the observation period;

[0066] The peak-to-valley difference index, intraday peak-to-valley change rate, peak duration index, trend change rate, acceleration characteristics, energy proportion, dynamic correlation coefficient and correlation fluctuation index are combined to construct a load characteristic vector; the feature importance is calculated based on the variance of each load characteristic vector, and the feature importance is determined by the ratio of the variance of the load characteristic vector to the sum of the variances of all load characteristic vectors, thereby generating load characteristic curves of the multiple energy-consuming units.

[0067] A method for generating a load characteristic curve of an energy-consuming unit driven by real-time load data comprises the following steps:

[0068] First, obtain historical load data and environmental parameter data. Historical load data includes the power consumption data of each energy-consuming unit in the past period of time. For example, hourly power consumption data for the past year can be collected. Environmental parameter data includes temperature, humidity, weather conditions, etc., which can be obtained through weather stations or sensors. For example, the average temperature, maximum temperature, minimum temperature and humidity data for each day in the past year can be collected.

[0069] Then, the deep neural network model is trained using the acquired historical load data and environmental parameter data. The deep neural network model can adopt a structure such as a convolutional neural network, a recurrent neural network or a variant thereof. During the training process, the historical load data is used as output, and the environmental parameter data and the corresponding time information are used as input. For example, the historical load data and environmental parameter data of the past year can be arranged in chronological order and input into the deep neural network model for training, so that the model can learn the relationship between the load data and the environmental parameters.

[0070] Next, the real-time load data is input into the trained deep neural network model to generate load characteristic curves for multiple energy-consuming units. Real-time load data refers to the current electricity consumption data, which can be collected in real time through devices such as smart meters. For example, the current electricity consumption data and the corresponding environmental parameter data are input into the trained deep neural network model, and the model will output the predicted load data for each energy-consuming unit in the next 24 hours, thereby generating a load characteristic curve.

[0071] Based on the output results of the deep neural network model, the maximum load, minimum load and average load within 24 hours corresponding to the real-time load data are calculated. For example, the maximum and minimum values ​​can be found from the predicted 24-hour load data, and the average value of the 24-hour load data can be calculated. Then, the peak-to-valley difference is calculated, that is, the difference between the maximum and minimum values. In order to obtain the rate of change of the peak-to-valley difference in the time dimension, the change in the peak-to-valley difference between adjacent moments can be calculated to obtain the intraday peak-to-valley change rate.

[0072] In order to calculate the peak duration index, the peak interval must first be determined. For example, the interval that is a certain percentage higher than the average load can be defined as the peak interval. Then, the duration of each peak interval is calculated separately. Finally, the duration of each peak interval is weighted and summed according to the pre-set weight coefficient to obtain the peak duration index. The setting of the weight coefficient can be adjusted according to the load characteristics. For example, a larger weight is given to the peak interval with a longer duration.

[0073] Multi-scale wavelet decomposition is performed on real-time load data. Wavelet decomposition can decompose the signal into components of different frequencies, thereby extracting the characteristics of the signal. Multi-scale wavelet decomposition can decompose the signal into multiple layers of detail components and an approximate component. The detail component represents the high-frequency part of the signal, and the approximate component represents the low-frequency part of the signal.

[0074] Based on the approximate components obtained by wavelet decomposition, the trend change rate and acceleration characteristics are calculated. The trend change rate can be obtained by calculating the difference of the approximate components at adjacent moments. The acceleration characteristics can be obtained by calculating the difference of the trend change rates at adjacent moments.

[0075] Perform energy density analysis on the detail components obtained by wavelet decomposition. Calculate the energy value of detail components at each scale and the proportion of detail component energy at each scale to the total energy to obtain the energy proportion. The total energy includes the sum of all detail component energies and approximate component energies.

[0076] The sliding window method is used to calculate the dynamic correlation coefficient between the loads of different energy-consuming units. The sliding window method refers to sliding a fixed-size window on the time series to calculate the correlation coefficient of the data in the window. The dynamic correlation coefficient is calculated based on the covariance and standard deviation of the load data in the window.

[0077] The correlation fluctuation index is calculated based on the dynamic correlation coefficient. The correlation fluctuation index characterizes the stability of the correlation coefficient during the observation period. For example, the fluctuation degree of the dynamic correlation coefficient can be measured by calculating the standard deviation of the dynamic correlation coefficient.

[0078] Finally, the peak-to-valley difference index, intraday peak-to-valley change rate, peak duration index, trend change rate, acceleration characteristics, energy proportion, dynamic correlation coefficient and correlation fluctuation index are combined to construct the load characteristic vector. The feature importance is calculated based on the variance of each load characteristic vector. The feature importance is determined by the ratio of the variance of the load characteristic vector to the sum of the variances of all load characteristic vectors. Finally, the load characteristic curves of multiple energy-consuming units are generated.

[0079] The beneficial effects of this method can be summarized in the following three aspects:

[0080] 1. High accuracy: Using technologies such as deep neural network models and multi-scale wavelet decomposition, the characteristics of load data can be extracted more accurately, thereby more accurately predicting future loads and generating more accurate load characteristic curves. For example, the accuracy of load forecasting can be improved by learning the complex relationship between historical data and environmental parameters through deep learning models.

