A distributed comprehensive optimization method for new energy access and dispatch

Abstract

The present application relates to a kind of distributed new energy access and scheduling comprehensive optimization method, belong to distribution network technical field;The method comprises: S1: obtain containing mass distributed energy monitoring system historical data, obtain regional historical load data, obtain regional historical environmental governance cost, power generation cost, reliability and toughness weight index;S2: utilize historical data initialization different types distributed energy generation proportion and margin;S3: utilize S1 different types distributed energy monitoring system historical data to predict future 24 hours power generation power data;S4: utilize predicted 24 hours power generation power data, historical environmental governance cost, power generation cost, reliability and toughness weight index and distributed energy generation proportion and margin, construct distributed energy environmental governance cost, power generation cost, reliability and toughness multi-objective and comprehensive cost single target optimal scheduling model and constraint condition;S5: with long-term operation optimal as target, relevant data is substituted into multi-objective and single target scheduling model, by solving, to obtain different types distributed energy long-term generation proportion and power generation margin, simultaneously generate short-term power generation trend constraint factor;S6: with short-term operation optimal as objective function, combine abnormal disturbance factor and trend constraint factor, solve short-time multi-objective and single target optimal function, to obtain the generation proportion and power generation margin of different distributed energy in short time;The substantial significance of the present application is: under the premise of guaranteeing distributed energy normal stable output and reasonable power generation margin, through the comprehensive optimization scheduling of different distributed energy, realize overall new energy reliable complementary optimization operation and comprehensive economic benefit maximization.

Description

Technical Field

[0001] This invention belongs to the field of microgrid access scheduling and management technology, and relates to a comprehensive optimization method for distributed new energy access and scheduling. Background Technology

[0002] With the rapid development of distributed energy sources such as wind power, photovoltaics, and modern bioenergy, their proportion in distribution networks is gradually increasing. For distribution networks with high penetration rates of new energy sources, traditional distribution network dispatching methods are becoming increasingly inadequate due to the randomness, volatility, and intermittency of wind and photovoltaic power generation. Therefore, it is urgent to study the safe and economical dispatching of distribution networks under the background of high new energy penetration. The purpose of this invention is to provide a comprehensive optimization method for the access and dispatching of distributed new energy sources, achieving optimal dispatching of distributed energy generation in terms of environmental governance costs, power generation costs, reliability, and resilience.

[0003] In the research on optimal dispatching of high-penetration distribution networks, existing technologies consider uncertain factors in micro-source output, prediction errors in wind and solar power output, and uncertainties in wind and solar power output prediction for optimal circuit models, establishing various models that better reflect real-world factors. Regarding the factors considered in optimal dispatching models, economic cost is usually the most important indicator in microgrid optimal operation. With the increasing prevalence of microgrid applications and the national advocacy for energy conservation and emission reduction, more and more literature is beginning to consider environmental objectives. One approach is to transform multi-objective models into single-objective models through collaboration, or to use multi-objective models for modeling. In summary, existing optimal dispatching research often uses minimizing the sum of generation costs and environmental governance costs as the optimal function, with reliability and safety margin as constraints, without comprehensively considering resilience and the impact of anomalies. Introducing reliability and resilience indicators can effectively address optimal dispatching under abnormal conditions while achieving reasonable configuration strategies for distributed power sources. Summary of the Invention

[0004] The purpose of this invention is to provide a comprehensive optimization method for distributed renewable energy access and scheduling, so as to achieve optimal scheduling of distributed energy generation in terms of environmental governance costs, power generation costs, reliability and resilience.

[0005] The first aspect of this application provides a comprehensive optimization method for distributed new energy access and scheduling, the method comprising:

[0006] S1: Obtain historical data from a massive distributed energy monitoring system, historical load data for the region, historical environmental governance costs, power generation costs, and reliability and resilience weighting indicators for the region.

[0007] S2: Initialize the generation ratio and margin of different types of distributed energy using historical data;

[0008] S3: Predict power generation data for the next 24 hours using historical data from different types of distributed energy monitoring systems in S1;

[0009] S4: Using the predicted 24-hour power generation data, historical environmental governance costs, power generation costs, reliability and resilience weight indicators, and the proportion and margin of distributed energy generation, construct a multi-objective and comprehensive cost single-objective optimal scheduling model and constraints for distributed energy environmental governance costs, power generation costs, reliability and resilience.

[0010] S5: With the goal of long-term optimal operation, relevant data are substituted into multi-objective and single-objective scheduling models. By solving the model, the long-term power generation ratio and power generation margin of different types of distributed energy are obtained, and trend constraint factors for short-term power generation are generated.

[0011] S6: Using short-term optimal operation as the objective function, and combining abnormal disturbance factors and trend constraint factors, solve the short-term multi-objective and single-objective optimal functions to obtain the power generation ratio and power generation margin of different energy sources in the short term.

