Multi-element new energy access power distribution network comprehensive planning method, system and equipment

Abstract

The invention is suitable for the technical field of power distribution networks, and provides a multi-element new energy access power distribution network comprehensive planning method, which comprises the following steps: respectively extracting energy storage characteristics of different energy storage types, and generating an energy storage characteristic parameter table; analyzing load data to obtain peak valley characteristics of different load types, and generating a load characteristic matrix; according to the load characteristic matrix, combining an energy storage characteristic parameter table, and adopting a rule engine algorithm to perform matching of an energy storage type and a load characteristic, so as to generate an energy storage combination scheme; based on a Monte Carlo simulation model, simulating and optimizing an energy storage combination scheme in combination with historical load data, energy storage technical economic parameters and power grid operation constraint conditions; the optimized energy storage combination scheme is quantitatively evaluated, if the evaluation result does not meet the set target threshold value, a compensation mechanism or a dynamic optimization mechanism is executed, the load balance capacity of the power distribution network is effectively improved, and efficient operation of a power system and deep utilization of renewable energy sources are achieved.

Description

Technical Field

[0001] This application belongs to the technical field of distribution networks, and particularly relates to a comprehensive planning method, system, and equipment for distribution networks with multiple new energy sources connected. Background Art

[0002] With the continuous advancement of the energy structure transformation, the grid connection of a high proportion of renewable energy and the growth of diversified loads pose dual challenges to the stable operation of distribution networks. In the existing technology, although the use of multiple energy storage technologies such as electrochemical energy storage and mechanical energy storage (including pumped-storage energy storage and compressed-air energy storage) is regarded as an effective means to improve the flexibility of the power grid, there are still multiple technical bottlenecks in actual engineering applications.

[0003] Specifically, different energy storage technologies such as electrochemical energy storage, pumped-storage energy storage, and compressed-air energy storage have differences in key parameters such as charge-discharge efficiency, capacity, and cycle life. At the same time, different types of loads such as industrial, commercial, and residential loads also have their own characteristics in terms of peak-valley differences, duration, and volatility. However, the existing technology lacks the precise matching of different energy storage technologies and load characteristics, resulting in the utilization rate of energy storage assets being lower than the design expectation.

[0004] In addition, on the premise of considering multiple factors such as the characteristics of energy storage technologies, load demands, grid constraints (such as line capacity and frequency stability), and economy (including costs and electricity prices), it is necessary to optimize the capacity configuration of the multiple energy storage system and synergistically optimize its charge-discharge strategy to maximize the system benefits. In the existing technology, historical operation data and simulation means are usually used to evaluate the peak shaving and valley filling and load smoothing effects of the multiple energy storage combination scheme in actual operation. However, when the actual effect does not meet the expectation, how to adjust the charge-discharge strategy (involving aspects such as power, time, introduction of prediction and intelligent algorithms, etc.) and energy storage configuration (including capacity, type, power, etc.) to optimize the system performance and ensure the achievement of the preset peak shaving and valley filling and load smoothing goals has become another problem to be solved. Summary of the Invention

[0005] The embodiments of this application provide a comprehensive planning method, system, and equipment for distribution networks with multiple new energy sources connected, which can solve one of the above-mentioned existing technical problems.

[0006] In the first aspect, the embodiments of this application provide a comprehensive planning method for distribution networks with multiple new energy sources connected, including:

[0007] In the second aspect, the embodiments of this application provide a comprehensive planning system for distribution networks with multiple new energy sources connected, including:

[0008] The first processing module is used to obtain the historical operation data set of the distribution network multi - energy storage system, extract the energy storage characteristics of different energy storage types respectively, and generate an energy storage characteristic parameter table, where the energy storage types include electrochemical energy storage, pumped - storage energy storage, and compressed - air energy storage;

[0009] The second processing module is used to obtain the load data of the power system, analyze the load data to obtain the peak - valley characteristics of different load types, and generate a load characteristic matrix, where the load types include industrial load, commercial load, and residential load;

[0010] The third processing module is used to match the energy storage type with the load characteristics by using a rule - engine algorithm according to the load characteristic matrix and in combination with the energy storage characteristic parameter table, and generate an energy storage combination plan;

[0011] The fourth processing module is used to simulate and optimize the energy storage combination plan based on the Monte Carlo simulation model, in combination with historical load data, energy storage technical and economic parameters, and grid operation constraint conditions;

[0012] The fifth processing module is used to calculate the peak - shaving amount, valley - filling amount, and load smoothness based on the optimized energy storage combination plan;

[0013] The sixth processing module is used to execute a compensation mechanism if one of the peak - shaving amount and the valley - filling amount is lower than the target peak - shaving and valley - filling threshold;

[0014] The seventh processing module is used to execute a dynamic optimization mechanism if the load smoothness is lower than the preset target load smoothness threshold.

[0015] In a third aspect, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above - mentioned comprehensive planning method for a distribution network with multiple new - energy accesses is implemented.

[0016] The beneficial effects of the embodiments of the present application compared with the prior art are:

[0017] The present invention discloses a comprehensive planning method for a distribution network with multiple new energy accesses. The method obtains the energy storage characteristics of different energy storage types and the load data of the power system, analyzes the load data to form a load characteristic matrix, matches the energy storage types with the load characteristics to form an energy storage combination plan, and optimizes the energy storage combination plan based on historical load data, energy storage technical and economic parameters and grid operation constraints. At the same time, by calculating key indicators such as peak shaving volume, valley filling volume and load smoothness, the optimized energy storage combination plan is quantitatively evaluated. If the evaluation result does not meet the set target threshold, a compensation mechanism or a dynamic optimization mechanism is respectively executed to ensure that the energy storage system can continuously meet the requirements of grid operation. The present invention effectively improves the load balancing ability of the distribution network through the intelligent scheduling and optimization of the multiple energy storage systems, realizes the efficient operation of the power system and the deep utilization of renewable energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0019] Figure 1 It is a schematic flowchart of a comprehensive planning method for a distribution network with multiple new energy accesses provided by an embodiment of the present invention;

[0020] Figure 2 It is a schematic structural diagram of a comprehensive planning system for a distribution network with multiple new energy accesses provided by an embodiment of the present invention;

[0021] Figure 3 It is a schematic structural diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] In the following description, specific details such as specific system structures and technologies are presented for the purpose of illustration rather than limitation, so as to thoroughly understand the embodiments of the present application. However, those skilled in the art should clearly understand that the present application can also be implemented in other embodiments without these specific details. In other cases, the detailed descriptions of well-known systems, devices, circuits and methods are omitted to avoid unnecessary details from interfering with the description of the present application.

[0023] It should be understood that when used in the specification and the appended claims of the present application, the term "comprising" indicates the presence of the described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their combinations.

[0024] It should also be understood that the term "and / or" used in the specification and appended claims of this application refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0025] As used in the specification and appended claims of this application, the term "if" may be interpreted as "when", "once", "in response to determining", or "in response to detecting" depending on the context. Similarly, the phrases "if determined" or "if [the described condition or event] is detected" may be interpreted as meaning "once determined", "in response to determining", "once [the described condition or event] is detected", or "in response to detecting [the described condition or event]" depending on the context.