[0081] 2. Comprehensiveness: This method comprehensively considers the multi-dimensional characteristics of load data, including peak-to-valley difference, change rate, duration, trend change, energy distribution, and correlation between different energy-consuming units, and can more comprehensively describe the load characteristics of energy-consuming units. For example, through wavelet decomposition, the load change characteristics on different time scales can be analyzed, and through correlation analysis, the mutual influence between different energy-consuming units can be revealed.

[0082] 3. Good practicability: The load characteristic curve generated by this method can provide an important reference for power system planning, operation and management. For example, the future load peak can be predicted based on the load characteristic curve, thereby optimizing the power dispatch strategy and improving energy efficiency. At the same time, this method can also be used to identify abnormal power consumption behavior and improve the safety of the power system.

[0083] In an optional implementation, based on the load characteristic curve, a multi-objective optimization model of a park distributed energy equipment group is constructed, wherein the optimization objectives of the multi-objective optimization model include minimizing system operation costs and maximizing energy utilization efficiency, including:

[0084] The equipment fuel cost is calculated based on the output power, equipment efficiency and unit fuel price of each equipment in the park distributed energy equipment group at each moment, and the equipment fuel cost, the pre-acquired equipment operation and maintenance cost and the pre-acquired equipment start-up and shutdown cost are added together to obtain the single equipment operation cost. The operation costs of all equipment are accumulated in the time dimension and the equipment dimension to obtain the total system operation cost;

[0085] Obtain the output energy and input energy of each device, calculate the ratio of the output energy to the input energy to obtain the system comprehensive efficiency; obtain the available energy and total energy in various energy forms, calculate the weighted sum of the ratios of the available energy and total energy in various energy forms based on a preset weight coefficient to obtain an energy quality coefficient; multiply the system comprehensive efficiency by the energy quality coefficient to obtain the system energy utilization efficiency;

[0086] A multi-objective optimization model is constructed, with the total operating cost of the system as the first optimization objective and the inverse of the system energy efficiency as the second optimization objective; the constraints of the multi-objective optimization model include equipment operating parameter constraints, energy balance constraints and environmental constraints.

[0087] A multi-objective optimization method for constructing a distributed energy equipment group in a park based on load characteristic curves aims to minimize system operating costs and maximize energy efficiency. The specific steps of this method are as follows:

[0088] First, obtain the park load characteristic curve, which describes the park's power, cooling and heating energy demand at different times. For example, hourly power demand data and cooling and heating demand data for the next week can be obtained based on historical data, prediction models or actual measurements.

[0089] Next, determine the composition of the park's distributed energy equipment group, such as gas internal combustion engines, gas boilers, photovoltaic power generation systems, energy storage systems, etc. And obtain the parameters of each device, including rated power, efficiency characteristic curve, fuel type, unit fuel price, operation and maintenance cost, start-up and shutdown cost, etc. For example, the efficiency characteristic curve of a gas internal combustion engine describes the efficiency value at different output powers; the unit fuel price of a gas boiler is the price per cubic meter of natural gas; the operation and maintenance cost of a photovoltaic power generation system includes daily maintenance and cleaning costs.

[0090] Then, the distributed energy equipment group in the park is dispatched and optimized to meet the load demand of the park. Dispatching optimization needs to consider the constraints of equipment operation parameters, such as the output power range of the equipment, the number of start and stop times, etc.; energy balance constraints, that is, at each moment, the sum of the output energy of all equipment is equal to the load demand of the park; and environmental constraints, such as emission limits. The result of dispatch optimization is the output power of each device at each moment. For example, at 9 am, the output power of the photovoltaic power generation system is 100kW, and the output power of the gas internal combustion engine is 50kW, which together meet the 150kW power demand of the park.

[0091] According to the scheduling optimization results, calculate the total operating cost of the system. First, calculate the equipment fuel cost based on the output power, equipment efficiency and unit fuel price of each device at each time. Then, add the equipment fuel cost, the pre-acquired equipment operation and maintenance cost, and the pre-acquired equipment start and stop cost to get the single equipment operating cost. Finally, add the operating costs of all equipment in the time dimension and the equipment dimension to get the total operating cost of the system. For example, the output power of a gas internal combustion engine at a certain moment is 50kW, the efficiency is 40%, and the unit fuel price is 3 yuan / cubic meter. The fuel cost at that moment is 50kW / 40%*3 yuan / cubic meter*operating time.

[0092] At the same time, calculate the energy utilization efficiency of the system. First, obtain the output energy and input energy of each device, and calculate the ratio of output energy to input energy to obtain the overall efficiency of the system. Then, obtain the available energy and total energy of various energy forms, for example, the ratio of available energy of natural gas to its total energy, the ratio of available energy of electricity to its total energy, etc. Calculate the weighted sum of the ratios of available energy to total energy in various energy forms based on the preset weight coefficient to obtain the energy quality coefficient. Finally, multiply the overall efficiency of the system by the energy quality coefficient to obtain the energy utilization efficiency of the system. For example, if the overall efficiency of the system is 80% and the energy quality coefficient is 90%, the energy utilization efficiency of the system is 72%.

[0093] Finally, a multi-objective optimization model is constructed, with the total system operating cost as the first optimization objective and the opposite of the system energy efficiency as the second optimization objective. The model is solved using a multi-objective optimization algorithm, such as a genetic algorithm or a particle swarm algorithm. Finally, a set of Pareto optimal solutions are obtained, which represent different trade-offs between system operating costs and energy efficiency. Decision makers can choose the appropriate solution based on actual needs.

