Desert area-oriented hybrid energy storage and wind-solar prediction collaborative optimization method

Abstract

The invention discloses a desert-area-oriented hybrid energy storage and wind-light prediction collaborative optimization method, which comprises the following steps of: firstly, carrying out equipment modeling constraint on a multi-energy complementary system including a wind power generation system, a photovoltaic power generation system, a hybrid energy storage system and the like; establishing a wind and light prediction model by adopting BP neural network classification prediction optimized by a particle swarm algorithm; and performing joint optimization on the wind energy and illumination prediction model and the multi-energy complementary system optimization model containing the hybrid energy storage by using a differentiation creative search algorithm, and taking the wind energy and illumination prediction result as the input of the multi-energy complementary system optimization model containing the hybrid energy storage to obtain an optimal energy supply and energy storage scheduling scheme. The method has the advantages that the carbon emission is reduced, the consumption of new energy is promoted, the operation cost of the comprehensive energy system is reduced, the energy supply pressure is reduced, and the like.

Description

Technical Field

[0001] The present invention relates to a hybrid energy storage and wind-solar forecasting collaborative optimization method, and in particular to a hybrid energy storage and wind-solar forecasting collaborative optimization method for desert areas, and belongs to the field of energy technology. Background Art

[0002] Desert Gobi regions are rich in energy resources, but due to the complex geographical environment and unstable climatic conditions, traditional energy supply methods are difficult to meet demand. Hybrid energy storage systems for desert Gobi regions have high application potential, can effectively integrate renewable energy such as wind energy and solar energy, and improve energy utilization efficiency. Battery energy storage system (BESS), as a mature and widely used energy storage solution, is a vital component of hybrid energy storage systems. At the same time, hydrogen energy storage system (HES) and thermal energy storage system (TES) have unique advantages in desert Gobi regions, which can improve the flexibility and stability of the system. Optimization of wind power prediction model is the key to achieving wind and solar synergy. By accurately predicting meteorological factors such as wind speed, wind direction, and air pressure, the efficiency of wind power generation can be improved and dependence on traditional energy can be reduced. Considering the coordinated optimization of carbon emissions, carbon emissions in the energy supply process can be effectively reduced, which meets the requirements of environmental protection and sustainable development. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a hybrid energy storage and wind-solar forecasting coordinated optimization method for desert areas.

[0004] To achieve the above object, the present invention adopts the following technical solution:

[0005] A hybrid energy storage and wind-solar forecasting coordinated optimization method for desert areas, comprising the following steps:

[0006] Step S1: Establishing a multi-energy complementary system model:

[0007] The multi-energy complementary system model includes wind power generation module, photovoltaic power generation module, hybrid energy storage module, user load module, gas boiler module, cogeneration unit module, and power grid purchase and sale module. The wind power generation module and photovoltaic power generation module convert renewable energy into electrical energy, which is directly supplied to the user load module or traded through the power grid purchase and sale module. The gas boiler module and cogeneration unit module provide thermal energy to the user load module to meet the user's thermal load demand. The hybrid energy storage module includes BESS module, HES module, and TES module. The hybrid energy storage module stores the remaining electricity when there is excess electricity.

[0008] Step S2: Establishing a wind-solar prediction model: The wind-solar prediction model consists of a wind-solar output uncertainty module and a BP neural network; the output uncertainty generated by the wind-solar output uncertainty model is input into the BP neural network together with wind speed and light intensity as input features;

[0009] Step S3: Cascading the wind-solar forecast model with the multi-energy complementary system model, and using the differentiated creative search algorithm for joint optimization, to obtain the best energy supply and energy storage scheduling solution, including the following steps:

[0010] Step S3-1: Initialize parameters: First, set the population size and the range of decision variables, and then randomly generate the initial population X i , the initial population X i It consists of more than one individual; the variables contained in the individual include: the power P of hydrogen produced by the electrolyzer in the HES Bc , the charging power P of BESS Bd , the electrical power P obtained by the electrolyzer in HES ELE,in , the power P of hydrogen produced by electrolyzer in HES ELE，out , the electrical power P obtained by the fuel cell in the HES FC，in , the electrical power P generated by the fuel cell in the HES FC，out , the thermal storage power P of TES Tc , the heat release power P of TES Td , the output power of the combined heat and power unit P CHP , the actual output power of the wind turbine P PV , the actual output power P of the photovoltaic unit PW , objective function C tot The total operating cost of the multi-energy complementary system is:

[0011] x i，d =LB D +U(0,1)×(UB d -LB d ) (1)

[0012] In the formula, x i，d represents the i-th candidate solution X i =[x i，1 x i，2 …x i，D ] is the dth parameter of the candidate solution, D is the number of parameters of the candidate solution; the number of candidate solutions is NP; U(0,1) represents the uniform distribution on the interval (0,1), LB d and UB d Respectively represent the lower and upper bounds of the dth parameter;

[0013] Step S3-2: Fitness evaluation: Take the objective function value C corresponding to each individual totAs the initial fitness value:

[0014] f(X i )=C tot (X i ) (2)

[0015] Step S3-3: Select the individual corresponding to the maximum fitness value as the best individual X best ;

[0016] Step S3-4: Differentiated knowledge acquisition: Randomly select two individuals X r1 , X r2 , calculate η and coefficient:

[0017] The calculation method of η coefficient is:

[0018]

[0019] The coefficient is calculated as:

[0020]

[0021]

[0022] Where t is the number of iterations; the symbol [·] represents rounding, R i，t is the rank of the i-th individual at the beginning of the t-th iteration; x r1，d and x r2，d In two individuals X r1 , X r2 The dth parameter value of ;

[0023] Update parameters:

[0024]

[0025] Where: v i，d is the updated value of the dth parameter value of the ith individual;

[0026] Step S3-5: Knowledge integration and new solution generation:

[0027] Compute the convergence factor:

[0028]

[0029] In the formula, λ t is the convergence factor at the tth iteration; NFE t Indicates the current number of function calculations at the tth iteration, NFE max Indicates the maximum number of function calculations;

[0030] Update individual location:

[0031]

[0032] In the formula, x best，d is the dth parameter of the best individual in the current iteration;

[0033] Generate random numbers using Linique distribution to update individual positions:

[0034]

[0035] In the formula, x e，d is the dth parameter of the individual, Lk(α,σ) represents the random number generator of Linnik distribution with control parameters α and σ;

[0036] Step S3-6: Individual diversification: Randomly generate N new individuals:

[0037]

[0038] Where V np is the npth new individual, LB and UB are the lower and upper bounds;

[0039] Step S3-7: Retrospective evaluation: Replace the screened individuals with new individuals:

[0040]

[0041]

[0042] Where, X i,t+1 represents the i-th individual in X at the t+1 iteration; X best,t represents the optimal solution of X at t iterations;

[0043] Step S3-8: Determine whether the maximum number of iterations has been reached or the stop condition has been met. If so, output the final optimal solution and its corresponding optimal fitness value; otherwise, go to step S3-7;

[0044] Furthermore, in the step S1, the equipment of the multi-energy complementary system is modeled, including the following process:

[0045] S1-1. The multi-energy complementary system includes wind power generation module, photovoltaic power generation module, hybrid energy storage system module, user load module, gas boiler module and cogeneration unit module, and grid power purchase and sales module;

[0046] S1-2. Each module is modeled as follows:

[0047] Wind power generation is modeled as:

[0048]

[0049] Where P PW and P PW,N are the actual output power and rated output power of the wind turbine respectively; V PW and V PW,N are the actual wind speed and the rated wind speed respectively; V in and V out are the cut-in wind speed and cut-out wind speed of the fan respectively;

[0050] Photovoltaic power generation is modeled as:

[0051]

[0052] T C =T a +0.0138(1+0.0031T a )·(1-0.0042v)·S (17)

[0053] Where P PV and P PV,N is the actual output power and rated output power of the photovoltaic unit; f PV is the photovoltaic derating factor, which takes into account the effects of dirt, snow accumulation and battery aging, and is usually taken as 0.9; S is the light intensity in the actual environment; S T is the light intensity under standard working conditions, the value is 1000W / m 2 ; k photovoltaic power temperature coefficient; T C is the temperature of the photovoltaic unit in the actual environment; T T is the standard test working environment temperature, which is 25℃; T a is the actual ambient temperature (℃); v is the wind speed at the photovoltaic panel installation location (m / s); T T is the light intensity in the actual environment (W / m 2 );

[0054] The BESS model is:

[0055]

[0056] SOC min ≤SOC t ≤SOC max (20)

[0057] In the formula, SOC t , SOC t-1 is the remaining power of BESS at time t and t-1; β, C e is the self-discharge rate and rated capacity of BESS; P Bc , P Bd is the charging power and discharging power of BESS; ηBc , η Bd is the charging efficiency coefficient and discharging efficiency coefficient of BESS; P max，Bc , P max，Bd are the maximum charging power and maximum discharging power of BESS at time t respectively; P max，c , P max，d are the maximum continuous charging power and maximum continuous discharging power of BESS at time t respectively; Δt is the time interval;

[0058] HES is modeled as:

[0059]

[0060] P min，ELE ≤P ELE，t ≤P max，ELE (twenty three)

[0061] P min，FC ≤P FC，t ≤P max，FC (twenty four)

[0062] λ min，ELE ≤λ ELE，t ≤λ min，ELE (25)

[0063] Where P ELE,in , P ELE，out are the electric power obtained by the electrolyzer in HES and the electric power for hydrogen production by the electrolyzer; P FC，in , P FC，out are the electric power obtained and generated by the fuel cell in the HES, respectively; H ELE , H FC are the thermal power generated by the electrolyzer and the fuel cell respectively; η ELE , η FC is the working efficiency of electrolyzer and fuel cell; P max，ELE , P min，ELE are the upper and lower limits of the power of the electrolyzer respectively; P max，FC , P min，FC are the upper and lower limits of the fuel cell power respectively; max，EH , min，EH They are the upper and lower limits of the energy storage state of the hydrogen storage tank in the HES;

[0064] TES is modeled as:

[0065]

[0066] 0≤P Tc，t ≤P max，Tc ε Tc，t (27)

[0067] 0≤P Td，t ≤P max，Td ε Td，t (28)

[0068] H min，T ≤H T，t ≤H max，T (29)

[0069] 0≤ε Tc，t +ε Td，t ≤1 (30)