[0026] In addition, in the description of the specification and appended claims of this application, the terms "first", "second", "third", etc. are only used for differential description and cannot be understood as indicating or implying relative importance.

[0027] Referring to "one embodiment" or "some embodiments" described in the specification of this application means that a specific feature, structure, or characteristic described in connection with that embodiment is included in one or more embodiments of this application. Thus, statements such as "in one embodiment", "in some embodiments", "in other some embodiments", "in still other embodiments", etc. that appear in different places in this specification do not necessarily all refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in another way. The terms "comprising", "including", "having" and their variants all mean "including but not limited to", unless otherwise specifically emphasized in another way.

[0028] Please refer to Figure 1 As shown, the present invention is a comprehensive planning method for a distribution network with multiple new energy accesses, including the following steps:

[0029] S100. Obtain the historical operation data set of the distribution network's multiple energy storage systems, extract the energy storage characteristics of different energy storage types respectively, and generate an energy storage characteristic parameter table, where the energy storage types include electrochemical energy storage, pumped-storage energy storage, and compressed air energy storage;

[0030] In some of these embodiments, the above step S100 includes:

[0031] Obtain the charge and discharge voltage and current values of electrochemical energy storage, the water level and power values of pumped-storage energy storage, and the pressure and temperature values of compressed air energy storage respectively, and generate an original operation data record table based on the energy storage type identification code;

[0032] Based on the original operation data record table, the charge-discharge efficiency parameters of each energy storage type are calculated, and the charge-discharge efficiency curve equation is established by using the least square method;

[0033] Based on the charge-discharge efficiency curve equation, the capacity parameters of each energy storage type are calculated, and the capacity parameters include the stored electricity of the electrochemical energy storage unit, the stored energy of the pumped-storage unit, and the stored gas volume of the compressed air energy storage unit;

[0034] The response time parameters of each energy storage type are obtained, and based on the charge-discharge efficiency parameters, the capacity parameters, and the response time parameters, an energy storage characteristic parameter table is generated.

[0035] In this embodiment, a data collector is used to obtain the charging voltage value, charging current value, discharging voltage value, and discharging current value from the electrochemical energy storage unit, the upper water level value, lower water level value, pumping power value, and power generation power value from the pumped-storage unit, and the air charging pressure value, air discharging pressure value, air charging temperature value, and air discharging temperature value from the compressed air energy storage unit, and an original operation data record table is generated according to the energy storage type identification code. Further, the charge-discharge efficiency value of the electrochemical energy storage is calculated according to the voltage and current values in the original operation data record table, the charge-discharge efficiency value of the pumped-storage is calculated according to the water level value and power value, and the charge-discharge efficiency value of the compressed air energy storage is calculated according to the pressure and temperature values. Specifically, during the operation of the electrochemical energy storage unit, the charging voltage value and charging current value are sampled in real time by the data collector, the charging input power is calculated by multiplying the voltage and current, the discharging voltage value and discharging current value are sampled and the discharging output power is calculated, and the charge-discharge efficiency value is obtained by dividing the output power by the input power. For example, in a data collection, a discharge occurs when the voltage drops from 48V to 44V, and the current drops from 20A to 15A, and the calculated charge-discharge efficiency is 92%. For the pumped-storage unit, the water level difference is calculated by recording the upper water level value and the lower water level value. In a data record, a discharge process occurs when the water level difference changes from 500 meters to 450 meters, the pumping power is 800MW, and the power generation power is 720MW, and the calculated charge-discharge efficiency is 90%. For the compressed air energy storage unit, in a data collection, its air charging pressure rises from 5MPa to 10MPa, the air charging temperature rises from 25°C to 35°C, and after a complete charge-discharge cycle, the air discharging pressure drops to 4MPa and the air discharging temperature drops to 20°C, then the calculated charge-discharge efficiency is 85%.

[0036] In this embodiment, for the charge-discharge efficiency curve equations of different types of energy storage devices, the least squares method is used to fit the charge-discharge efficiency data points. For example, in a possible embodiment, the electrochemical energy storage efficiency curve shows a relatively gentle downward trend, gradually decreasing from an initial 95% to 88%. The pumped-storage energy storage efficiency curve shows a correlation with the water level difference, and when the water level difference is greater than 400 meters, the efficiency remains above 90%. The compressed air energy storage efficiency curve is affected by both pressure and temperature, and reaches the optimal efficiency point at a pressure of 8 MPa and a temperature of 30°C.

[0037] In this embodiment, the stored electricity of the electrochemical energy storage unit is calculated according to the electrochemical energy storage efficiency curve, the stored energy of the pumped-storage energy storage unit is calculated according to the pumped-storage energy storage efficiency curve, and the gas storage volume of the compressed air energy storage unit is calculated through the compressed air energy storage efficiency curve. At the same time, the response time parameters, charge-discharge efficiency parameters, and capacity parameters of the electrochemical energy storage are extracted from the energy storage database, the response time parameters, charge-discharge efficiency parameters, and capacity parameters of the pumped-storage energy storage, and the response time parameters, charge-discharge efficiency parameters, and capacity parameters of the compressed air energy storage. Combining the stored electricity of the electrochemical energy storage unit, the stored energy of the pumped-storage energy storage unit, and the gas storage volume of the compressed air energy storage unit generates an energy storage characteristic parameter table for use in the subsequent generation of the energy storage combination scheme. It can be understood that the energy storage database is a database specifically used to store data related to the energy storage system, which specifically includes detailed parameters, performance data, historical operation records, and other information of various energy storage types.

[0038] S200. Obtain the load data of the power system, and by analyzing the load data, obtain the peak-valley characteristics of different load types to generate a load characteristic matrix. The load types include industrial load, commercial load, and residential load.

[0039] In some of these embodiments, the above step S200 includes:

[0040] Obtain the three-phase active power values of the distribution transformer according to the sampling time identifier, and classify and mark the three-phase active power values with the load type identification code to generate an original load data record table.

[0041] Based on the original load data record table, use the maximum-minimum value detection algorithm to identify the peak points and valley points in the daily load curve, and generate a peak-valley characteristic data table based on the load type identification code.

[0042] Based on the peak-valley characteristic data table, obtain the peak-valley differences of each load type, and perform normalization processing on the peak-valley differences through the maximum-minimum value normalization method to generate a normalized peak-valley difference data table.

[0043] Based on the time identifiers of the peak points and valley points in the peak-valley characteristic data table, calculate and obtain the peak duration and valley duration to generate a peak-valley duration data table for each load type.

[0044] Calculate the power change amount based on the three-phase active power values at adjacent moments, obtain the volatility data through the ratio of the power change amount to the average power value, and classify and label the volatility data with the load type identification code to generate a volatility data table.

[0045] Construct a feature vector based on the standardized peak-valley difference data table, the peak-valley duration data table, and the volatility data table, and use the principal component analysis method to extract the load characteristics to generate a load characteristic matrix.