[0094] The beneficial effects of this method can be summarized in the following three aspects:

[0095] 1. Reduce the energy cost of the park: By optimizing the operation strategy of the distributed energy equipment group, the energy consumption of the park can be effectively reduced, thereby reducing energy costs.

[0096] 2. Improve energy utilization efficiency: This method takes into account the comprehensive efficiency and energy quality of energy, and can effectively improve the comprehensive utilization efficiency of energy.

[0097] 3. Reduce environmental pollution: By optimizing the operation of the energy system, pollutant emissions can be reduced, thereby protecting the environment.

[0098] In an optional implementation, the improved particle swarm algorithm is used to solve the multi-objective optimization model to obtain the optimal operation plan of the park distributed energy equipment group, including:

[0099] Constructing particle position vectors and velocity vectors that characterize the operating status of the device in different time periods, limiting the velocity vector within a preset velocity range, and initializing the particle swarm using a time-based adaptive inertia weight, wherein the adaptive inertia weight decreases nonlinearly with the increase in the number of iterations, to obtain an initialized particle swarm;

[0100] For each particle in the initialized particle swarm, a dual-objective fitness function is constructed based on system operation cost and energy utilization efficiency for evaluation, a dominance level of each particle is determined by non-dominated sorting, and a crowding distance between particles is calculated as a diversity index; based on the dominance level and the crowding distance, a dynamic learning factor is used to update the particle position vector and velocity vector;

[0101] For the updated particle position vector, a mixed Gaussian perturbation containing standard deviations of different scales is constructed for local search. The weights of perturbations of different scales are adjusted by adaptive mixing coefficients. A solution space density function is constructed based on the congestion level of the current solution. The local search direction is dynamically adjusted using the solution space density function to obtain a local optimal solution.

[0102] Establishing an elite solution evaluation system including dominance level, distance diversity and target space distribution for the local optimization solution, and calculating the comprehensive score of the solution based on the elite solution evaluation system; maintaining the elite solution set by using a dynamic capacity strategy that is adaptively adjusted according to the population size of the current particle swarm;

[0103] The convergence index is calculated based on the continuous change of the particle position of the elite solution set; the density analysis and uniformity evaluation are performed on the converged non-dominated solution set, and finally an optimal Pareto solution set with uniform distribution is output, and the optimal Pareto solution set corresponds to the optimal operation plan of the distributed energy equipment group.

[0104] The distributed energy equipment group optimization operation method aims to reduce system operation costs and improve energy efficiency. This method uses an improved particle swarm algorithm to search for the optimal operation plan, and improves the algorithm's optimization performance and solution quality through a series of strategies.

[0105] First, the particle swarm is initialized. In order to characterize the operating status of the device in different time periods, the particle position vector and velocity vector are constructed. The velocity vector is limited to a preset range. For example, the rate of change of the device output power is limited to a certain range. The particle swarm is initialized using a time-based adaptive inertia weight. The inertia weight decreases nonlinearly with the increase in the number of iterations. For example, an exponential decreasing function is used to enable the algorithm to have a strong global search capability in the early stage, and focus on local optimization in the later stage. Assuming there are 10 devices, each time period is 1 hour, and the operating cycle is 24 hours, the position vector of each particle contains 240 elements, each element represents the output power of the corresponding device in the corresponding time period.

[0106] Then, the fitness of each particle in the initialized particle swarm is evaluated. A dual-objective fitness function is constructed based on the system operating cost and energy efficiency. For example, the operating cost objective can consider fuel cost, maintenance cost, etc., and the energy efficiency objective can consider energy conversion efficiency, energy loss, etc. The dominance level of each particle is determined by non-dominated sorting, and the crowding distance between particles is calculated as a diversity indicator. The crowding distance reflects the distribution density of solutions around the particle. Assuming there are two objective functions, each particle is non-dominatedly sorted according to its objective function value to obtain its dominance level.

[0107] Next, update the particle position vector and velocity vector. Use a dynamic learning factor to update the particle position vector and velocity vector according to the particle's dominance level and crowding distance. The learning factor can be adjusted according to the particle's dominance level. For example, a particle with a high dominance level has a larger learning factor, which enables it to better maintain its own advantages. Assuming that a particle's dominance level is 1 and its crowding distance is 0.5, the learning factor of the particle is calculated according to the preset dynamic learning factor formula.

[0108] Perform local search for the updated particle position vector. Construct a mixed Gaussian perturbation containing standard deviations of different scales for local search. For example, set three standard deviations of different scales to represent small-scale, medium-scale, and large-scale perturbations. Adjust the weights of perturbations of different scales through adaptive mixing coefficients. The mixing coefficient can be adjusted according to the degree of crowding of the current solution. For example, in crowded areas, increase the weight of larger-scale perturbations to escape the local optimum. Construct a solution space density function based on the degree of crowding of the current solution, and use this function to dynamically adjust the local search direction to obtain a local optimal solution. For example, in sparse areas, prioritize searching areas with higher density.

[0109] An elite solution evaluation system is established for local optimization solutions. The system includes dominance level, distance diversity, and target space distribution. The comprehensive score of the solution is calculated based on the evaluation system. For example, the comprehensive score of the solution can be calculated based on the weights of dominance level, distance diversity, and target space distribution. A dynamic capacity strategy is used to maintain the elite solution set, and the dynamic capacity is adaptively adjusted according to the population size of the current particle swarm. For example, the larger the population size, the larger the capacity of the elite solution set.