[0012] In one possible implementation of the first aspect, the system acquires historical data from a massive distributed energy monitoring system, historical regional load data, historical regional environmental governance costs, power generation costs, and reliability and resilience weighting indicators, specifically:

[0013] S11: For distributed energy, the distributed energy monitoring system data from the past 3 years, specifically the 10 days before and after that date, is used as the long-term reference dataset D. i,long Using the monitoring system data from the five days prior to that date as the recent reference dataset D i,short , where i represents the photovoltaic power generation dataset D in different types of distributed energy sources. PV,t Wind power generation dataset D WV,t Energy storage battery dataset D ESB,t Electric vehicle dataset D EV,t Gas turbine dataset D MT,t Where t = long represents 10 days before and after the date in the past 3 years, t = short represents 5 days before the date, and t = day represents the next 24 hours. The monitoring system data are different for different distributed energy sources.

[0014] S12: Use the daily electricity load data for the 10 days before and after this date over the past 3 years as a reference. long Using the load data from the previous 5 days as a reference L short ;

[0015] S13: Perform weighted fusion processing based on the long-term and recent data in S11:

[0016] D i，new=θ i，long D i，long +θ i，short D i，short (1)

[0017] The load is also weighted accordingly:

[0018] L new =λ long L long +λ short L short (2)

[0019] In equation (1) θ i，j (j = long, short) represents the weighting of historical monitoring data of different distributed energy sources in different historical periods; Equation (2) λ j (j = long, short) represents the weighting of load data for different historical periods;

[0020] S14: Statistically analyze the environmental governance costs, power generation costs, reliability, and resilience of different distributed energy sources at different time periods, and calculate the initial weight w0 = [w c,t,i w cos,t,i w rel,t,i w adj,t,i ], t can represent long execution, short execution, and day execution for the next 24 hours, that is, t = [long, short, day]; w c,t,i This represents the weight of governance costs for different distributed energy operating environments at different times; w cos,t,i This indicates the weight of different distributed energy operating costs at different times; w rel,t,i The weights representing the operational reliability of different distributed energy sources at different times; w adj,t,i The weights represent the operational resilience of different distributed energy sources at different times.

[0021] In one possible implementation of the first aspect, historical data is used to initialize the generation ratio and margin of different types of distributed energy resources, specifically as follows:

[0022] S21: Calculate the initial power generation ratio R for different distributed energy sources at different time periods according to the relevant data in claim 1. i，t =[R PV，0 R WV，0 R ESB，0 R EV，0 R MT，0 ], i represents different distributed energy sources, t=0 represents the initial value, t=long represents long-term operation, t=short represents short-term operation, R PV,0 R represents the initial proportion of photovoltaic power generation. WV,0 R represents the initial proportion of wind power generation.ESB,0 R represents the initial percentage of power output from the energy storage battery. EV,0 R represents the initial percentage of power output from electric vehicles. MT,0 This indicates the initial percentage of power output from the gas turbine;

[0023] S22: Determine the initial power generation margin F for different energy sources and different time periods based on seasonality and diurnal rhythms. i，t =[F PV，0 F WV，0 F ESB，0 F EV，0 F MT，0 ], i represents different distributed energy sources; t = 0 represents the initial value, t = long represents long-term operation, t = short represents short-term operation, F PV,0 F represents the initial power generation margin of photovoltaic power generation. WV,0 F represents the initial power generation margin of wind power generation. ESB,0 F represents the initial power generation margin of the energy storage battery. EV,0 F represents the initial power generation margin of an electric vehicle. MT,0 This indicates the initial power generation margin of the gas turbine output.

[0024] In one possible implementation of the first aspect, the power generation data for the next 24 hours is predicted as follows:

[0025] S31: Perform preprocessing and normalization on the historical data of different types of distributed energy monitoring systems obtained in S13;

[0026] S32: The data in S31 are analyzed using kernel canonical correlation analysis, with non-output power data from the monitoring system data used as the sample matrix X. im =[x i,1 x i,2 …x i,m The output power is represented by matrix P. im =[p i,1 p i,2 …p i,m ], where m is the number of different types of distributed energy sources;

[0027] S33: Divide the data into training sample set D 训练 (x i,j ,p i,j ) and test sample set D 测试 (x i,j ,p i,j );

[0028] S34: Train a lightweight gradient boosting tree model using training samples;

[0029] S35: Predict power P for the next 24 hours using the test sample set and the trained model. i,day ;

[0030] S36: Based on the predicted P i,day Dispatch different types of distributed energy sources to generate electricity;

[0031] In one possible implementation of the first aspect, a multi-objective and comprehensive cost-single-objective optimal scheduling model and constraints for distributed energy environmental governance costs, power generation costs, reliability, and resilience are constructed, specifically as follows:

[0032] S41: Establish a comprehensive optimal scheduling model for distributed energy resources, considering environmental governance costs, power generation costs, reliability, and resilience.

[0033] Power balance formula:

[0034]

[0035] In equation (3), P i P represents the daily power generation capacity of different types of distributed energy sources. i,day,max R represents the maximum daily output power of different distributed energy sources, k is the total number of different distributed energy sources included, and R i,t F represents the proportion of power generation from different distributed energy sources at different times. i,t This represents the power generation margin for different energy sources and different time periods. t=0 represents the initial value, t=long represents long-term operation, t=short represents short-term operation, and δ is the system resilience threshold value.