[0046] In this embodiment, the three-phase active power values are collected from the distribution transformer by the load data collector at 15-minute intervals, and the three-phase active power values of industrial loads, commercial loads, and residential loads are classified and labeled with the load type identification code. The active power values are arranged in time series according to the sampling time identification, and an original load data record table is generated. For the original load data record table, the abnormal three-phase active power values in the original load data record table are filtered by using a sliding median filter, the peak points and valley points in the daily load curve are identified according to the maximum-minimum value detection algorithm, a peak-valley characteristic data table is generated according to the load type identification code, the peak-valley differences of industrial loads, commercial loads, and residential loads are calculated based on the peak-valley characteristic data table, and the above peak-valley differences are standardized by using the maximum-minimum value normalization method to generate a standardized peak-valley difference data table.

[0047] In this embodiment, obtain the sampling time identification of the peak points and valley points of each load type in the peak-valley characteristic data table, calculate the peak duration and valley duration of each load type, and use the time window statistical method to obtain the peak-valley duration data table of industrial loads, commercial loads, and residential loads.

[0048] In this embodiment, calculate the power change amount by using the power values at adjacent moments recorded by the data collector, obtain the volatility value through the ratio of the change amount to the average power value, distinguish the volatility data of industrial loads, commercial loads, and residential loads according to the load type identification code, and generate a volatility data table.

[0049] In this embodiment, construct a feature vector based on the standardized peak-valley difference data table, the peak-valley duration data table, and the volatility data, and use the principal component analysis method to extract the load characteristics and generate a load characteristic matrix.

[0050] Specifically, the load data collector monitors the distribution transformer in real time and records the three-phase active power values every 15 minutes. In one embodiment, the specific data it collects is as follows: The industrial load shows high-load characteristics in the interval from 8:00 to 17:00 on weekdays, with a peak power reaching 1200 kW, and operates at a low-load state from 22:00 at night to 6:00 the next day, with the valley power dropping to 300 kW; The commercial load maintains a relatively high power consumption level from 10:00 to 22:00, with a peak power of 800 kW, and the power drops to 200 kW during other periods; The residential load shows a double-peak characteristic from 7:00 to 9:00 in the morning and from 18:00 to 22:00 in the evening, with a peak power of 500 kW, and drops to the lowest valley value of 100 kW from 3:00 to 5:00 in the early morning. By analyzing the above-collected data, there are outliers in the original load data caused by equipment start-stop and load mutations. Therefore, smoothing processing is performed through a 5-point sliding median filter to filter out the mutation data points beyond the normal range. Then, peak-valley detection is performed on the processed power curve, and the peak-valley difference of the industrial load is obtained as 900 kW, the peak-valley difference of the commercial load is 600 kW, and the peak-valley difference of the residential load is 400 kW. Further, the maximum-minimum normalization method is used to map the peak-valley difference to the interval from 0 to 1. At this time, the standardized peak-valley difference of the industrial load is 1.0, the commercial load is 0.67, and the residential load is 0.44, generating a standardized peak-valley difference data table. In terms of the duration, the peak period of the industrial load lasts for 9 hours, and the valley period lasts for 8 hours, showing a regular power consumption characteristic; The peak period of the commercial load lasts for 12 hours, and the valley period lasts for 10 hours, with a relatively large span of power consumption periods; The morning and evening peak periods of the residential load each last for 2 to 3 hours, and the valley period lasts for 3 hours, showing an obvious double-peak characteristic. At the same time, the load volatility is used to reflect the severity of power changes. The industrial load is affected by the start-stop of production equipment, and the power change amount between adjacent moments can reach 200 kW, with a maximum volatility of 0.25. The commercial load fluctuation mainly comes from equipment such as air conditioners and lighting, with a power change amount of about 100 kW and a volatility of 0.15. The residential load is affected by lighting and household appliance usage habits, with a power change amount of 50 kW and a volatility of 0.12. The principal component analysis method is used to extract load characteristics from three dimensions: standardized peak-valley difference, peak-valley duration, and volatility. The industrial load characteristic vector shows the characteristics of a large peak-valley difference and a stable duration, the commercial load characteristic vector shows the characteristics of a medium peak-valley difference and a long duration, and the residential load characteristic vector shows the rule of a small peak-valley difference and a short double-peak duration. The load characteristic matrix forms a complete description of the power consumption behaviors of different load types by combining the above load characteristic vectors.

[0051] S300. According to the load characteristic matrix, combined with the energy storage characteristic parameter table, use the rule engine algorithm to match the energy storage type with the load characteristics and generate an energy storage combination plan;

[0052] In some of these embodiments, the above step S300 includes:

[0053] Establish a capacity fitness calculation function, an efficiency fitness calculation function, and a response fitness calculation function, and based on the energy storage characteristic parameter table, use a deep neural network to train a fitness prediction model;

[0054] Input the load characteristic matrix into the fitness prediction model to obtain the energy storage fitness values of each load type, where the energy storage fitness values include capacity fitness values, efficiency fitness values, and response fitness values;

[0055] Construct an energy storage type discriminator, use the energy storage fitness values of each load type as input features, classify and evaluate the electrochemical energy storage, the pumped-storage energy storage, and the compressed air energy storage, and generate an energy storage type fitness evaluation table;

[0056] Based on the energy storage type fitness evaluation table, generate an energy storage capacity allocation ratio threshold, and generate an electrochemical energy storage configuration capacity value, a pumped-storage energy storage configuration capacity value, and a compressed air energy storage configuration capacity value according to the energy storage capacity allocation ratio threshold to form an energy storage combination plan.

[0057] In this embodiment, a capacity fitness calculation function is established through peak-valley difference data, an efficiency fitness calculation function is established through peak-valley duration data, a response fitness calculation function is established through volatility data, and based on the energy storage characteristics of each energy storage type in the energy storage characteristic parameter table, a fitness prediction model is trained using a deep neural network for matching the energy storage type with the load characteristics.

[0058] In this embodiment, an industrial load characteristic vector, a commercial load characteristic vector, and a residential load characteristic vector are input into the above fitness prediction model. The fitness prediction model calculates the energy storage capacity fitness value, the efficiency fitness value, and the response fitness value for the industrial load characteristic vector, calculates the energy storage capacity fitness value, the efficiency fitness value, and the response fitness value for the commercial load characteristic vector, and calculates the energy storage capacity fitness value, the efficiency fitness value, and the response fitness value for the residential load characteristic vector.

[0059] In this embodiment, a decision tree algorithm is used to construct an energy storage type discriminator. Using the above fitness values as input features, the electrochemical energy storage, the pumped-storage energy storage, and the compressed air energy storage are classified and evaluated to generate an energy storage type fitness evaluation table. Further, an energy storage capacity allocation ratio threshold is set according to the energy storage type fitness evaluation table, and an electrochemical energy storage configuration capacity value, a pumped-storage energy storage configuration capacity value, and a compressed air energy storage configuration capacity value are generated according to the capacity allocation ratio, and finally an energy storage combination plan is formed. Optionally, for load characteristics with large peak-valley differences and long durations, pumped-storage energy storage and compressed air energy storage are used, and for load characteristics with high volatility, electrochemical energy storage is used.