[0070] In the formula, H T，t is the heat storage capacity of TES at time t; μ is the self-heat release rate of TES; η Tc , η Td are the heat storage and heat release efficiencies of TES respectively; P Tc，t , P Td，t are the heat storage and heat release powers of TES at time t; ε Tc,t , ε Td，t are binary variables of heat storage and heat release of TES at time t; H max，T , H min，T are the maximum and minimum heat storage of TES respectively;

[0071] The gas boiler unit is modeled as:

[0072] Q GB =P GB η GB (31)

[0073] Where: Q GB P is the thermal power output of the gas boiler unit; GB is the gas consumption power of the gas boiler unit; η GB is the thermal efficiency of the gas boiler unit;

[0074] The combined heat and power unit is modeled as:

[0075] P CHP，g =G CHP H g (32)

[0076] P CHP =P CHP，g η CHP (33)

[0077] P CHP，h =P CHP，g η CHP，h (34)

[0078] Where: P CHP P is the electric power output of the cogeneration unit; CHP,hP is the thermal power output of the cogeneration unit; CHP,g The natural gas power input to the cogeneration unit; G CHP is the gas consumption of the cogeneration unit; H g is the lower calorific value of natural gas; η CHP and η CHP，g are the electrical efficiency and thermal efficiency of the cogeneration unit, respectively;

[0079] The model of the power grid purchasing and selling module is:

[0080] 0≤P be，t ≤P max，be (35)

[0081] 0≤P se，t ≤P max，se (36)

[0082] Where P be，t , P max，be are the amount of electricity purchased from the upper power grid and the upper limit of the purchased electricity; P se，t , P max，se The maximum amount of electricity sold to the upper grid and the amount of electricity sold;

[0083] Furthermore, in step S2, modeling is performed on the prediction of wind and solar energy, including the following process:

[0084] S2-1: Modeling of wind and solar power output uncertainty:

[0085]

[0086] In the formula, and is an uncertain parameter, is the net output power of the wind turbine and the photovoltaic unit; and Predict output values ​​for wind turbines and photovoltaic units; i It is expressed as the uncertainty in predicting the output of wind turbines and photovoltaic units, and its range is between 0 and 1; k1 and k2 are the influence of the error between the actual output and the predicted output of wind turbines and photovoltaic units in the overall uncertainty;

[0087] Step S2-2: Collect data: including historical power generation data of wind speed, light intensity, temperature, total power generation of wind turbine units and photovoltaic units, and wind and solar power output uncertainty factors, and perform data cleaning, normalization and feature extraction preprocessing on the collected data; with maximizing wind and solar power utilization as the objective function, use the particle swarm algorithm to optimize the connection weights and bias terms of the BP neural network;

[0088] The optimization objective function is:

[0089]

[0090] Wherein, Pwind(w, t) and Psolar(s, t) represent the power generation of wind turbines and photovoltaic generators when the wind force is w and the light intensity is s, respectively; F(w, t) represents the predicted value of wind energy when the wind force is w at time t; G(s, t) represents the predicted value of light energy when the light intensity is s at time t;

[0091] S2-3: Establishing the operation cost optimization model of multi-energy complementary system:

[0092] minC tot =C p +C ma +C op (40)

[0093] The corresponding power balance constraint is:

[0094] P load +P ELE，out +P Bc +P FC，in =P be +P CHP +P ELE，in +P Bd +P FC，out +P PV +P PW (41)

[0095] P HL +ε Tc P Tc =P CHP，h +P GB +ε Td P Td (42)

[0096] Where P Load , P HL , P be , P CHP，h , P GB , P CHP , P ELE，in , P ELE，out , P Bc , P Bd , P FC，in , P FC，out , P Tc , P Td , P PV , P PWRespectively represent the total load power demand of the multi-energy complementary system, the total heat load demand of the system, the electric power purchased from the power grid, the thermal power output of the cogeneration unit, the thermal power output of the gas boiler unit, the electric power output of the cogeneration unit, the electric power obtained by the electrolyzer in the HES, the electric power produced by the electrolyzer in the HES, the charging power of the BESS, the discharging power of the BESS, the electric power obtained by the fuel cell in the HES, the electric power generated by the fuel cell in the HES, the thermal storage power of the TES, the heat release power of the TES, the actual output power of the wind turbine unit, and the actual output power of the photovoltaic unit; ε Tc , ε Td are binary variables for heat storage and heat release of TES, respectively;

[0097] in:

[0098]

[0099] C ma =∑ i C ma，i S cap，i (44)

[0100] Where: i is 1, which means photovoltaic unit; i is 2, which means wind turbine unit; i is 3, which means battery energy storage system; i is 4, which means thermal energy storage system; i is 5, which means cogeneration unit; i is 6, which means electrolyzer (ELC); i is 7, which means hydrogen storage tank; i is 8, which means fuel cell; i is 9, which means gas boiler; i is 10, which means absorption refrigerator; i is 11, which means voltage refrigerator; r is the base discount rate; l i is the life cycle of equipment i; C p，i is the unit investment cost of equipment i; S p，i is the unit capacity of equipment i; C ma，i is the annual fixed maintenance cost of equipment i, S cap，i is the unit maintenance cost of equipment i;

[0101] Build an operating cost model:

[0102] C op =C ele +C bg +C en +C ab +C ob +C oe +C ot (45)