[0036] Constructing a multi-objective function:

[0037] Objective function for environmental governance cost indicators:

[0038] Electricity generation cost objective function:

[0039] Reliability objective function:

[0040] Resilience objective function:

[0041] The objective function above can be expressed as follows:

[0042]

[0043] Single-objective function: The above multi-objective function is converted into a comprehensive cost single-objective function. Since environmental governance costs and power generation costs have the same dimensions, no conversion is needed. Reliability and resilience are converted into cost dimensions, and the corresponding weights become w. relfc，t，i and w adjfc，t，iTherefore, environmental governance costs, operating costs, reliability, and resilience are all converted into comprehensive cost conditions, and the single objective function is:

[0044]

[0045] In equation (5), w relfc，t，i As a weight for converting reliability into cost, w adjfc，t，i Weighting resilience in relation to cost;

[0046] S42: Constraints for establishing the comprehensive optimal scheduling model:

[0047] P i，day min <P i，day <P i，day max

[0048]

[0049] In one possible implementation of the first aspect, with the goal of long-term optimal operation, relevant data is substituted into the model to determine the generation ratio of different types of distributed energy resources, while simultaneously generating constraint factors for short-term generation. The specific steps are as follows:

[0050] S51: Taking the optimization of long-term power generation as the objective, substitute the long-term data into formulas (3) and (4):

[0051] Power balance formula:

[0052]

[0053] Multi-objective function:

[0054]

[0055] In equation (8), w c,long,i w represents the weight of long-term operating environment governance costs. cost,long,i w represents the weight of long-term operating costs. rel,long,i w represents the long-term operational reliability weight. adj,long,i Indicates the weighting for long-term operational resilience;

[0056] S52: The long-term multi-objective optimal function for power generation is solved using the alternating direction multiplier method to obtain the optimal scheduling strategy for each distributed energy source. For multi-objective optimal scheduling, the optimal strategy is:

[0057] W m，OPT =[R m，i，opt F m，i，opt γ m，i ];

[0058] Among them, R m,i,opt For the long-term optimal proportion of multiple objectives, Fm,i,opt For multi-objective long-term optimal margin, γ m，i This is a constraint factor for the long-term optimal solution of short-term power generation.

[0059] S53: Taking the optimal single objective of long-term power generation as the objective function, substitute the long-term data into formulas (3) and (4);

[0060] Single-objective function: By converting environmental governance costs, operating costs, reliability, and resilience into comprehensive cost conditions, the multi-objective problem becomes a single objective.

[0061]

[0062] In equation (9), w c，long，i w represents the weight of long-term operating environment governance costs. cost，long，i w represents the weight of long-term operating costs. relfc，long，i w represents the long-term operational reliability weight. adj，long，i Indicates the weighting for long-term operational resilience;

[0063] S54: The alternating direction multiplier method is used to solve the long-term generation single-objective optimal function, obtaining the optimal scheduling strategy for each distributed energy source. For the comprehensive cost single-objective optimal scheduling, the optimal strategy is W. s，OPT =[R si，opt， F si，opt， γ s，i ];

[0064] Among them, R s，i，opt For the long-term optimal proportion of a single objective, F m，i，opt For the long-term optimal margin of a single objective, γ s，i This is a constraint factor for finding the optimal solution for short-term power generation over a long-term problem.

[0065] In one possible implementation of the first aspect, short-term operational optimization is used as the objective function. By combining anomaly disturbance factors and trend constraint factors, short-term multi-objective and single-objective optimal functions are solved to obtain the short-term power generation ratio and power generation margin of different energy sources. Specifically:

[0066] S61: Determine the anomalous disturbance factor ζ m，i or ζ s，i Under normal circumstances, the abnormal disturbance factor is 1. In abnormal situations, the disturbance factor is updated in real time. The abnormal disturbance factor is mainly determined by experts.

[0067] or

[0068] S62: Combining the abnormal disturbance factor and the trend constraint factor, determine the short-term power generation optimization as the objective function;

[0069] Power balance formula:

[0070]

[0071] Multi-objective function:

[0072]

[0073] Single objective function:

[0074]

[0075] In equation (13), w c,short,i w represents the weight of short-term operating environment governance costs. cost,short,i w represents the weight of short-term operating costs. relfc,short,i w represents the weight by which short-term operational reliability is converted into economic cost. adjfc,short,i The weight representing the conversion of short-term operational resilience into cost;

[0076] S63: Using the alternating direction multiplier method to solve for the optimal functions of short-term power generation multi-objective and single-objective functions, the optimal scheduling strategy W for each distributed energy source is obtained. m，short，opt =[R m，i，short，opt F m，i，short，opt For a single-objective optimal scheduling with comprehensive cost, the optimal strategy is W. s，short，opt =[R s，i，short，opt F s，i，short，opt ];

[0077] S64: Based on the short-term multi-objective and single-objective optimal scheduling strategies calculated in S63, the corresponding scheduling is carried out in combination with the actual application scenario; multi-objective can be used in specific situations, while single-objective has the lowest comprehensive cost and can be used as the daily scheduling strategy.