[0060] In some embodiments, for the energy storage combination scheme, the total energy storage capacity value is calculated, and the capacity matching degree of the energy storage configuration scheme is verified by using the capacity balance constraint function.

[0061] In one embodiment, the energy storage characteristics of different energy storage types recorded in the energy storage technology characteristic data table are as follows: the response time of electrochemical energy storage is in the millisecond level, up to 50 milliseconds at the fastest, the charge-discharge efficiency reaches 95%, and the stored electricity of the electrochemical energy storage unit is between 100 kWh and 1000 kWh. The response time of pumped storage is in the minute level, the start-up time is about 2 minutes, the charge-discharge efficiency is 80%, and the stored energy of the pumped storage unit is 300 MWh. The response time of compressed air energy storage is in the second level, the start-up time is 30 seconds, the charge-discharge efficiency is 75%, and the gas storage volume of the compressed air energy storage unit is between 10 MWh and 100 MWh. For the peak-valley difference data in the load characteristic matrix, the energy storage capacity parameters of the above-mentioned various energy storage types and the peak-valley difference data are matched and evaluated through the capacity adaptability calculation function. For example, the energy storage capacity demand corresponding to the industrial load peak-valley difference of 900 kW is 1200 kWh, the energy storage capacity demand corresponding to the commercial load peak-valley difference of 600 kW is 800 kWh, and the energy storage capacity demand corresponding to the residential load peak-valley difference of 400 kW is 500 kWh. The efficiency adaptability calculation function is evaluated based on the peak-valley duration. The 9-hour duration of industrial load is suitable for pumped storage, the 12-hour duration of commercial load is suitable for compressed air energy storage, and the 2-3-hour short-term fluctuation of residential load is suitable for electrochemical energy storage. Based on this, based on the capacity adaptability calculation function, the efficiency adaptability calculation function, the response adaptability calculation function, and the energy storage characteristics of each energy storage type in the energy storage characteristic parameter table, an adaptability prediction model is trained by a deep neural network. Through the adaptability prediction model, the capacity adaptability of the electrochemical energy storage corresponding to the industrial load feature vector is 0.3, the efficiency adaptability is 0.4, and the response adaptability is 0.8. The three adaptabilities of pumped storage are 0.9, 0.85, and 0.4 respectively, and the adaptability of compressed air energy storage is 0.6, 0.7, and 0.6. The commercial load and the residential load also obtain corresponding adaptability values. Further, the energy storage types are classified according to the adaptability values through the decision tree algorithm. For example, in the above embodiment, the pumped storage with the highest matching degree for industrial load is selected. Therefore, the recommended configuration ratio is 60%, followed by compressed air energy storage accounting for 30%, and electrochemical energy storage accounting for 10%.Similarly, the same method is used to analyze commercial loads and residential loads to further obtain the energy storage configuration ratios for commercial loads and residential loads. For example, commercial loads mainly use compressed air energy storage, accounting for 50%, pumped hydro energy storage and electrochemical energy storage each account for 25%. Residential loads mainly use electrochemical energy storage, accounting for 70%, compressed air energy storage accounts for 20%, and pumped hydro energy storage accounts for 10%. Then, an energy storage type adaptability evaluation table is generated. Subsequently, the capacity allocation ratio is determined according to the energy storage type adaptability evaluation table. In a possible embodiment, the industrial load is configured with 1000 MWh of pumped hydro energy storage, 500 MWh of compressed air energy storage, and 100 MWh of electrochemical energy storage. The commercial load is configured with 300 MWh of compressed air energy storage, 150 MWh of pumped hydro energy storage, and 150 MWh of electrochemical energy storage. The residential load is configured with 200 MWh of electrochemical energy storage, 60 MWh of compressed air energy storage, and 40 MWh of pumped hydro energy storage. Then, an energy storage combination plan is generated. Verification by the capacity balance constraint function shows that this energy storage combination plan meets the peak-valley filling requirements, and the matching degree between the energy storage type and the load characteristics reaches more than 90%.

[0062] S400. Based on the Monte Carlo simulation model, combined with historical load data, energy storage technology economic parameters, and grid operation constraint conditions, simulate and optimize the energy storage combination plan;

[0063] In some of these embodiments, the above step S400 includes:

[0064] Obtain the historical load power data of the distribution transformer through a load collector, and generate a daily load power curve data table according to the historical load power data at the sampling time interval;

[0065] According to the daily load power curve data table, use a deep neural network to train a load sequence predictor to obtain a load prediction curve data table;

[0066] Based on the load prediction curve data table, mark the peak period and the trough period, and use the Monte Carlo method to generate a set of energy storage charge and discharge sequences;

[0067] Based on the grid operation constraint conditions, verify the constraint conditions for each energy storage charge and discharge sequence to generate a set of valid charge and discharge sequences;

[0068] Based on the set of charge and discharge sequences and the energy storage technology economic parameters, use the particle swarm algorithm to construct a multi-objective optimizer, with the goal of minimizing the investment cost and maximizing the peak shaving and valley filling amount, to optimize the energy storage combination plan.

[0069] In this embodiment, historical load power data is obtained from a distribution transformer through a load collector, and the historical load power data is arranged in time series according to the sampling time interval to generate a daily load power curve data table. A load sequence predictor is established using a deep neural network, and the predictor parameters are trained based on the daily load power curve data table to generate a load prediction curve data table. The peak period and valley period are marked according to the load prediction curve data table, and a charge-discharge sequence of energy storage is generated based on the Monte Carlo method. The constraint conditions of each charge-discharge sequence are verified to generate a set of valid charge-discharge sequences. A multi-objective optimizer is constructed using the particle swarm algorithm, with the minimization of investment cost and the maximization of peak shaving and valley filling as the optimization objectives, to generate an optimized energy storage operation plan.

[0070] In this embodiment, for the load sequence predictor, 30-day historical load data is used for training, that is, the daily load power curve data tables for thirty days are used as input features. The number of hidden layer nodes is set to 128, the learning rate is 0.001, and the number of training epochs is 1000. After training, a load sequence predictor is generated, and a load prediction data table can be obtained, with a prediction accuracy rate reaching 95%.

[0071] In this embodiment, the technical and economic parameters of energy storage include energy storage cost parameters and energy storage technology parameters. Among them, the energy storage cost parameters include the investment cost per unit capacity of electrochemical energy storage, the investment cost per unit capacity of pumped-storage energy storage, and the investment cost per unit capacity of compressed air energy storage. The energy storage technology parameters include charge-discharge efficiency and cycle life. For the grid operation constraint conditions, the maximum transmission capacity of the line, the power factor limit, and the allowable frequency deviation are extracted according to the distribution network operation regulations to generate a grid constraint condition data table.