[0103]

[0104] In the formula, C ele is the transaction cost with the power grid, C bg is the gas purchase cost, Cen is the environmental cost of the gas unit, C ab The penalty cost for abandoning solar and wind power is C ob is the charge and discharge loss and maintenance cost of EES, C oe is the heat storage and charging and discharging heat loss cost of HES, C ot is the heat storage and charging and discharging heat loss cost of TES; be，t , se，t are the time-of-use electricity purchase price and the time-of-use electricity sales price traded with the power grid at time t, P be，t , P se，t are the purchased power and sold power at time t respectively; f g is the purchase price of natural gas; P CHP,t , P GB，t are the output power of the cogeneration unit and the gas boiler unit at time t respectively; η GB is the working efficiency of the gas boiler unit; η g Environmental costs of gas boiler units and power grid electricity purchase; w , pv are the penalty factors for wind and solar power abandonment respectively; ΔP aw，t , ΔP ap，t are the amount of wind and solar power abandoned at time t respectively.

[0105] The present invention combines the design of a hybrid energy storage system with battery energy storage and hydrogen energy storage to improve energy storage efficiency; uses a BP neural network classification prediction optimized by a particle swarm algorithm to establish a wind and solar prediction model, effectively improving the accuracy and reliability of energy supply; takes into account the environmental friendliness of the integrated energy system, which is in line with the current goals of carbon neutrality and carbon peak; uses a differentiated creative search algorithm to jointly optimize the prediction model of wind energy and light with the optimization model of a multi-energy complementary system containing hybrid energy storage, which can more effectively reduce carbon emissions, make more full use of the abundant wind and solar resources in the western region, and improve the reliability and economy of energy supply; at the same time, the energy management algorithm based on multi-objective optimization achieves a balance between economy and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0106] Figure 1 This is the structural diagram of the multi-energy complementary system in the western desert area.

[0107] Figure 2 It is a flow chart of the solution method of the present invention. DETAILED DESCRIPTION

[0108] The present invention will be further described below in conjunction with the accompanying drawings.

[0109] Reference Figure 1-2, a hybrid energy storage and wind-solar forecasting synergistic optimization method for the western desert region, the structure diagram of the multi-energy complementary system in the western desert region invented and studied is in Figure 1 It was demonstrated in the paper. Among them, wind and solar power generation and other renewable energy sources are used as the main energy input. At the same time, the system can also purchase electricity from the power grid and purchase external energy inputs such as natural gas from the gas grid. The power conversion and energy storage subsystems are used to meet the power demand; the electric energy is converted into hydrogen energy through hydrogen production, fuel cells, etc. to meet the hydrogen energy demand; at the same time, the remaining heat is stored through heat exchange equipment to meet the thermal energy demand. The energy trading mode of the integrated energy system allows energy transactions between the various subsystems within the system to achieve the goal of minimizing costs. In addition, the system can also trade energy with external power grids and gas grids to meet the diversification of the system's energy needs. The design of this integrated energy system not only improves energy utilization efficiency, but also effectively reduces carbon emissions, meeting the requirements of environmental protection and sustainable development.

[0110] A hybrid energy storage and wind-solar forecasting synergistic optimization method for western desert regions includes the following steps:

[0111] S1: Establish a multi-energy complementary system model:

[0112] The multi-energy complementary system model includes wind power generation module, photovoltaic power generation module, hybrid energy storage module, user load module, gas boiler module, cogeneration unit module, and power grid purchase and sale module. The wind power generation module and photovoltaic power generation module convert renewable energy into electrical energy, which is directly supplied to the user load module or traded through the power grid purchase and sale module. The gas boiler module and cogeneration unit module provide thermal energy to the user load module to meet the user's thermal load demand. The hybrid energy storage module includes BESS module, HES module, and TES module. The hybrid energy storage module stores the remaining electricity when there is excess electricity.

[0113] In the step S1, the equipment of the multi-energy complementary system is modeled, including the following process:

[0114] S1-1. The multi-energy complementary system includes wind power generation, photovoltaic power generation, hybrid energy storage system, user load, gas boiler and cogeneration unit, and power purchase and sales module of the power grid;

[0115] S1-2. Each module is modeled as follows:

[0116] Wind power generation is modeled as:

[0117]

[0118] Where P PW and P PW,N are the actual output power and rated output power of the wind turbine respectively; V PW and V PW,Nare the actual wind speed and the rated wind speed respectively; V in and V out are the cut-in wind speed and cut-out wind speed of the fan respectively;

[0119] Photovoltaic power generation is modeled as:

[0120]

[0121] Where P PV and P PV,N is the actual output power and rated output power of the photovoltaic unit; f PV is the photovoltaic derating factor, which takes into account the effects of dirt, snow accumulation and battery aging, and is usually taken as 0.9; S is the light intensity in the actual environment; S T is the light intensity under standard working conditions, the value is 1000W / m 2 ; k photovoltaic power temperature coefficient; T C is the temperature of the photovoltaic unit in the actual environment; T T is the standard test working environment temperature, which is 25℃; T a is the actual ambient temperature (℃); v is the wind speed at the photovoltaic panel installation location (m / s); T T is the light intensity in the actual environment (W / m 2 );

[0122] The BESS model is:

[0123]

[0124] SOC min ≤SOC t ≤SOC max (6)