[0078] The beneficial effects of this invention are as follows: This invention adopts a comprehensive optimization method for distributed renewable energy access and scheduling, incorporating reliability and resilience factors into the optimization function and including power generation margin as part of the optimization strategy. This not only enables optimized scheduling under normal conditions but also achieves optimized scheduling under abnormal conditions. The algorithm first uses historical data from the distributed energy monitoring system to statistically analyze the power generation, power generation ratio, power generation margin, environmental governance costs, operating costs, operational reliability, and system resilience indicators of different types of distributed energy. It then predicts the power generation and load data for the next 24 hours. Next, it establishes a system model and constraints. Finally, based on demand, it can generate optimal operating schemes for different modes, including long-term multi-objective comprehensive optimization, short-term multi-objective optimization, long-term single-objective optimization, and short-term single-objective optimization. The introduction of a long-term trend influence factor on the short-term makes short-term predictions more reasonable. Furthermore, the design of an anomaly disturbance factor ensures the optimal operating mode to the greatest extent possible in the event of abnormal situations. Therefore, the method proposed in this invention can further improve the timely and on-demand optimal operation of the system.

[0079] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0080] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein:

[0081] Figure 1 This is a block diagram of a comprehensive optimization method for distributed new energy access and scheduling provided in an embodiment of the present invention;

[0082] Figure 2 This is a flowchart illustrating the process of obtaining historical monitoring data of distributed energy sources and related indicators such as power generation ratio and power generation margin, according to an embodiment of the present invention.

[0083] Figure 3 This is a flowchart of a method for predicting the power generation of different types of distributed energy sources in the next 24 hours, provided by an embodiment of the present invention.

[0084] Figure 4 This is an implementation process for optimal scheduling of distributed energy access provided by an embodiment of the present invention. Detailed Implementation

[0085] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0086] Please see Figures 1-4 . Figure 1 This is a block diagram of a comprehensive optimization method for distributed new energy access and scheduling provided by an embodiment of the present invention, as shown below. Figure 1 As shown, a preferred comprehensive optimization method for distributed renewable energy access and scheduling according to the present invention includes the following steps:

[0087] S1: Obtain historical data from a massive distributed energy monitoring system, historical load data for the region, historical environmental governance costs, power generation costs, and reliability and resilience weighting indicators for the region.

[0088] S2: Initialize the generation ratio and margin of different types of distributed energy using historical data;

[0089] S3: Predict power generation data for the next 24 hours using historical data from different types of distributed energy monitoring systems in S1;

[0090] S4: Using the predicted 24-hour power generation data, historical environmental governance costs, power generation costs, reliability and resilience weight indicators, and the proportion and margin of distributed energy generation, construct a multi-objective and comprehensive cost single-objective optimal scheduling model and constraints for distributed energy environmental governance costs, power generation costs, reliability and resilience.

[0091] S5: With the goal of long-term optimal operation, relevant data are substituted into multi-objective and single-objective scheduling models. By solving the model, the long-term power generation ratio and power generation margin of different types of distributed energy are obtained, and trend constraint factors for short-term power generation are generated.

[0092] S6: Using short-term optimal operation as the objective function, and combining abnormal disturbance factors and trend constraint factors, solve the short-term multi-objective and single-objective optimal functions to obtain the power generation ratio and power generation margin of different energy sources in the short term.

[0093] Figure 2 This is a flowchart illustrating the process of obtaining historical monitoring data and related indicators such as power generation ratio and power generation margin for distributed energy resources, as provided in an embodiment of the present invention. Figure 2 Therefore, the specific steps include:

[0094] S1: Obtain historical data of energy monitoring systems in areas containing massive distributed energy resources, obtain historical load data of the area, and obtain historical environmental governance costs, power generation costs, reliability and resilience weight indicators of the area.

[0095] S11: For distributed energy, the distributed energy monitoring system data from the past 3 years, specifically the 10 days before and after that date, is used as the long-term reference dataset D. i,long Using the monitoring system data from the five days prior to that date as the recent reference dataset D i,short , where i represents the photovoltaic power generation dataset D in different types of distributed energy sources. PV,t Wind power generation dataset D WV,t Energy storage battery dataset D ESB,t Electric vehicle dataset D EV,t Gas turbine dataset D MT,t Where t = long represents 10 days before and after the date in the past 3 years, t = short represents 5 days before the date, and t = day represents the next 24 hours. The monitoring system data will be different for different distributed energy sources.