[0072] In one embodiment, the load collector collects the power data of the distribution transformer at 15-minute intervals and records the 24-hour load curve. Among them, the power during the peak period on weekdays reaches 1200 kW, and drops to 300 kW during the valley period. The peak power on holidays is 800 kW, and the valley power is 200 kW. The daily load curve is formed according to the time series arrangement to record the power change law within one year. For the technical and economic parameters of energy storage, in terms of the energy storage investment cost, the investment cost per unit capacity of electrochemical energy storage is 1500 yuan / kWh, which has the characteristics of fast response speed and high regulation accuracy and is suitable for dealing with short-term power fluctuations. The investment cost per unit capacity of pumped-storage energy storage is 2000 yuan / kWh, which has the advantages of large energy storage capacity and long life and is suitable for large-scale energy transfer. The investment cost per unit capacity of compressed air energy storage is 1200 yuan / kWh, which has the characteristics of low construction cost and flexible regulation and is suitable for medium-scale peak shaving and valley filling. The distribution network operation constraints include the maximum transmission capacity of the line of 1500 kW, the power factor not less than 0.95, and the allowable system frequency deviation of ±0.2 Hz. The above constraint conditions determine the upper limit of the charge-discharge power of the energy storage device to avoid line overload or system frequency overlimit caused by energy storage regulation.

[0073] Specifically, 1000 sets of energy storage charge and discharge sequences are randomly generated based on the Monte Carlo method. Each sequence contains the charge and discharge power values at 96 time points. The maximum charging power is set to 400 kW, and the maximum discharging power is set to 500 kW. 800 sets of effective sequences are screened out through constraint verification. During the peak hours from 8 to 11 and from 14 to 17, energy storage is discharged, and during the valley hours from 23 to 5 the next day, energy storage is charged. The calculated average peak shaving amount is 300 kW, and the valley filling amount is 250 kW. The particle swarm algorithm sets the population size to 100 and the maximum number of iterations to 500. An optimization objective function is constructed with the energy storage cost parameters, energy storage technology parameters, and peak shaving and valley filling amounts of each energy storage type. The optimization results show that the scheme of configuring 300 kWh of electrochemical energy storage, 500 kWh of compressed air energy storage, and 1000 kWh of pumped-storage energy storage has better comprehensive benefits. Thus, using the above method, the energy storage configuration schemes for industrial loads, commercial loads, and residential loads are optimized in sequence, and then the energy storage combination scheme generated in the above step S300 is optimized.

[0074] S500. Based on the optimized energy storage combination scheme, calculate the peak shaving amount, valley filling amount, and load smoothness;

[0075] In some of these embodiments, the above step S500 includes:

[0076] Obtain the configuration ratios of each energy storage type from the optimized energy storage combination scheme;

[0077] According to the peak-valley difference threshold of the load curve, perform time period division within a preset time period to obtain a peak interval and a valley interval;

[0078] In the peak interval, calculate the energy storage discharge power, and in the valley interval, calculate the energy storage charging power. Based on the energy storage discharge power and the energy storage charging power, generate an energy storage power adjustment data table;

[0079] According to the energy storage power adjustment data table, calculate the peak shaving amount during peak hours and the valley filling amount during valley hours;

[0080] Based on the load sequence predictor, obtain a load prediction curve data table, calculate the load mean square error value through the load prediction curve data, and generate the load smoothness.

[0081] In this embodiment, the electrochemical energy storage capacity ratio value, pumped-storage energy storage capacity ratio value, and compressed air energy storage capacity ratio value of each load type are obtained from the optimized energy storage combination scheme, and the peak interval and valley interval within the preset time period are divided according to the preset peak-valley difference threshold of the load curve. For example, if a preset peak-valley difference threshold of the load curve is 400 kW, according to this peak-valley difference threshold of the load curve, the time periods from 8:00 to 11:00 and from 14:00 to 17:00 in a day are marked as peak intervals, and the time period from 23:00 to 5:00 the next day is marked as the valley interval.

[0082] In this embodiment, during the peak interval, the energy storage discharge power is dynamically adjusted according to the load demand. For example, in a possible embodiment, the electrochemical energy storage discharge power is 200 kW, with a duration of 3 hours, the pumped-storage energy storage discharge power is 500 kW, with a duration of 6 hours, and the compressed air energy storage discharge power is 150 kW, with a duration of 4 hours. And for the valley interval, energy storage charging is performed. The charging powers of the electrochemical energy storage, pumped-storage energy storage, and compressed air energy storage are 180 kW, 450 kW, and 130 kW respectively, thereby obtaining a peak shaving of 850 kW and a valley filling of 760 kW.

[0083] In this embodiment, for the load smoothness, through the load sequence predictor, a load prediction curve data table is obtained, and based on this load prediction curve data table, the load smoothness is calculated through the power mean square deviation.

[0084] S600. If one of the peak shaving amount and the valley filling amount is lower than the target peak shaving and valley filling threshold, then execute the compensation mechanism;

[0085] In this embodiment, if one of the peak shaving value and the valley filling value is lower than the target peak shaving and valley filling threshold, then obtain the charge and discharge data of the energy storage combination scheme within the preset time period, analyze the volatility of the peak shaving value and the valley filling value, calculate the difference between the peak shaving and valley filling effect and the preset target, determine the change range of the load rate, and adjust the peak shaving amount and the valley filling amount according to the analysis result of the volatility of the peak shaving value and the valley filling value.

[0086] In some of these embodiments, the execution of the compensation mechanism includes:

[0087] Adopt a deep neural network to construct a power fluctuation predictor for extracting the fluctuation characteristics of the peak shaving amount and the valley filling amount to generate a power fluctuation characteristic table;

[0088] Based on the power fluctuation characteristic table, extract the fluctuation period value and the fluctuation amplitude, optimize the boundaries of the charging period and the discharging period to generate a period optimization data table;

[0089] Based on the historical load data, obtain the upper and lower limit intervals of the load rate, determine the load regulation interval, and generate a charging power correction value and a discharging power correction value based on the power fluctuation characteristic table;

[0090] Based on the charging power correction value and the discharging power correction value, and in combination with the time period optimization data table, charging power adjustment and discharging power adjustment are carried out.

[0091] In this embodiment, a long short-term memory network is used to construct a load interval predictor. Using historical load data as training samples, the upper and lower limits of the load rate are calculated to determine the load regulation interval. For example, in a possible embodiment, the long short-term memory network is trained with 30 days of historical load data, obtaining an upper limit of the load rate of 85% and a lower limit of 35%, leaving a 15% upward adjustment space and a 5% downward adjustment space.