[0110] The convergence index is calculated based on the continuous change of the particle position in the elite solution set. For example, the variance of the particle position in the elite solution set is calculated. If the variance is less than the preset threshold, the algorithm is considered to have converged. The non-dominated solution set after convergence is subjected to density analysis and uniformity evaluation, and finally an optimal Pareto solution set with uniform distribution is output. For example, the distance between solutions in the non-dominated solution set is calculated. If the distance distribution is uniform, the solution set is considered to be uniformly distributed. The optimal Pareto solution set corresponds to the optimal operation plan of the distributed energy device group. For example, based on the particle position vector in the optimal Pareto solution set, the output power of each device in each time period can be determined.

[0111] The beneficial effects of this method can be summarized in the following three aspects:

[0112] 1. Improve optimization performance: Through strategies such as adaptive inertia weight, dynamic learning factor and mixed Gaussian perturbation, the algorithm's global search ability and local optimization ability are improved, and better solutions can be found faster.

[0113] 2. Improve the quality of the solution set: Through methods such as non-dominated sorting, crowding distance and elite solution evaluation system, the diversity and uniformity of the Pareto solution set are guaranteed, providing more diverse options.

[0114] 3. Enhanced practicality: This method can flexibly adjust parameters according to actual conditions, such as the number of equipment, operation cycle, objective function, etc., and is suitable for the optimized operation of different types of distributed energy systems.

[0115] In an optional implementation, the method further includes:

[0116] A dynamic search step is constructed based on the number of iterations, and the dynamic search step decreases as the number of iterations increases in a power function; a mixed Gaussian distribution with different scale standard deviations is constructed, and the mixed Gaussian distributions with different scale standard deviations are combined using a mixing coefficient that is adaptively adjusted according to the iteration process to obtain a mixed Gaussian perturbation vector; the distance distribution between the solutions in the current solution set is calculated, and a density function that reflects the degree of congestion in the solution space is established, and the dynamic search step, the mixed Gaussian perturbation vector and the density function are multiplied to construct an enhanced local search operator;

[0117] For the solution set optimized by the enhanced local search operator, the dominance level is obtained by calculating the number of solutions dominated by other solutions, the distance diversity index is calculated based on the minimum Euclidean distance between the solutions, and the target space distribution index is obtained by calculating the deviation of each solution in the target space relative to the average value; the dominance level, distance diversity index and target space distribution index are weightedly combined by weight coefficients to construct a comprehensive scoring function for elite solutions;

[0118] According to the population size of the current particle swarm, the upper limit of the dynamic capacity of the elite solution set is linearly determined within the preset minimum capacity and maximum capacity range; the solutions with the best dominance level and whose number does not exceed the upper limit of the dynamic capacity are selected to form the updated elite solution set;

[0119] For the updated elite solution set, the position change of the solution between adjacent iterations is calculated, and a population convergence index is constructed based on the statistical characteristics of the position change; non-dominated solutions that meet the population convergence index and are evenly distributed are selected from the updated elite solution set, and finally an optimal Pareto solution set with even distribution is formed.

[0120] A multi-objective optimization method based on the elite solution set strategy is used to find multiple optimal solutions in a complex search space. This method constructs an enhanced local search operator through dynamic search step size, mixed Gaussian perturbation and density function, and combines dominance level, distance diversity and target space distribution indicators to screen elite solutions, and finally obtains a uniformly distributed Pareto optimal solution set.

[0121] First, an initial population of solutions is initialized, each of which represents a potential solution to the multi-objective optimization problem. For example, in the case of designing an aircraft, each solution can represent a different design of the aircraft, including parameters such as wing shape, fuselage length, and engine thrust.

[0122] Next, the iterative search process is performed. In each iteration, the dynamic search step is first calculated according to the current iteration number. The dynamic search step decreases in a power function as the number of iterations increases. For example, the step can be set to the initial value of 0.9 times the number of iterations. This ensures that a larger range of exploration is performed in the early stage of the search, and a more refined local search is performed in the later stage.

[0123] Then, a mixed Gaussian distribution is constructed, which is composed of multiple Gaussian distributions with different standard deviations. For example, three Gaussian distributions with standard deviations of 0.1, 0.5, and 1 can be used. The mixing coefficients are used to control the proportion of Gaussian distributions with different standard deviations in the mixed Gaussian distribution, and these mixing coefficients are adaptively adjusted during the iteration process. For example, in the early stage of the search, a Gaussian distribution with a larger standard deviation can be given a higher weight for a more extensive exploration; and in the later stage of the search, a Gaussian distribution with a smaller standard deviation can be given a higher weight for a more refined local search. Multiply the dynamic search step size by the mixed Gaussian perturbation vector to obtain a perturbation vector.

[0124] At the same time, the distance distribution between the solutions in the current solution set is calculated, and a density function is established to reflect the degree of crowding in the solution space. For example, the Gaussian kernel density estimation method can be used to construct the density function. The perturbation vector is multiplied by the density function to construct an enhanced local search operator. This operator can guide the search process to avoid crowded areas and explore sparser areas.

[0125] Apply the enhanced local search operator to each solution in the current population to generate new solutions. Evaluate the newly generated solutions and the original solutions to calculate the objective function value for each solution. For example, in the case of aircraft design, the objective function can be the flight speed and fuel efficiency of the aircraft.