[0096] S12: Use the daily electricity load data for the 10 days before and after this date over the past 3 years as a reference. long Using the load data from the previous 5 days as a reference L short ;

[0097] S13: Perform weighted fusion processing based on the long-term and recent data in S11:

[0098] D i，new =θ i，long D i，long θ i，short D i，short (1)

[0099] The load is also weighted accordingly:

[0100] L new =λ long L long +λ short L short (2)

[0101] In equation (1) θ i，j (j = long, short) represents the weighting of historical monitoring data of different distributed energy sources in different historical periods; Equation (2) λ j (j = long, short) represents the weighting of load data for different historical periods;

[0102] S14: Statistically analyze the environmental governance costs, power generation costs, reliability, and resilience of different distributed energy sources at different time periods, and calculate the initial weight w0 = [w c,t,i w cos,t,i w rel,t,i w adj,t,i ], t can represent long execution, short execution, and day execution for the next 24 hours, that is, t = [long, short, day]; w c,t,i This represents the weight of governance costs for different distributed energy operating environments at different times; w cos,t,i This indicates the weight of different distributed energy operating costs at different times; w rel,t,i The weights representing the operational reliability of different distributed energy sources at different times; w adj,t,i The weights represent the operational resilience of different distributed energy sources at different times.

[0103] Figure 3 This is a flowchart of a method for predicting the power generation of different types of distributed energy sources in the next 24 hours, provided by an embodiment of the present invention. Figure 3 Therefore, the implementation steps include:

[0104] S31: Perform preprocessing and normalization on the fused datasets of different types of distributed energy monitoring systems obtained in step S13, including both long-term and short-term datasets.

[0105] S32: The data in S31 are analyzed using kernel canonical correlation analysis, with non-output power data from the monitoring system data used as the sample matrix X. im =[x i,1 x i,2 …x i,m The output power is represented by matrix P. im =[p i,1 p i,2 …p i,m ];

[0106] S33: The data in S32 are analyzed using kernel canonical correlation analysis, and the kernel function K is taken.

[0107]

[0108] In formula (3) and It is a transformation applied to the non-output power dataset and the power dataset, finding two vectors a and b in the transformed space such that and The correlation coefficient is the largest.

[0109] The objective function for the nuclear canonical correlation analysis problem is:

[0110]

[0111] In formula (4) Canonical correlation analysis of α and β kernels is transformed into an input eigenvalue problem:

[0112] P -1 MN -1 M T α=μ 2 α,N -1 M T P -1 Mβ=μ 2 β (5)

[0113] Where, μ = μ x =μ y Using α and β obtained from equation (5), the typical variables of X and Y can be determined as follows: Predict the output power Y based on the first k canonical correlation vectors;

[0114] S33: Group the canonical variables u and v obtained in S32 into training sample sets D. 训练 (u i,j ,v i,j ) and test sample set D 测试 (u i,j ,v i,j );

[0115] S34: Train a lightweight gradient boosting tree model using training samples. The specific steps are as follows:

[0116] (1) Construct a boosting tree model;

[0117] (2) Input training data D 训练 (u i,j ,v i,j );

[0118] (3) Adjustment of relevant parameters;

[0119] (4) Lightweight gradient boosting tree models are easy to train;

[0120] S35: Predict power P for the next 24 hours using the test sample set and the trained model. i,day .

[0121] Figure 4 This is an implementation process for optimal scheduling of distributed energy access provided by an embodiment of the present invention, consisting of... Figure 4 Therefore, its implementation steps include establishing the model and constraints, solving for the optimal operation in different time periods and modes, and the specific implementation process includes:

[0122] According to claim 4, by utilizing predicted 24-hour power generation data, historical environmental governance costs, power generation costs, reliability and resilience weight indicators, and the proportion and margin of distributed energy generation, a multi-objective and comprehensive cost single-objective optimal scheduling model and constraints for distributed energy environmental governance costs, power generation costs, reliability and resilience are constructed. Step S4 specifically includes the following steps:

[0123] S41: Establish a comprehensive optimal scheduling model for distributed energy resources, considering environmental governance costs, (generation) costs, reliability, and resilience.

[0124] Power balance formula:

[0125]

[0126] In the formula P i P represents the daily power generation capacity of different types of distributed energy sources. i,day,max R represents the maximum daily output power of different distributed energy sources, k is the total number of different distributed energy sources included, and R i,t F represents the proportion of power generation from different distributed energy sources at different times. i,t This represents the power generation margin for different energy sources and different time periods. t=0 represents the initial value, t=long represents long-term operation, t=short represents short-term operation, and δ is the system resilience threshold.

[0127] Constructing a multi-objective function:

[0128] Objective function for environmental governance cost indicators:

[0129] Electricity generation cost objective function:

[0130] Reliability objective function:

[0131] Resilience objective function:

[0132] The objective function above can be expressed as follows:

[0133]

[0134] Single-objective function: The above multi-objective function is converted into a comprehensive cost single-objective function. Since environmental governance costs and power generation costs have the same dimensions, no conversion is needed. Reliability and resilience are converted into cost dimensions, and the corresponding weights become w. relfc,t,i and w adjfc,t,i Therefore, environmental governance costs, operating costs, reliability, and resilience are all converted into comprehensive cost conditions, and the single objective function is:

[0135]

[0136] In the formula w relfc，t，i As a weight for converting reliability into cost, w adjfc，t，i Weighting resilience in relation to cost;

[0137] S42: Constraints for establishing the comprehensive optimal scheduling model:

[0138] P i，dda min <P i，day <P i，day max

[0139]

[0140] Figure 4 With long-term operational optimization as the goal, relevant data are substituted into the model to determine the proportion of power generation for different types of distributed energy sources, while simultaneously generating constraint factors for short-term power generation. The specific steps are as follows:

[0141] S51: Taking the optimization of long-term power generation as the objective, substitute long-term data into the formulas for multi-objective and single-objective functions:

[0142] Power balance formula:

[0143]

[0144] Multi-objective function:

[0145]

[0146] In the formula w c,long,i w represents the weight of long-term operating environment governance costs. cost,long,i w represents the weight of long-term operating costs. rel,long,i w represents the long-term operational reliability weight. adj,long,i This represents the weighting for long-term operational resilience.