[0092] In one embodiment, during the charging period, i.e., the low valley interval from 23:00 to 5:00 the next day, the charging power of the electrochemical energy storage is 100 kW, the charging power of the pumped-storage energy storage is 400 kW, and the charging power of the compressed air energy storage is 150 kW. During the discharging period, i.e., the peak interval, the morning peak from 8:00 to 11:00 and the evening peak from 17:00 to 20:00, the discharging powers of the three energy storage types are 120 kW, 450 kW, and 180 kW respectively. The calculated peak shaving power is 750 kW, and the valley filling power is 650 kW, where the valley filling power is lower than the preset threshold of 700 kW. Thus, through the power fluctuation predictor constructed by the deep neural network, fluctuation analysis is carried out on the charging and discharging power sequences, obtaining that the fluctuation amplitude of the charging power of the electrochemical energy storage is 20 kW and the period is 1 hour, the fluctuation amplitude of the charging power of the pumped-storage energy storage is 50 kW and the period is 2 hours, the fluctuation amplitude of the charging power of the compressed air energy storage is 30 kW and the period is 1.5 hours, while the power fluctuation characteristics during the discharging period show that the fluctuation amplitude of the electrochemical energy storage is 25 kW, the fluctuation amplitude of the pumped-storage energy storage is 60 kW, and the fluctuation amplitude of the compressed air energy storage is 35 kW. Based on the above fluctuation characteristics, boundary optimization is carried out for the charging period and the discharging period. The charging period of the electrochemical energy storage is adjusted to 22:00 to 4:00 the next day, the pumped-storage energy storage remains unchanged from 23:00 to 5:00 the next day, and the compressed air energy storage is adjusted to 23:00 to 4:00 the next day. In terms of the discharging period, the morning peak of the electrochemical energy storage is adjusted to 7:30 to 10:30, the evening peak remains at 17:00 to 20:00, and the discharging periods of the pumped-storage energy storage and the compressed air energy storage remain unchanged. And through the load interval predictor, the upper limit of the load rate is obtained as 85% and the lower limit is 35%. On this basis, it is set that the charging power of the electrochemical energy storage is increased to 120 kW, the discharging power is increased to 140 kW, the charging power of the pumped-storage energy storage remains unchanged at 400 kW, the discharging power is increased to 470 kW, the charging power of the compressed air energy storage is increased to 170 kW, and the discharging power is increased to 200 kW.

[0093] In this embodiment, after adjusting the charging power and discharging power, it is necessary to verify the charging and discharging effects through operation. As in the above embodiment, when the charging power of the electrochemical energy storage is increased to 120 kW, the discharging power is increased to 140 kW, the charging power of the pumped-storage energy storage remains unchanged at 400 kW, the discharging power is increased to 470 kW, and the charging power of the compressed air energy storage is increased to 170 kW and the discharging power is increased to 200 kW, the total charging power of each energy storage type during the charging period reaches 690 kW, which is 40 kW higher than the original value, and the total discharging power during the discharging period reaches 810 kW, which is 60 kW higher than the original value. After one day of operation verification, the new charging and discharging power configuration increases the valley filling power to 720 kW, exceeding the preset threshold requirement, and the peak shaving power remains at the level of 750 kW, meeting the dual requirements of peak shaving and valley filling. At the same time, the amplitude of the charging and discharging power fluctuations of each energy storage type is also significantly improved. The fluctuation of the electrochemical energy storage drops to 15 kW, the pumped-storage energy storage drops to 40 kW, and the compressed air energy storage drops to 25 kW.

[0094] S700. If the load smoothness is lower than the preset target load smoothness threshold, execute the dynamic optimization mechanism.

[0095] In this embodiment, if the load smoothness value is lower than the preset target load smoothness threshold, it indicates that the load fluctuation is large and needs to be smoothed through energy storage regulation, that is, optimize the charging and discharging strategies of the energy storage, such as adjusting the charging and discharging power and time, introducing control strategies based on load and renewable energy prediction, using reinforcement learning algorithms to increase the energy storage capacity, optimizing the energy storage type combination or adjusting the energy storage power.

[0096] In some of these embodiments, the execution of the dynamic optimization mechanism includes:

[0097] Based on the target load smoothness threshold, generate a charging and discharging power sequence, use a deep neural network to construct a charging and discharging period divider, segment the charging and discharging power sequence, and generate an energy storage scheduling period table;

[0098] Extract the power generation prediction values from the photovoltaic power curve and the wind power curve, and establish a power balance constraint function through the power generation prediction values;

[0099] Based on the power balance constraint function, use a reinforcement learning algorithm to construct an energy storage capacity optimizer to dynamically adjust the configuration of each energy storage type.

[0100] In this embodiment, based on the target load smoothing threshold, a charge-discharge power sequence is generated, that is, the generated charge-discharge power sequence needs to meet the target of load smoothness. Then, a deep neural network is used to construct a charge-discharge period divider to optimally segment the charge-discharge power sequence and generate an energy storage scheduling period table. Renewable energy such as photovoltaic power generation and wind power generation is introduced, the power generation prediction values are extracted from the photovoltaic power generation curve and the wind power curve, and a power balance constraint function is established in combination with the load prediction value. A reinforcement learning algorithm is used to construct an energy storage capacity optimizer, and the power balance constraint function is used as the basis for calculating the reward to dynamically adjust the configurations of electrochemical energy storage, pumped-storage energy storage, and compressed-air energy storage.

[0101] In one embodiment, a deep neural network is used to perform segmented optimization on the generated charge-discharge power sequence. For example, the charging period is set from 23:00 to 5:00 the next day, and the discharging period is divided into the morning peak from 8:00 to 11:00 and the evening peak from 17:00 to 20:00. During the morning peak period on weekdays, the discharging power of the energy storage is set to 400 kW and lasts for 3 hours. During the evening peak, the discharging power is 350 kW and lasts for 3 hours. During the low valley period at night, the charging power is 300 kW and lasts for 6 hours. For the renewable energy power generation side, by analyzing the photovoltaic power generation curve and the wind power curve, it is obtained that the photovoltaic power generation curve of the photovoltaic power station presents a typical single-peak characteristic, and the power generation reaches the maximum value of 500 kW from 10:00 to 14:00, and the power generation is relatively low in the early morning and evening. The power generation of the wind farm shows a random fluctuation characteristic, with an average power generation of 300 kW and a power fluctuation range between 100 kW and 500 kW. Then, the power generation prediction values of photovoltaic power generation and wind power generation are obtained. Further, with the power balance constraint as the reward benchmark through the reinforcement learning algorithm, a positive reward is given when the standard deviation of the load curve after energy storage regulation decreases, and a negative reward is given otherwise. The optimization results show that 2 additional 100 kWh energy storage units are required for electrochemical energy storage, the pumped-storage energy storage maintains a 2-unit 1000 kWh installed capacity, and 1 additional 200 kWh energy storage unit is required for compressed-air energy storage. The total energy storage capacity increases from the original 2800 kWh to 3200 kWh, fully meeting the load regulation requirements.

[0102] In some embodiments, the configurations of each generated energy storage type are simulated and verified. Specifically, 1000 groups of simulation scenarios are generated by the Monte Carlo method and verified. As in the above embodiments, after adding 2 sets of 100 kWh energy storage units for electrochemical energy storage, maintaining 2 sets of 1000 kWh installed capacity for pumped-storage energy storage, and adding 1 set of 200 kWh energy storage unit for compressed air energy storage, the simulation verification results show that the load standard deviation in more than 95% of the scenarios is reduced to below 150 kW. After 3 rounds of iterative optimization, the scheme achieves a load peak shaving of 900 kW, valley filling of 850 kW, and the load smoothing index value is reduced to 0.12, which is better than the preset threshold requirement. The energy storage system completes a full charge-discharge cycle every day, the matching degree between the charging amount and the discharging amount reaches 95%, and the energy storage capacity margin is maintained at about 20%, ensuring the safe and stable operation of the energy storage system.