[0125] In the formula, SOC t , SOC t-1 is the remaining power of BESS at time t and t-1; β, C e is the self-discharge rate and rated capacity of BESS; P Bc , P Bd is the charging power and discharging power of BESS; η Bc , η Bd is the charging efficiency coefficient and discharging efficiency coefficient of BESS; P max，Bc , P mdx，Bd are the maximum charging power and maximum discharging power of BESS at time t respectively; P max，c , P max，d are the maximum continuous charging power and maximum continuous discharging power of BESS at time t respectively; Δt is the time interval;

[0126] HES is modeled as:

[0127]

[0128] P min，ELE ≤P ELE，t ≤P max，ELE (9)

[0129] P min，FC ≤P FC，t ≤P max，FC (10)

[0130] λ min，WLE ≤λ ELE，t ≤λ min，ELE (11)

[0131] Where P ELE,in , P ELE，out are the electric power obtained by the electrolyzer in HES and the electric power for hydrogen production by the electrolyzer; P FC，in , P FC，out are the electric power obtained and generated by the fuel cell in the HES, respectively; H ELE , H FC are the thermal power generated by the electrolyzer and the fuel cell respectively; η ELE , η FC is the working efficiency of electrolyzer and fuel cell; P max，ELE , P min，ELE are the upper and lower limits of the power of the electrolyzer respectively; P max，FC , P min，FC are the upper and lower limits of the fuel cell power respectively; max，EH , min，EH They are the upper and lower limits of the energy storage state of the hydrogen storage tank in the HES;

[0132] TES is modeled as:

[0133]

[0134] 0≤P Tc，t ≤P max，Tc ε Tc，t (13)

[0135] 0≤P Td，t ≤P max，Td ε Td，t (14)

[0136] H min， T≤H T，t ≤H max，T (15)

[0137] 0≤ε Tc，t +ε Td，t ≤1 (16)

[0138] In the formula, H T，t is the heat storage capacity of TES at time t; μ is the self-heat release rate of TES; η Tc , η Td are the heat storage and heat release efficiencies of TES respectively; P Tc，t , P Td，t are the heat storage and heat release powers of TES at time t; ε Tc,t , ε Td，t are binary variables of heat storage and heat release of TES at time t; H max，T , H min，T are the maximum and minimum heat storage of TES respectively;

[0139] The gas boiler unit is modeled as:

[0140] Q GB =P GB η GB (17)

[0141] Where: Q GB P is the thermal power output of the gas boiler unit; GB is the gas consumption power of the gas boiler unit; η GB is the thermal efficiency of the gas boiler unit;

[0142] The combined heat and power unit is modeled as:

[0143] P CHP，g =G CHP H g (18)

[0144] P CHP =P CHP，g η CHP (19)

[0145] P CHP，h =P CHP，g η CHP，h _____(20)

[0146] Where: P CHP P is the electric power output of the cogeneration unit; CHP,h P is the thermal power output of the cogeneration unit; CHP,g The natural gas power input to the cogeneration unit; G CHP is the gas consumption of the cogeneration unit; H g is the lower calorific value of natural gas; η CHP and η CHP，g are the electrical efficiency and thermal efficiency of the cogeneration unit, respectively;

[0147] The model of the power grid purchasing and selling module is:

[0148] 0≤P be，t≤P max，be (twenty one)

[0149] 0≤P se，t ≤P max，se (twenty two)

[0150] Where P be，t , P max，be are the amount of electricity purchased from the upper power grid at time t and the upper limit of the purchased electricity; P se，t , P max，se The power sold to the upper grid at time t and the maximum value of the power sold;

[0151] S2: Establishing a wind-solar prediction model: The wind-solar prediction model consists of a wind-solar output uncertainty module and a BP neural network. The output uncertainty factor output by the wind-solar output uncertainty model is input into the BP neural network together with wind speed and light intensity as input features.

[0152] In step S2, modeling is performed for the prediction of wind and solar energy, including the following process:

[0153] S2-1. Modeling of wind and solar power output uncertainty:

[0154]

[0155] In the formula, and is an uncertain parameter, is the net output power of the wind turbine and the photovoltaic unit; and Predict output values ​​for wind turbines and photovoltaic units; i It is expressed as the uncertainty in predicting the output of wind turbines and photovoltaic units, and its range is between 0 and 1; k1 and k2 are the influence of the error between the actual output and the predicted output of wind turbines and photovoltaic units in the overall uncertainty;

[0156] S2-2. BP neural network classification prediction modeling based on particle swarm algorithm optimization

[0157] 1) Collect wind and solar-related data, including wind speed, light intensity, temperature and historical power generation data, and perform preprocessing such as data cleaning, normalization and feature extraction;

[0158] 2) Integrate the uncertainty model of wind and solar power output into the BP neural network. The specific steps include: first, identify and quantify the wind speed fluctuation and light intensity change factors that affect wind and solar power output by statistically analyzing the historical power generation data; second, extract these factors as additional features and input them into the BP neural network together with wind speed and light intensity as input features; finally, establish a classification prediction model to predict the future power generation of wind and solar energy;

[0159] 3) Use the particle swarm algorithm to optimize the connection weights and bias terms of the BP neural network to maximize the objective function of wind and solar utilization. The specific steps include: first, initialize the particle swarm to randomly generate a set of possible weight and bias combinations, then the particle swarm evaluates the effect of each combination based on the fitness function. The fitness function takes into account the real-time data of wind and solar energy and the historical power generation. Through continuous iteration, the particle swarm updates its position according to the best fitness and gradually converges to the optimal solution.