[0147] S52: The long-term multi-objective optimal function for power generation is solved using the alternating direction multiplier method to obtain the optimal scheduling strategy for each distributed energy source. For multi-objective optimal scheduling, the optimal strategy is:

[0148] W m，OPT =[R m，i，opt F m，i，opt γ m，i ].

[0149] R m,i,opt For the long-term optimal proportion of multiple objectives, F m,i,opt For multi-objective long-term optimal margin, γ m，i This is a constraint factor for the long-term optimal solution of short-term power generation.

[0150] S53: Using the optimal single objective of long-term power generation as the objective function, substitute long-term data into the formula for multi-objective and single-objective functions.

[0151] Single-objective function: By converting environmental governance costs, operating costs, reliability, and resilience into comprehensive cost conditions, the multi-objective problem becomes a single objective.

[0152]

[0153] In the formula w c，long，i w represents the weight of long-term operating environment governance costs. cost，long，i w represents the weighting of long-term operating power generation costs. relfc，long，i w represents the weight that converts long-term operational reliability into economic cost. adj，long，i This represents the weight by which long-term operational resilience is converted into cost.

[0154] S54: The alternating direction multiplier method is used to solve the long-term generation single-objective optimal function, obtaining the optimal scheduling strategy for each distributed energy source. For the comprehensive cost single-objective optimal scheduling, the optimal strategy is W. s，OPT =[R si，opt F s，i，opt，γs，i ].

[0155] R s，i，opt For the long-term optimal proportion of a single objective, F m，i，opt For the long-term optimal margin of a single objective, γ s，i This is a constraint factor for finding the optimal solution for short-term power generation over a long-term problem.

[0156] Furthermore, using short-term optimal power generation as the objective function, and combining anomaly disturbance factors and trend constraint factors, the short-term optimal function is solved to obtain the power generation ratio and margin of different energy sources in the short term. The specific steps are as follows:

[0157] S61: Determine the anomalous disturbance factor ζ m，i or ζ s，i Under normal circumstances, the abnormal disturbance factor is 1. In abnormal situations, the disturbance factor is updated in real time. The abnormal disturbance factor is mainly determined by experts.

[0158] or

[0159] S62: Combining the abnormal disturbance factor and the trend constraint factor, determine the short-term power generation optimization as the objective function;

[0160] Power balance formula:

[0161]

[0162] Multi-objective function:

[0163]

[0164] Single objective function:

[0165]

[0166] In the formula w c,short,i w represents the weight of short-term operating environment governance costs. cost,short,i w represents the weight of short-term operating costs. relfc,short,i w represents the weight by which short-term operational reliability is converted into economic cost. adjfc,short,i This represents the weight by which short-term operational resilience is converted into cost.

[0167] S63: Using the alternating direction multiplier method to solve for the optimal functions of short-term power generation multi-objective and single-objective functions, the optimal scheduling strategy W for each distributed energy source is obtained. m，short，opt =[R m，i，short，opt F m，i，short，opt For a single-objective optimal scheduling with comprehensive cost, the optimal strategy is W. s，short，opt =[R s，i，short，opt F s，i，short，opt ].

[0168] S64: Based on the short-term multi-objective and single-objective optimal scheduling strategies calculated in S63, the corresponding scheduling is performed in conjunction with the actual application scenario. Multi-objective scheduling can be used in specific situations, while single-objective scheduling, which has the lowest overall cost, can be used as the daily scheduling strategy.