[0103] It can be understood that when implementing the compensation mechanism and the dynamic optimization mechanism, it is necessary to optimize the energy storage configuration schemes for industrial loads, commercial loads, and residential loads in sequence, and then generate a new energy storage combination scheme.

[0104] Please refer to Figure 2 As shown in the figure, the present invention also provides a comprehensive distribution network planning system for multi-source new energy access, and the system includes:

[0105] A first processing module 201, configured to obtain the historical operation data set of the distribution network multi-energy storage system, extract the energy storage characteristics of different energy storage types respectively, and generate an energy storage characteristic parameter table, where the energy storage types include electrochemical energy storage, pumped-storage energy storage, and compressed air energy storage;

[0106] A second processing module 202, configured to obtain the load data of the power system, analyze the load data to obtain the peak-valley characteristics of different load types, and generate a load characteristic matrix, where the load types include industrial loads, commercial loads, and residential loads;

[0107] A third processing module 203, configured to match the energy storage type with the load characteristics by using a rule engine algorithm according to the load characteristic matrix and in combination with the energy storage characteristic parameter table, and generate an energy storage combination scheme;

[0108] A fourth processing module 204, configured to simulate and optimize the energy storage combination scheme based on a Monte Carlo simulation model in combination with historical load data, energy storage technical and economic parameters, and power grid operation constraint conditions;

[0109] A fifth processing module 205, configured to calculate the peak shaving amount, valley filling amount, and load smoothness based on the optimized energy storage combination scheme;

[0110] A sixth processing module 206, configured to execute a compensation mechanism if one of the peak shaving amount and the valley filling amount is lower than the target peak shaving and valley filling threshold;

[0111] A seventh processing module 207, configured to execute a dynamic optimization mechanism if the load smoothness is lower than a preset target load smoothness threshold.

[0112] It can be understood that the content in the embodiments of the integrated distribution network planning method for multi-source new energy access as Figure 1 shown is applicable to the embodiments of the integrated distribution network planning system for multi-source new energy access in this application. The functions specifically implemented in the embodiments of the integrated distribution network planning system for multi-source new energy access in this application are the same as those in the embodiments of the integrated distribution network planning method for multi-source new energy access as Figure 1 shown, and the beneficial effects achieved are also the same as those achieved in the embodiments of the integrated distribution network planning method for multi-source new energy access as Figure 1 shown.

[0113] It should be noted that for the information interaction, execution process, etc. between the above systems, since they are based on the same concept as the method embodiments of the present invention, for their specific functions and the technical effects brought, reference can be specifically made to the method embodiment part, and details are not described here again.

[0114] Those skilled in the art can clearly understand that for the convenience and brevity of description, only the above division of each functional unit and module is used as an example for illustration. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the system is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiments can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction and do not limit the protection scope of this application. The specific working processes of the units and modules in the above system can refer to the corresponding processes in the foregoing method embodiments, and details are not described here again.

[0115] Please refer to Figure 3 shown. An embodiment of the present invention further provides a computer device 3, including: a memory 302, a processor 301, and a computer program 303 stored on the memory 302. When the computer program 303 is executed on the processor 301, the integrated distribution network planning method for multi-source new energy access as described in any one of the above methods is implemented.

[0116] The computer device 3 may be a computing device such as a desktop computer, a notebook, a palm computer, or a cloud server. The computer device 3 may include, but is not limited to, a processor 301 and a memory 302. Those skilled in the art can understand that Figure 3 merely examples of the computer device 3, which do not constitute a limitation on the computer device 3, may include more or fewer components than shown in the figure, or combine certain components, or different components. For example, it may also include input / output devices, network access devices, etc.

[0117] The so-called processor 301 may be a central processing unit (CPU), and the processor 301 may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0118] The memory 302 may be an internal storage unit of the computer device 3 in some embodiments, such as the hard disk or memory of the computer device 3. The memory 302 may also be an external storage device of the computer device 3 in some other embodiments, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the computer device 3. Further, the memory 302 may also include both the internal storage unit and the external storage device of the computer device 3. The memory 302 is used to store an operating system, application programs, a boot loader, data, and other programs, such as the program code of the computer program. The memory 302 may also be used to temporarily store data that has been output or will be output.

[0119] The embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is run by a processor, it implements the integrated planning method for a distribution network with multiple new energy accesses as described in any one of the above methods.

[0120] In this embodiment, if the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, to implement all or part of the processes in the above-described embodiment methods of this application, a computer program can be used to instruct the relevant hardware to complete. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-described various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can at least include: any entity or device that can carry the computer program code to the photographing device / computer device, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium. For example, a USB flash drive, a portable hard drive, a magnetic disk, or an optical disc, etc. In some jurisdictions, according to legislation and patent practice, the computer-readable medium cannot be an electrical carrier signal and a telecommunication signal.

[0121] The above-described embodiments are only used to illustrate the technical solutions of this application, rather than to limit them; although this application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A comprehensive planning method for a distribution network with multiple new energy accesses, characterized in that Including: Obtain the historical operation data set of the distribution network multi - energy storage system, extract the energy storage characteristics of different energy storage types respectively, and generate an energy storage characteristic parameter table, where the energy storage types include electrochemical energy storage, pumped - storage energy storage, and compressed - air energy storage; Obtain the load data of the power system, analyze the load data to obtain the peak - valley characteristics of different load types, and generate a load characteristic matrix, where the load types include industrial load, commercial load, and residential load; According to the load characteristic matrix, combined with the energy storage characteristic parameter table, use the rule - engine algorithm to match the energy storage type and the load characteristics, and generate an energy storage combination plan; Based on the Monte Carlo simulation model, combined with historical load data, energy storage technical - economic parameters, and grid operation constraint conditions, simulate and optimize the energy storage combination plan; Based on the optimized energy storage combination plan, calculate the peak - shaving volume, valley - filling volume, and load smoothness; If one of the peak - shaving volume and the valley - filling volume is lower than the target peak - shaving and valley - filling threshold, execute the compensation mechanism; If the load smoothness is lower than the preset target load smoothness threshold, execute the dynamic optimization mechanism.

2. The method according to claim 1, wherein The obtaining the historical operation data set of the distribution network multi - energy storage system, extracting the energy storage characteristics of different energy storage types respectively, and generating an energy storage characteristic parameter table, where the energy storage types include electrochemical energy storage, pumped - storage energy storage, and compressed - air energy storage, includes: Obtain the charge - discharge voltage and current values of electrochemical energy storage, the water - level and power values of pumped - storage energy storage, and the pressure and temperature values of compressed - air energy storage respectively, and generate an original operation data record table based on the energy storage type identification code; Based on the original operation data record table, calculate the charge - discharge efficiency parameters of each energy storage type, and establish a charge - discharge efficiency curve equation using the least - squares method; Based on the charge - discharge efficiency curve equation, calculate the capacity parameters of each energy storage type, where the capacity parameters include the stored electric energy of the electrochemical energy storage unit, the stored energy of the pumped - storage energy storage unit, and the stored gas volume of the compressed - air energy storage unit; Obtain the response - time parameters of each energy storage type, and generate an energy storage characteristic parameter table based on the charge - discharge efficiency parameters, the capacity parameters, and the response - time parameters.