[0160] 4) Use the optimized BP neural network model to train the training data, and tune the model according to the performance of the validation set to obtain the best prediction performance;

[0161] Optimizing the objective function

[0162]

[0163] Where, Pwind(w, t) and Psolar(s, t) represent the power generation of the wind turbine and the photovoltaic unit when the wind force is w and the light intensity is s, respectively; F(w, t) represents the predicted wind energy value when the wind force is w at time t; G(s, t) represents the predicted light energy value when the light intensity is s at time t;

[0164] This objective function should be used in the optimization module of the prediction model. Specifically, it is used to guide the particle swarm algorithm to adjust the weights and biases when training the BP neural network. By maximizing this objective function, the model can effectively integrate the forecast data of wind and solar energy, thereby achieving the optimal scheduling of renewable energy generation capacity and improving the overall energy efficiency of the system.

[0165] S2-3. Construct a cost model and corresponding constraints for the operation of the multi-energy complementary system. The process is as follows:

[0166] Constructing investment and maintenance cost model of multi-energy complementary system

[0167] minC tot =C p +C ma +C op (26)

[0168] The corresponding power balance constraint is:

[0169] P load +P ELE，out +P Bc +P FC，in =P be +P CHP +P ELE，in +P Bd +P FC，out +PPV +P PW (27)

[0170] P HL +ε Tc P Tc =P CHP，h +P GB +ε Td P Td (28)

[0171] Where P Load , P HL , P be , P CHP，h , P GB , P CHP , P ELE，in , P ELE，out , P Bc , P Bd , P FC，in , P FC，out , P Tc , P Td , P PV , P PW Respectively represent the total load power demand of the multi-energy complementary system, the total heat load demand of the system, the electric power purchased from the power grid, the thermal power output of the cogeneration unit, the thermal power output of the gas boiler unit, the electric power output of the cogeneration unit, the electric power obtained by the electrolyzer in the HES, the electric power produced by the electrolyzer in the HES, the charging power of the BESS, the discharging power of the BESS, the electric power obtained by the fuel cell in the HES, the electric power generated by the fuel cell in the HES, the thermal storage power of the TES, the heat release power of the TES, the actual output power of the wind turbine unit, and the actual output power of the photovoltaic unit; ε Tc , ε Td are binary variables for heat storage and heat release of TES, respectively;

[0172] in:

[0173]

[0174] C ma =∑ i C ma，i S cap，i (30)

[0175] Where: i is 1, indicating a photovoltaic unit; i is 2, indicating a wind turbine unit; i is 3, indicating a battery energy storage system; i is 4, indicating a thermal energy storage system; i is 5, indicating a combined heat and power unit; i is 6, indicating an electrolyzer (ELC); i is 7, indicating a hydrogen storage tank; i is 8, indicating a fuel cell; i is 9, indicating a gas boiler unit; i is 10, indicating an absorption refrigerator; i is 11, indicating a voltage refrigerator; r is the base discount rate; l i is the life cycle of equipment i; C p，i is the unit investment cost of equipment i; S p，i is the unit capacity of equipment i; C ma，i is the annual fixed maintenance cost of equipment i, S cap，i is the unit maintenance cost of equipment i;

[0176] C op =C ele +C bg +C en +C ab +C ob +C oe +C ot (31)

[0177]

[0178] In the formula, C ele is the transaction cost with the power grid, C bg is the gas purchase cost, C en is the environmental cost of the gas unit, C ab Penalty costs for abandoning solar and wind power, C ob is the charge and discharge loss and maintenance cost of EES, C oe is the heat storage and charging and discharging heat loss cost of HES, C ot is the heat storage and charging and discharging heat loss cost of TES; be，t , se，t are the purchase price and sales price of electricity traded with the power grid at time t, P be，t , P se，t are the amount of electricity purchased from the upper power grid and the amount of electricity sold to the upper power grid at time t; f g is the purchase price of natural gas; H g is the lower calorific value of natural gas; P CHP,t , P GB，t are the output power of the cogeneration unit and the gas boiler unit at time t respectively; η GB is the thermal efficiency of the gas boiler unit; η g Environmental costs of gas boiler units and power grid electricity purchase; w , pv are the penalty factors for wind and solar power abandonment respectively; ΔPaw，t , ΔP ap，t are the amount of wind and solar power abandoned at time t respectively.