[0126] Then, the number of solutions dominated by other solutions is calculated to obtain the dominance level. The lower the dominance level, the better the performance of the solution. At the same time, the distance diversity index is calculated based on the minimum Euclidean distance between the solutions. The distance diversity index reflects the diversity of the solution set. In addition, the target space distribution index is obtained by calculating the deviation of each solution in the target space relative to the average value. The target space distribution index reflects the distribution of the solution set in the target space. The dominance level, distance diversity index and target space distribution index are weighted and combined by weight coefficients to construct an elite solution comprehensive scoring function. For example, the weight of the dominance level can be set to 0.5, the weight of the distance diversity index can be set to 0.3, and the weight of the target space distribution index can be set to 0.2.

[0127] According to the current population size, the upper limit of the dynamic capacity of the elite solution set is linearly determined within the preset minimum capacity and maximum capacity range. For example, assuming the minimum capacity is 10, the maximum capacity is 100, and the current population size is 50, the upper limit of the dynamic capacity is 60. The solutions with the best dominance level and whose number does not exceed the upper limit of the dynamic capacity are selected to form the updated elite solution set.

[0128] For the updated elite solution set, the position change of the solution between adjacent iterations is calculated, and the population convergence index is constructed based on the statistical characteristics of the position change. For example, the mean value or variance of the position change can be calculated as the population convergence index. Non-dominated solutions that meet the population convergence index and are evenly distributed are selected from the updated elite solution set, and finally the optimal Pareto solution set with even distribution is formed.

[0129] The beneficial effects of this method can be summarized in the following three aspects:

[0130] 1. Improved search efficiency: Through dynamic search step size and mixed Gaussian perturbations, the search space can be explored more efficiently and the optimal solution can be found faster.

[0131] 2. Enhanced population diversity: Through density function and distance diversity indicators, the diversity of the population can be effectively maintained and the search process can be prevented from falling into local optimality.

[0132] 3. Obtaining a uniformly distributed Pareto optimal solution set: Through the elite solution set strategy and the target space distribution index, a uniformly distributed Pareto optimal solution set can be obtained, providing decision makers with more choices.

[0133] In an optional implementation, according to the optimal operation plan, sending a control instruction to each distributed energy device through the park energy management system includes:

[0134] Constructing a device start-stop instruction set, a power control instruction set and a parameter adjustment instruction set, wherein the device start-stop instruction set includes a start-stop status identifier of each device, the power control instruction set includes a target power value of each device, and the parameter adjustment instruction set includes an operating parameter vector of each device;

[0135] Based on the energy efficiency coefficient, response speed coefficient and flexibility coefficient of each device, the energy efficiency coefficient, response speed coefficient and flexibility coefficient are weighted and combined by a weight coefficient to obtain an importance index of each device, and the control instruction execution priority of each device is determined according to the importance index;

[0136] Establishing a device startup timing matrix, wherein the matrix elements in the device startup timing matrix represent the startup response delay between the corresponding two devices, dividing the instruction execution time window based on the device startup timing matrix, and determining the key execution time nodes in the time window in sequence according to the constraint relationship of the device startup timing matrix;

[0137] According to the importance index of each device and the startup timing matrix, the start and stop state identifiers in the device start and stop instruction set are decomposed in order of priority to obtain a device start and stop instruction sequence, and the target power value in the power control instruction set is decomposed in order of priority to obtain a power control instruction sequence;

[0138] The device start and stop instruction sequence and the power control instruction sequence are combined according to the divided time windows and key execution time nodes to form a phased control instruction containing an execution sequence, wherein the phased control instruction contains the start and stop status identification, target power value and operating parameter vector of each device at each key time node.

[0139] The park energy management system achieves optimal operation of energy by sending control instructions to distributed energy devices. The specific implementation of the control method is described in detail below:

[0140] First, the system collects real-time operating data of all distributed energy devices in the park, including device type, current power, energy efficiency coefficient, response speed, adjustable parameter range, etc. For example, the current power of gas turbine A is 50kW, the energy efficiency coefficient is 0.4, the response speed is 10kW / min, and the adjustable parameters include gas flow and speed, etc.; the current power of photovoltaic array B is 20kW, the energy efficiency coefficient is 0.9, the response speed is 5kW / min, and the adjustable parameters include inverter output voltage and current, etc.

[0141] Next, construct the device control instruction set. The instruction set consists of three parts: the device start and stop instruction set, which records the target start and stop status of each device, such as "start" or "stop"; the power control instruction set, which records the target power value of each device, such as the target power of gas turbine A is 60kW, and the target power of photovoltaic array B is 30kW; the parameter adjustment instruction set, which records the target operating parameter vector of each device, such as adjusting the gas flow of gas turbine A to 80% and the speed to 6000rpm.

[0142] Then, evaluate the importance of each device. According to the energy efficiency coefficient, response speed coefficient and flexibility coefficient of the equipment, a weighted combination calculation is performed to obtain the importance index. For example, the energy efficiency coefficient weight is set to 0.5, the response speed coefficient weight is set to 0.3, and the flexibility coefficient weight is set to 0.2. The importance index of gas turbine A is 0.5*0.4+0.3*10+0.2*1=3.2; the importance index of photovoltaic array B is 0.5*0.9+0.3*5+0.2*1=2.15. According to the importance index, the control instruction execution priority of the device is determined. The larger the value of the importance index, the higher the priority of the device. For example, the priority of gas turbine A is higher than that of photovoltaic array B.