3. The method according to claim 1, characterized in that, The obtaining the load data of the power system, analyzing the load data to obtain the peak - valley characteristics of different load types, and generating a load characteristic matrix, where the load types include industrial load, commercial load, and residential load, includes: Obtain the three - phase active power values of the distribution transformer according to the sampling - moment identifier, classify and mark the three - phase active power values using the load - type identification code, and generate an original load data record table; Based on the original load data record table, use the maximum - minimum value detection algorithm to identify the peak points and valley points in the daily load curve, and generate a peak - valley characteristic data table based on the load - type identification code; Based on the peak - valley characteristic data table, obtain the peak - valley differences of each load type, and perform normalization processing on the peak - valley differences through the maximum - minimum value normalization method to generate a normalized peak - valley difference data table; Based on the moment identifiers of the peak points and valley points in the peak - valley characteristic data table, calculate the peak - duration period and valley - duration period, and generate a peak - valley duration data table for each load type. Calculate the power change amount based on the three-phase active power values at adjacent moments, obtain the volatility data through the ratio of the power change amount to the average power value, and classify and label the volatility data through the load type identification code to generate a volatility data table; Construct a feature vector based on the standardized peak-valley difference data table, the peak-valley duration data table, and the volatility data table, and use the principal component analysis method to extract the load characteristics to generate a load characteristic matrix.

4. The method according to claim 1, wherein According to the load characteristic matrix, combined with the energy storage characteristic parameter table, use the rule engine algorithm to match the energy storage type and the load characteristics to generate an energy storage combination plan, including: Establish a capacity adaptability calculation function, an efficiency adaptability calculation function, and a response adaptability calculation function, and use a deep neural network to train an adaptability prediction model based on the energy storage characteristic parameter table; Input the load characteristic matrix into the adaptability prediction model to obtain the energy storage adaptability values of each load type, and the energy storage adaptability values include capacity adaptability values, efficiency adaptability values, and response adaptability values; Construct an energy storage type discriminator, use the energy storage adaptability values of each load type as input features, classify and evaluate the electrochemical energy storage, the pumped-storage energy storage, and the compressed air energy storage to generate an energy storage type adaptability evaluation table; Based on the energy storage type adaptability evaluation table, generate an energy storage capacity allocation ratio threshold, and generate an electrochemical energy storage configuration capacity value, a pumped-storage energy storage configuration capacity value, and a compressed air energy storage configuration capacity value according to the energy storage capacity allocation ratio threshold to form an energy storage combination plan.

5. The method according to claim 1, characterized in that Based on the Monte Carlo simulation model, combined with historical load data, energy storage technical and economic parameters, and grid operation constraint conditions, simulate and optimize the energy storage combination plan, including: Obtain the historical load power data of the distribution transformer through the load collector, and generate a daily load power curve data table according to the historical load power data at the sampling time interval; Train a load sequence predictor using a deep neural network according to the daily load power curve data table to obtain a load prediction curve data table; Mark the peak period and the valley period based on the load prediction curve data table, and use the Monte Carlo method to generate a set of energy storage charge and discharge sequences; Based on the grid operation constraint conditions, verify the constraint conditions for each energy storage charge and discharge sequence to generate a set of effective charge and discharge sequences; Based on the set of charge and discharge sequences and the energy storage technical and economic parameters, use the particle swarm algorithm to construct a multi-objective optimizer, and optimize the energy storage combination plan with the goal of minimizing the investment cost and maximizing the peak shaving and valley filling amount.

6. The method according to claim 5, wherein Based on the optimized energy storage combination plan, calculate the peak shaving amount, the valley filling amount, and the load smoothness, including: Obtain the configuration ratio of each energy storage type from the optimized energy storage combination plan; According to the peak-valley difference threshold of the load curve, perform time period division within a preset time period to obtain a peak interval and a valley interval; In the peak interval, calculate the energy storage discharge power, and in the valley interval, calculate the energy storage charge power. Based on the energy storage discharge power and the energy storage charge power, generate an energy storage power regulation data table; Calculate the peak shaving amount during peak hours and the valley filling amount during off-peak hours according to the energy storage power regulation data table; Based on the load sequence predictor, obtain the load prediction curve data table, calculate the load mean square error value through the load prediction curve data, and generate the load smoothness.

7. The method according to claim 1, characterized in that The execution of the compensation mechanism includes: Use a deep neural network to construct a power fluctuation predictor to extract the fluctuation characteristics of the peak shaving amount and the valley filling amount, and generate a power fluctuation characteristic table; Based on the power fluctuation characteristic table, extract the fluctuation period value and the fluctuation amplitude, optimize the boundaries of the charging period and the discharging period, and generate a period optimization data table; Based on the historical load data, obtain the upper and lower limit intervals of the load rate, determine the load regulation interval, and generate a charging power correction value and a discharging power correction value based on the power fluctuation characteristic table; Based on the charging power correction value and the discharging power correction value, combined with the period optimization data table, perform charging power adjustment and discharging power adjustment.

8. The method according to claim 1, characterized in that, The execution of the dynamic optimization mechanism includes: Based on the target load smoothness threshold, generate a charge and discharge power sequence, use a deep neural network to construct a charge and discharge period divider to segment the charge and discharge power sequence, and generate an energy storage scheduling period table; Extract the power generation prediction value from the photovoltaic power curve and the wind power curve, and establish a power balance constraint function through the power generation prediction value; Based on the power balance constraint function, use a reinforcement learning algorithm to construct an energy storage capacity optimizer to dynamically adjust the configuration of each energy storage type.

9. A comprehensive planning system for a distribution network with multiple new energy accesses, characterized in that, It includes: The first processing module is used to obtain the historical operation data set of the distribution network multi-energy storage system, extract the energy storage characteristics of different energy storage types respectively, and generate an energy storage characteristic parameter table, where the energy storage types include electrochemical energy storage, pumped storage energy storage, and compressed air energy storage; The second processing module is used to obtain the load data of the power system, analyze the load data to obtain the peak-valley characteristics of different load types, and generate a load characteristic matrix, where the load types include industrial load, commercial load, and residential load; The third processing module is used to match the energy storage type with the load characteristics by using the rule engine algorithm according to the load characteristic matrix and in combination with the energy storage characteristic parameter table, and generate an energy storage combination plan; The fourth processing module is used to simulate and optimize the energy storage combination plan based on the Monte Carlo simulation model, in combination with historical load data, energy storage technical and economic parameters, and grid operation constraint conditions; The fifth processing module is used to calculate the peak shaving amount, the valley filling amount, and the load smoothness based on the optimized energy storage combination plan; The sixth processing module is used to execute the compensation mechanism if one of the peak shaving amount and the valley filling amount is lower than the target peak shaving and valley filling threshold; The seventh processing module is used to execute the dynamic optimization mechanism if the load smoothness is lower than the preset target load smoothness threshold.

10. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method according to any one of claims 1 to 8.

Images