[0179] Step S3: Cascading the wind-solar forecast model with the multi-energy complementary system model, and using the differentiated creative search algorithm for joint optimization, to obtain the best energy supply and energy storage scheduling solution, including the following steps:

[0180] Step S3-1: Initialize parameters: First, set the population size and the range of decision variables, and then randomly generate the initial population X i , the initial population X i It consists of more than one individual; the variables contained in the individual include: the power P of hydrogen produced by the electrolyzer in the HES Bc , the charging power P of BESS Bd , the electrical power P obtained by the electrolyzer in HES ELE,in , the power P of hydrogen produced by electrolyzer in HES ELE，out , the electrical power P obtained by the fuel cell in the HES FC，in , the electrical power P generated by the fuel cell in the HES FC，out , the thermal storage power P of TES Tc , the heat release power P of TES Td , the output power of the combined heat and power unit P CHP , the actual output power of the wind turbine P PV , the actual output power P of the photovoltaic unit PW , objective function C tot The total operating cost of the multi-energy complementary system is:

[0181] x i，d =LB d +U(0,1)×(UB d -LB d ) (36)

[0182] In the formula, x i，d represents the i-th candidate solution X i =[x i，1 x i，2 …x i，D ] is the dth parameter of the candidate solution, D is the number of parameters of the candidate solution; the number of candidate solutions is NP; U(0,1) represents the uniform distribution on the interval (0,1), LB d and UB d Respectively represent the lower and upper bounds of the dth parameter;

[0183] Step S3-2: Fitness evaluation: Take the objective function value C corresponding to each individual tot As the initial fitness value:

[0184] f(X i)=C tot (X i ) (37)

[0185] Step S3-3: Select the individual corresponding to the maximum fitness value as the best individual X best ;

[0186] Step S3-4: Differentiated knowledge acquisition: Randomly select two individuals X r1 , X r2 , calculate η and coefficient:

[0187] The calculation method of η coefficient is:

[0188]

[0189] The coefficient is calculated as:

[0190]

[0191] Where t is the number of iterations; the symbol [·] represents rounding, R i，t is the rank of the i-th individual at the beginning of the t-th iteration; x r1,d and x r2，d In two individuals X r1 , X r2 The dth parameter value of ;

[0192] Update parameters:

[0193]

[0194] Where: v i，d is the updated value of the dth parameter value of the ith individual;

[0195] Step S3-5: Knowledge integration and new solution generation:

[0196] Compute the convergence factor:

[0197]

[0198] In the formula, λ t is the convergence factor at the tth iteration; NFE t Indicates the current number of function calculations at the tth iteration, NFE max Indicates the maximum number of function calculations;

[0199] Update individual location:

[0200]

[0201] In the formula, x best，dis the dth parameter of the best individual in the current iteration;

[0202] Generate random numbers using Linique distribution to update individual positions:

[0203]

[0204] In the formula, x e，d is the dth parameter of the individual, Lk(α,σ) represents the random number generator of Linnik distribution with control parameters α and σ;

[0205] Step S3-6: Individual diversification: Randomly generate N new individuals:

[0206]

[0207] Where V np is the npth new individual, LB and UB are the lower and upper bounds;

[0208] Step S3-7: Retrospective evaluation: Replace the screened individuals with new individuals:

[0209]

[0210] Where, X i，t+1 represents the i-th individual in X at the t+1 iteration; X best，t represents the optimal solution of X at t iterations;

[0211] Step S3-8: Determine whether the maximum number of iterations has been reached or the stopping condition has been met. If so, output the final optimal solution and its corresponding optimal fitness value; otherwise, go to step S3-7.

Claims

1. A hybrid energy storage and wind-solar forecasting collaborative optimization method for desert areas, characterized in that: The following steps are involved: Step S1: Establishing a multi-energy complementary system model: The multi-energy complementary system model includes wind power generation module, photovoltaic power generation module, hybrid energy storage module, user load module, gas boiler module, cogeneration unit module, and power grid purchase and sale module. The wind power generation module and photovoltaic power generation module convert renewable energy into electrical energy, which is directly supplied to the user load module or traded through the power grid purchase and sale module. The gas boiler module and cogeneration unit module provide thermal energy to the user load module to meet the user's thermal load demand. The hybrid energy storage module includes BESS module, HES module, and TES module. The hybrid energy storage module stores the remaining electricity when there is excess electricity. Step S2: Establishing a wind-solar prediction model: The wind-solar prediction model consists of a wind-solar output uncertainty module and a BP neural network; the output uncertainty generated by the wind-solar output uncertainty model is input into the BP neural network together with wind speed and light intensity as input features; Modeling of wind and solar power output uncertainty: In the formula, and is an uncertain parameter, is the net output power of the wind turbine and the photovoltaic unit; and Predict output values ​​for wind turbines and photovoltaic units; i It is expressed as the uncertainty in predicting the output of wind turbines and photovoltaic units, and its range is between 0 and 1; k1 and k2 are the influence of the error between the actual output and the predicted output of wind turbines and photovoltaic units in the overall uncertainty; Step S3: Cascading the wind-solar forecast model with the multi-energy complementary system model, and using the differentiated creative search algorithm for joint optimization, to obtain the best energy supply and energy storage scheduling solution, including the following steps: Step S3-1: Initialize parameters: First, set the population size and the range of decision variables, and then randomly generate the initial population X i , the initial population X i It consists of more than one individual; the variables contained in the individual include: the power P of hydrogen produced by the electrolyzer in the HES Bc , the charging power P of BESS Bd , the electrical power P obtained by the electrolyzer in HES ELE,in , the power P of hydrogen produced by electrolyzer in HES ELE,out , the electrical power P obtained by the fuel cell in the HES FC,in , the electrical power P generated by the fuel cell in the HES FC,out , the thermal storage power P of TES Tc , the heat release power P of TES Td , the output power of the combined heat and power unit P CHP , the actual output power of the wind turbine P PV , the actual output power P of the photovoltaic unit PW , objective function C tot The total operating cost of the multi-energy complementary system is: x i,d =LB d +U(0,1)×(UB d -LB d ) (3) In the formula, x i,d represents the i-th candidate solution X i =[x i,1 x i,2 ...x i,D ] is the dth parameter of the candidate solution, D is the number of parameters of the candidate solution; the number of candidate solutions is NP; U(0,1) represents the uniform distribution on the interval (0,1), LB d and UB d Respectively represent the lower and upper bounds of the dth parameter; Step S3-2: Fitness evaluation: Take the objective function value C corresponding to each individual tot As the initial fitness value: f(X i )=C tot (X i ) (4) Step S3-3: Select the individual corresponding to the maximum fitness value as the best individual X best ; Step S3-4: Differentiated knowledge acquisition: Randomly select two individuals X r1 , X r2 , calculate η and coefficient: The calculation method of η coefficient is: The coefficient is calculated as: Where t is the number of iterations; the symbol [·] represents rounding, R i,t is the rank of the i-th individual at the beginning of the t-th iteration; Update parameters: Where: x r1,d and x r2,d In two individuals X r1 , X r2 The dth parameter value of i,d is the updated value of the dth parameter value of the ith individual; Step S3-5: Knowledge integration and new solution generation: Compute the convergence factor: In the formula, λ t is the convergence factor at the tth iteration; NFE t Indicates the current number of function calculations at the tth iteration, NFE max Indicates the maximum number of function calculations; Update individual location: In the formula, x best,d is the dth parameter of the best individual in the current iteration; Generate random numbers using Linique distribution to update individual positions: In the formula, x e,d is the dth parameter of the individual, Lk(α,σ) represents the random number generator of Linnik distribution with control parameters α and σ; Step S3-6: Individual diversification: Randomly generate N new individuals: Where V np is the npth new individual, LB and UB are the lower and upper bounds; Step S3-7: Retrospective evaluation: Replace the screened individuals with new individuals: Where, X i,t+1 represents the i-th individual in X at the t+1 iteration; X best,t represents the optimal solution of X at t iterations; Step S3-8: Determine whether the maximum number of iterations has been reached or the stopping condition has been met. If so, output the final optimal solution and its corresponding optimal fitness value; otherwise, go to step S3-7.