[0143] Establish an equipment startup timing matrix. This matrix describes the startup dependencies and time delays between devices. For example, after gas turbine A is started, it takes 5 minutes to start waste heat boiler C, so the element value from A to C in the matrix is ​​5. Based on this matrix, divide the instruction execution time window, and determine the key execution time nodes in the time window in sequence according to the constraint relationship of the equipment startup timing matrix. For example, if the time window is set to 10 minutes, 0 minutes, 5 minutes, and 10 minutes can be set as key execution time nodes.

[0144] According to the importance index of the equipment and the startup sequence matrix, the equipment start-stop instruction set and power control instruction set are decomposed into instruction sequences. According to the priority order, the equipment start-stop status identification and target power value are assigned to each key execution time node. For example, at 0 minutes, gas turbine A starts and the power is set to 55kW; at 5 minutes, waste heat boiler C starts; at 10 minutes, the power of gas turbine A is adjusted to 60kW, and the power of photovoltaic array B is adjusted to 30kW.

[0145] Finally, the equipment start and stop instruction sequence, power control instruction sequence and parameter adjustment instruction set are combined according to the divided time windows and key execution time nodes to form a phased control instruction containing the execution sequence. For example, at 0 minutes, the control instruction of gas turbine A is "start, power 55kW, gas flow 75%, speed 5500rpm"; at 5 minutes, the control instruction of waste heat boiler C is "start"; at 10 minutes, the control instruction of gas turbine A is "run, power 60kW, gas flow 80%, speed 6000rpm", and the control instruction of photovoltaic array B is "run, power 30kW, inverter output voltage 500V, current 60A".

[0146] The beneficial effects of this method are reflected in three aspects:

[0147] First, improve energy efficiency. By considering the energy efficiency coefficient and operating status of the equipment, optimize energy distribution and utilization, minimize energy waste, and improve overall energy efficiency.

[0148] Second, enhance the system response speed. By considering the response speed coefficient of the equipment, prioritize the equipment with fast response speed, quickly respond to changes in energy demand, and ensure the stability and reliability of energy supply.

[0149] Third, improve the flexibility of system operation. By considering the flexibility coefficient of the equipment and the adjustable parameter range, the equipment operation status can be flexibly adjusted to adapt to different energy demand scenarios and improve the overall flexibility of the system operation.

[0150] Figure 2FIG. 1 is a schematic diagram of a structure of a park distributed device group intelligent linkage system based on load characteristics according to an embodiment of the present invention. Figure 2 As shown, the system comprises:

[0151] The first unit is used to obtain real-time load data of multiple energy-consuming units in the park, wherein the real-time load data includes power load data, heating load data, and cooling load data; the real-time load data is input into a pre-trained deep neural network model, wherein the deep neural network model is trained based on historical load data and environmental parameter data to generate load characteristic curves of the multiple energy-consuming units, wherein the load characteristic curves include peak-valley distribution characteristics, load change trend characteristics, and load correlation characteristics;

[0152] The second unit is used to construct a multi-objective optimization model of the distributed energy equipment group in the park based on the load characteristic curve, the optimization objectives of the multi-objective optimization model include minimizing the system operation cost and maximizing the energy utilization efficiency, and the constraints of the multi-objective optimization model include equipment operation parameter constraints, energy balance constraints and environmental constraints; using an improved particle swarm algorithm to solve the multi-objective optimization model to obtain the optimal operation plan of the distributed energy equipment group in the park, the optimal operation plan includes the power generation power of photovoltaic power generation equipment, the charging and discharging power of energy storage equipment and the cooling power of electric refrigeration equipment, wherein the improved particle swarm algorithm adopts an adaptive local search strategy and an elite solution dynamic maintenance mechanism on the basis of the particle swarm algorithm;

[0153] The third unit is used to send control instructions to each distributed energy device through the park energy management system according to the optimal operation plan, and the control instructions include the equipment start and stop status, operating power and operating parameters; monitor the operating status of each distributed energy device in real time, and when abnormal equipment operation or sudden load change is detected, trigger the dynamic optimization correction mechanism, which dynamically updates the optimal operation plan based on the rolling time domain to ensure the coordinated operation of the park's distributed energy equipment group and the dynamic balance of the load.

[0154] According to a third aspect of the embodiments of the present invention,

[0155] An electronic device is provided, comprising:

[0156] processor;

[0157] a memory for storing processor-executable instructions;

[0158] The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0159] According to a fourth aspect of the embodiments of the present invention,