[0169] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A comprehensive optimization method for distributed renewable energy access and scheduling, characterized in that, The method includes: S1: Obtain historical data from a massive distributed energy monitoring system, historical load data for the region, historical environmental governance costs, power generation costs, and reliability and resilience weighting indicators for the region. S2: Initialize the generation ratio and margin of different types of distributed energy using historical data; S3: Predict power generation data for the next 24 hours using historical data from different types of distributed energy monitoring systems in S1; S4: Using the predicted 24-hour power generation data, historical environmental governance costs, power generation costs, reliability and resilience weight indicators, and the proportion and margin of distributed energy generation, construct a multi-objective and comprehensive cost single-objective optimal scheduling model and constraints for distributed energy environmental governance costs, power generation costs, reliability and resilience. S5: With the goal of long-term optimal operation, relevant data are substituted into multi-objective and single-objective scheduling models. By solving the model, the long-term power generation ratio and power generation margin of different types of distributed energy are obtained, and trend constraint factors for short-term power generation are generated. S6: Using short-term optimal operation as the objective function, and combining abnormal disturbance factors and trend constraint factors, solve the short-term multi-objective and single-objective optimal functions to obtain the power generation ratio and power generation margin of different energy sources in the short term. Specifically, the construction of the multi-objective and comprehensive cost-single-objective optimal scheduling model and constraints for distributed energy environmental governance costs, power generation costs, reliability, and resilience are as follows: S41: Establish a comprehensive optimal scheduling model for distributed energy resources, considering environmental governance costs, power generation costs, reliability, and resilience. Power balance formula: （3） In equation (3), P i P represents the daily power generation capacity of different types of distributed energy sources. i,day,max This indicates the maximum daily output power of different distributed energy sources. k It is the total number of different distributed energy sources contained therein, R i,t F represents the proportion of power generation from different distributed energy sources at different times. i,t This indicates the power generation margin for different energy sources and at different times. t =0 represents the initial value, t= long Indicates long-term operation, t= short Indicates short-term operation. It is the system resilience threshold; Constructing a multi-objective function: Objective function for environmental governance cost indicators: ; Electricity generation cost objective function: ; Reliability objective function: ; Resilience objective function: ; The objective function above can be expressed as follows: （4） Single-objective function: The above multi-objective function is converted into a comprehensive cost single-objective function. Since environmental governance costs and power generation costs have the same dimensions, no conversion is needed. Reliability and resilience are converted into cost dimensions, and the corresponding weights become w. relfc,t,i and w adjfc,t,i Therefore, environmental governance costs, operating costs, reliability, and resilience are all converted into comprehensive cost conditions, and the single objective function is: （5） In equation (5), w relfc,t,i As a weight for converting reliability into cost, w adjfc,t,i Weighting resilience in relation to cost; S42: Constraints for establishing the comprehensive optimal scheduling model: （6） Among them, w c,t,i This represents the weight of governance costs for different distributed energy operating environments at different times; w cos,t,i This indicates the weight of different distributed energy operating costs at different times; w rel,t,i The weights representing the operational reliability of different distributed energy sources at different times; w adj,t,i The weights represent the operational resilience of different distributed energy sources at different times.

2. The comprehensive optimization method for distributed new energy access and scheduling according to claim 1, characterized in that, The acquisition of historical data from a massive distributed energy monitoring system, historical regional load data, historical regional environmental governance costs, power generation costs, and reliability and resilience weighting indicators specifically includes: S11: For distributed energy, the distributed energy monitoring system data from the 10 days before and after the date of the event over the past 3 years shall be used as the long-term reference dataset. D i,long The monitoring system data from the previous 5 days is used as the recent reference dataset. D i,short , i Data sets representing photovoltaic power generation in different types of distributed energy resources D PV,t Wind power generation dataset D WV, t Energy storage battery dataset D ESB, t Electric vehicle dataset D EV, t Gas turbine dataset D MT, t ,in t=long This indicates that within 10 days before or after this date in the past 3 years, t=short This indicates that 5 days prior to that date, t = day This indicates that the monitoring system data will differ for different distributed energy sources over the next 24 hours. S12: Use the daily electricity load data for the 10 days before and after this date over the past 3 years as a reference. L long The load data from the five days prior to that date will be used as a reference. L short ; S13: Perform weighted fusion processing based on the long-term and recent data in S11: （1） The load is also weighted accordingly: （2） In formula (1) This represents the weighting of historical monitoring data for different distributed energy sources at different historical periods; Equation (2) This indicates a weighted average of load data for different historical periods; S14: Statistically analyze the environmental governance costs, power generation costs, reliability, and resilience of different distributed energy sources at different time periods, and calculate the initial weight w0=[w c,t,i w cos,t,i w rel,t,i w adj,t,i ], t Indicates long-term operation long Short-term operation short Operation in the next 24 hours day ,Right now, t =[ long , short , day ];w c,t,i The weights represent the governance costs of different distributed energy operating environments at different times; w cos,t,i This indicates the weight of different distributed energy operating costs at different times; w rel,t,i The weights representing the operational reliability of different distributed energy sources at different times; w adj,t,i The weights represent the operational resilience of different distributed energy sources at different times.

3. The comprehensive optimization method for distributed new energy access and scheduling according to claim 2, characterized in that, The process of initializing the generation ratio and margin of different types of distributed energy resources using historical data specifically involves: S21: Calculate the initial power generation ratio of different distributed energy sources at different time periods according to the relevant data in claim 1. , i Representing different distributed energy sources, t =0 indicates the initial value. t = long Indicates long-term operation. t = short Indicates short-term operation. R PV,0 Indicates the initial proportion of photovoltaic power generation. R WV,0 This indicates the initial proportion of wind power generation. R ESB,0 This indicates the initial percentage of power output from the energy storage battery. R EV,0 This indicates the initial percentage of power output from electric vehicles. R MT,0 This indicates the initial percentage of power output from the gas turbine; S22: Determine the initial power generation margin for different energy sources and different time periods based on seasonality and diurnal rhythms. , i This represents different distributed energy sources; t =0 indicates the initial value. t = long Indicates long-term operation. t = short Indicates short-term operation. F PV,0 This indicates the initial power generation margin of photovoltaic power generation. F WV,0 This indicates the initial power generation margin of wind power generation. F ESB,0 This indicates the initial power generation margin of the energy storage battery. F EV,0 This indicates the initial power generation margin of the electric vehicle. F MT,0 This indicates the initial power generation margin of the gas turbine output.