[0160] A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0161] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0162] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent linkage of distributed equipment groups in a park based on load characteristics, characterized in that: include: Acquire real-time load data of multiple energy-consuming units in the park, wherein the real-time load data includes power load data, heating load data, and cooling load data; Inputting the real-time load data into a pre-trained deep neural network model, wherein the deep neural network model is trained based on historical load data and environmental parameter data to generate load characteristic curves of the multiple energy consumption units, wherein the load characteristic curves include peak-valley distribution characteristics, load change trend characteristics, and load correlation characteristics; Based on the load characteristic curve, a multi-objective optimization model of the distributed energy equipment group in the park is constructed, wherein the optimization objectives of the multi-objective optimization model include minimizing the system operation cost and maximizing the energy utilization efficiency, and the constraints of the multi-objective optimization model include equipment operation parameter constraints, energy balance constraints and environmental constraints; the multi-objective optimization model is solved by using an improved particle swarm algorithm to obtain the optimal operation plan of the distributed energy equipment group in the park, wherein the optimal operation plan includes the power generation power of photovoltaic power generation equipment, the charging and discharging power of energy storage equipment and the cooling power of electric refrigeration equipment, wherein the improved particle swarm algorithm adopts an adaptive local search strategy and an elite solution dynamic maintenance mechanism on the basis of the particle swarm algorithm; According to the optimal operation plan, a control instruction is sent to each distributed energy device through the park energy management system, and the control instruction includes the start and stop status of the device, the operating power and the operating parameters; The operating status of each distributed energy device is monitored in real time. When abnormal equipment operation or sudden load change is detected, a dynamic optimization correction mechanism is triggered. The dynamic optimization correction mechanism dynamically updates the optimal operation plan based on a rolling time domain to ensure the coordinated operation of the distributed energy equipment group in the park and the dynamic balance of the load.

2. The method according to claim 1, characterized in that Inputting the real-time load data into a pre-trained deep neural network model, wherein the deep neural network model is trained based on historical load data and environmental parameter data, and generating load characteristic curves of the plurality of energy consumption units includes: Calculate the peak-to-valley difference index of the maximum load, the minimum load and the average load within 24 hours corresponding to the real-time load data according to the deep neural network model, and obtain the change rate of the peak-to-valley difference in the time dimension to obtain the intra-day peak-to-valley change rate; obtain the peak duration index by weighted calculation of the duration of each peak interval, and the weight coefficient of each duration interval in the peak duration index is determined based on the load characteristics; Perform multi-scale wavelet decomposition on the real-time load data to obtain multi-layer detail components and approximate components; calculate the trend change rate and acceleration characteristics based on the approximate components; perform energy density analysis on the detail components, calculate the energy value of detail components of each scale and their proportion in the total energy, and obtain the energy proportion, where the total energy includes the sum of the detail component energy and the approximate component energy; The dynamic correlation coefficient between the loads of different energy-consuming units is calculated by using a sliding window method, and the dynamic correlation coefficient is calculated based on the covariance and standard deviation of the load data in the window; the correlation fluctuation index is calculated based on the dynamic correlation coefficient, and the correlation fluctuation index represents the stability of the correlation coefficient within the observation period; The peak-to-valley difference index, intraday peak-to-valley change rate, peak duration index, trend change rate, acceleration characteristics, energy proportion, dynamic correlation coefficient and correlation fluctuation index are combined to construct a load characteristic vector; the feature importance is calculated based on the variance of each load characteristic vector, and the feature importance is determined by the ratio of the variance of the load characteristic vector to the sum of the variances of all load characteristic vectors, thereby generating load characteristic curves of the multiple energy-consuming units.

3. The method according to claim 1, characterized in that Based on the load characteristic curve, a multi-objective optimization model of the park distributed energy equipment group is constructed. The optimization objectives of the multi-objective optimization model include minimizing the system operation cost and maximizing the energy utilization efficiency, including: The equipment fuel cost is calculated based on the output power, equipment efficiency and unit fuel price of each equipment in the park distributed energy equipment group at each moment, and the equipment fuel cost, the pre-acquired equipment operation and maintenance cost and the pre-acquired equipment start-up and shutdown cost are added together to obtain the single equipment operation cost. The operation costs of all equipment are accumulated in the time dimension and the equipment dimension to obtain the total system operation cost; Obtain the output energy and input energy of each device, calculate the ratio of the output energy to the input energy to obtain the system comprehensive efficiency; obtain the available energy and total energy in various energy forms, calculate the weighted sum of the ratios of the available energy and total energy in various energy forms based on a preset weight coefficient to obtain an energy quality coefficient; multiply the system comprehensive efficiency by the energy quality coefficient to obtain the system energy utilization efficiency; A multi-objective optimization model is constructed, with the total operating cost of the system as the first optimization objective and the inverse of the system energy efficiency as the second optimization objective; the constraints of the multi-objective optimization model include equipment operating parameter constraints, energy balance constraints and environmental constraints.

4. The method according to claim 3, characterized in that The improved particle swarm algorithm is used to solve the multi-objective optimization model, and the optimal operation plan of the park's distributed energy equipment group is obtained, including: Constructing particle position vectors and velocity vectors that characterize the operating status of the device in different time periods, limiting the velocity vector within a preset velocity range, and initializing the particle swarm using a time-based adaptive inertia weight, wherein the adaptive inertia weight decreases nonlinearly with the increase in the number of iterations, to obtain an initialized particle swarm; For each particle in the initialized particle swarm, a dual-objective fitness function is constructed based on system operation cost and energy utilization efficiency for evaluation, a dominance level of each particle is determined by non-dominated sorting, and a crowding distance between particles is calculated as a diversity index; based on the dominance level and the crowding distance, a dynamic learning factor is used to update the particle position vector and velocity vector; For the updated particle position vector, a mixed Gaussian perturbation containing standard deviations of different scales is constructed for local search. The weights of perturbations of different scales are adjusted by adaptive mixing coefficients. A solution space density function is constructed based on the congestion level of the current solution. The local search direction is dynamically adjusted using the solution space density function to obtain a local optimal solution. Establishing an elite solution evaluation system including dominance level, distance diversity and target space distribution for the local optimization solution, and calculating the comprehensive score of the solution based on the elite solution evaluation system; maintaining the elite solution set by using a dynamic capacity strategy that is adaptively adjusted according to the population size of the current particle swarm; The convergence index is calculated based on the continuous change of the particle position of the elite solution set; the density analysis and uniformity evaluation are performed on the converged non-dominated solution set, and finally an optimal Pareto solution set with uniform distribution is output, and the optimal Pareto solution set corresponds to the optimal operation plan of the distributed energy equipment group.