4. The comprehensive optimization method for distributed new energy access and scheduling according to claim 3, characterized in that, The predicted power generation data for the next 24 hours is as follows: S31: Perform preprocessing and normalization on the historical data of different types of distributed energy monitoring systems obtained in S13; S32: The data in S31 are analyzed using kernel canonical correlation analysis, with non-output power data from the monitoring system data used as the sample matrix X. im =[x i,1 x i,2 …x i,m The output power is represented by matrix P. im =[p i,1 p i,2 …p i,m ], m It represents the number of different types of distributed energy sources; S33: Divide the data into training sample set D 训练 (x i,j , p i,j ) and test sample set D 测试 (x i,j , p i,j ); S34: Train a lightweight gradient boosting tree model using training samples; S35: Predict power P for the next 24 hours using the test sample set and the trained model. i,day ; S36: Based on the predicted P i,day Dispatch different types of distributed energy sources to generate electricity.

5. The comprehensive optimization method for distributed new energy access and scheduling according to claim 1, characterized in that, The goal is to achieve optimal long-term operation. Relevant data is substituted into the model to determine the power generation ratio of different types of distributed energy sources, while simultaneously generating constraint factors for short-term power generation. The specific steps are as follows: S51: Taking the optimization of long-term power generation as the objective, substitute the long-term data into formulas (3) and (4): Power balance formula: （7） Multi-objective function: （8） In equation (8), w c,long,i w represents the weight of long-term operating environment governance costs. cost,long,i w represents the weight of long-term operating costs. rel,long,i w represents the long-term operational reliability weight. adj,long,i Indicates the weighting for long-term operational resilience; S52: The long-term multi-objective optimal function for power generation is solved using the alternating direction multiplier method to obtain the optimal scheduling strategy for each distributed energy source. For multi-objective optimal scheduling, the optimal strategy is: ； Among them, R m,i,opt For the long-term optimal proportion of multiple objectives, F m,i,opt For multi-objective long-term optimal margin, This is a constraint factor for the long-term optimal solution of short-term power generation. S53: Taking the optimal single objective of long-term power generation as the objective function, substitute the long-term data into formulas (3) and (4). Single-objective function: By converting environmental governance costs, operating costs, reliability, and resilience into comprehensive cost conditions, the multi-objective problem becomes a single objective. （9） In equation (9), w c,long,i w represents the weight of long-term operating environment governance costs. cost,long,i w represents the weight of long-term operating costs. rel,long,i w represents the long-term operational reliability weight. adj,long,i Indicates the weighting for long-term operational resilience; S54: Using the alternating direction multiplier method to solve the long-term generation single-objective optimal function, the optimal scheduling strategy for each distributed energy source is obtained. For the comprehensive cost single-objective optimal scheduling, the optimal strategy is: ; Among them, R s,i,opt For the long-term optimal proportion of a single objective, F m,i,opt This represents the long-term optimal margin for a single objective. This is a constraint factor for finding the optimal solution for short-term power generation over a long-term problem.

6. The comprehensive optimization method for distributed new energy access and scheduling according to claim 5, characterized in that, The method uses short-term operational optimization as the objective function, combined with anomaly disturbance factors and trend constraint factors, to solve for short-term multi-objective and single-objective optimal functions, in order to obtain the short-term power generation ratio and power generation margin of different energy sources, specifically: S61: Determine the anomalous disturbance factor or Under normal circumstances, the abnormal disturbance factor is 1. In abnormal situations, the disturbance factor is updated in real time. The abnormal disturbance factor is mainly determined by experts. （10） S62: Combining the abnormal disturbance factor and the trend constraint factor, determine the short-term power generation optimization as the objective function; Power balance formula: （11） Multi-objective function: （12） Single objective function: （13） In equation (13), w c,short,i w represents the weight of short-term operating environment governance costs. cost,short,i w represents the weight of short-term operating costs. relfc,short,i w represents the weight by which short-term operational reliability is converted into economic cost. adjfc,short,i The weight representing the conversion of short-term operational resilience into cost; S63: Using the alternating direction multiplier method to solve for the short-term multi-objective and single-objective optimal functions of power generation, the optimal scheduling strategy for each distributed energy source with multiple objectives is obtained. For optimal scheduling of a single objective with comprehensive cost, the optimal strategy is: ; S64: Based on the short-term multi-objective and single-objective optimal scheduling strategies calculated in S63, the corresponding scheduling is carried out in combination with the actual application scenario; multi-objective scheduling is used under specific circumstances, while single-objective scheduling is used as the daily scheduling strategy with the lowest overall cost.

Images