5. The method according to claim 4, characterized in that The method further comprises: A dynamic search step is constructed based on the number of iterations, and the dynamic search step decreases as the number of iterations increases in a power function; a mixed Gaussian distribution with different scale standard deviations is constructed, and the mixed Gaussian distributions with different scale standard deviations are combined using a mixing coefficient that is adaptively adjusted according to the iteration process to obtain a mixed Gaussian perturbation vector; the distance distribution between the solutions in the current solution set is calculated, and a density function that reflects the degree of congestion in the solution space is established, and the dynamic search step, the mixed Gaussian perturbation vector and the density function are multiplied to construct an enhanced local search operator; For the solution set optimized by the enhanced local search operator, the dominance level is obtained by calculating the number of solutions dominated by other solutions, the distance diversity index is calculated based on the minimum Euclidean distance between the solutions, and the target space distribution index is obtained by calculating the deviation of each solution in the target space relative to the average value; the dominance level, distance diversity index and target space distribution index are weightedly combined by weight coefficients to construct a comprehensive scoring function for elite solutions; According to the population size of the current particle swarm, the upper limit of the dynamic capacity of the elite solution set is linearly determined within the preset minimum capacity and maximum capacity range; the solutions with the best dominance level and whose number does not exceed the upper limit of the dynamic capacity are selected to form the updated elite solution set; For the updated elite solution set, the position change of the solution between adjacent iterations is calculated, and a population convergence index is constructed based on the statistical characteristics of the position change; non-dominated solutions that meet the population convergence index and are evenly distributed are selected from the updated elite solution set, and finally an optimal Pareto solution set with even distribution is formed.

6. The method according to claim 1, characterized in that According to the optimal operation plan, the control instructions sent to each distributed energy device through the park energy management system include: Constructing a device start-stop instruction set, a power control instruction set and a parameter adjustment instruction set, wherein the device start-stop instruction set includes a start-stop status identifier of each device, the power control instruction set includes a target power value of each device, and the parameter adjustment instruction set includes an operating parameter vector of each device; Based on the energy efficiency coefficient, response speed coefficient and flexibility coefficient of each device, the energy efficiency coefficient, response speed coefficient and flexibility coefficient are weighted and combined by a weight coefficient to obtain an importance index of each device, and the control instruction execution priority of each device is determined according to the importance index; Establishing a device startup timing matrix, wherein the matrix elements in the device startup timing matrix represent the startup response delay between the corresponding two devices, dividing the instruction execution time window based on the device startup timing matrix, and determining the key execution time nodes in the time window in sequence according to the constraint relationship of the device startup timing matrix; According to the importance index of each device and the startup timing matrix, the start and stop state identifiers in the device start and stop instruction set are decomposed in order of priority to obtain a device start and stop instruction sequence, and the target power value in the power control instruction set is decomposed in order of priority to obtain a power control instruction sequence; The device start and stop instruction sequence and the power control instruction sequence are combined according to the divided time windows and key execution time nodes to form a phased control instruction containing an execution sequence, wherein the phased control instruction contains the start and stop status identification, target power value and operating parameter vector of each device at each key time node.

7. A park distributed equipment group intelligent linkage system based on load characteristics, used to implement the method as described in any one of claims 1 to 6, characterized in that: include: The first unit is used to obtain real-time load data of multiple energy-consuming units in the park, wherein the real-time load data includes power load data, heating load data and cooling load data; Inputting the real-time load data into a pre-trained deep neural network model, wherein the deep neural network model is trained based on historical load data and environmental parameter data to generate load characteristic curves of the multiple energy consumption units, wherein the load characteristic curves include peak-valley distribution characteristics, load change trend characteristics, and load correlation characteristics; The second unit is used to construct a multi-objective optimization model of the distributed energy equipment group in the park based on the load characteristic curve, the optimization objectives of the multi-objective optimization model include minimizing the system operation cost and maximizing the energy utilization efficiency, and the constraints of the multi-objective optimization model include equipment operation parameter constraints, energy balance constraints and environmental constraints; using an improved particle swarm algorithm to solve the multi-objective optimization model to obtain the optimal operation plan of the distributed energy equipment group in the park, the optimal operation plan includes the power generation power of photovoltaic power generation equipment, the charging and discharging power of energy storage equipment and the cooling power of electric refrigeration equipment, wherein the improved particle swarm algorithm adopts an adaptive local search strategy and an elite solution dynamic maintenance mechanism on the basis of the particle swarm algorithm; The third unit is used to send control instructions to each distributed energy device through the park energy management system according to the optimal operation plan, and the control instructions include the equipment start and stop status, operating power and operating parameters; monitor the operating status of each distributed energy device in real time, and when abnormal equipment operation or sudden load change is detected, trigger the dynamic optimization correction mechanism, which dynamically updates the optimal operation plan based on the rolling time domain to ensure the coordinated operation of the park's distributed energy equipment group and the dynamic balance of the load.

8. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images