2. The hybrid energy storage and wind-solar forecasting collaborative optimization method for desert areas according to claim 1 is characterized by: In step S1: The BESS model is: SOC min ≤SOC t ≤SOC max (19) In the formula, SOC t , SOC t-1 is the remaining power of BESS at time t and t-1; β, C e is the self-discharge rate and rated capacity of BESS; P Bc , P Bd is the charging power and discharging power of BESS; η Bc , η Bd is the charging efficiency coefficient and discharging efficiency coefficient of BESS; P max,Bc , P max,Bd are the maximum charging power and maximum discharging power of BESS at time t respectively; P max,c , P max,d are the maximum continuous charging power and maximum continuous discharging power of BESS at time t respectively; Δt is the time interval; HES is modeled as: P min,ELE ≤P ELE,t ≤P max,ELE (22) P min,FC ≤P FC,t ≤P max,FC (23) l min,ELE ≤λ ELE,t ≤λ min,ELE (24) Where P ELE,in , P ELE,out are the electric power obtained by the electrolyzer in HES and the electric power for hydrogen production by the electrolyzer; P FC,in , P FC,out are the electric power obtained and generated by the fuel cell in the HES, respectively; H ELE , H FC are the thermal power generated by the electrolyzer and the fuel cell respectively; η ELE , η FC is the working efficiency of electrolyzer and fuel cell; P max,ELE , P min,ELE are the upper and lower limits of the power of the electrolyzer respectively; P max,FC , P min,FC are the upper and lower limits of the fuel cell power respectively; max,EH , min,EH They are the upper and lower limits of the energy storage state of the hydrogen storage tank in the HES; TES is modeled as: 0≤P Tc,t ≤P max,Tc e Tc,t (26) 0≤P Td,t ≤P max,Td e Td,t (27) H min,T ≤H T,t ≤H max,T (28) 0≤ε Tc,t +e Td,t ≤1 (29) In the formula, H T,t is the heat storage capacity of TES at time t; μ is the self-heat release rate of TES; η Tc , η Td are the heat storage and heat release efficiencies of TES respectively; P Td,t , P Td,t are the heat storage and heat release powers of TES at time t; ε Tc,t , ε Td,t are binary variables of heat storage and heat release of TES at time t; H max,T , H min,T are the maximum and minimum heat storage of TES respectively; The model of the power grid purchasing and selling module is: 0≤P be,t ≤P max,be (30) 0≤P se,t ≤P max,se (31) Where P be,t , P max,be are the amount of electricity purchased from the upper power grid and the upper limit of the purchased electricity; P se,t , P max,se It is the amount of electricity sold to the upper grid and the maximum amount of electricity sold.

3. The method for coordinated optimization of hybrid energy storage and wind-solar forecasting for desert areas according to claim 1 is characterized in that: The training process of the BP neural network in step S2 consists of the following specific steps: Step S2-1: Collect data: including historical power generation data of wind speed, light intensity, temperature and total power generation of wind turbine units and photovoltaic units, and perform data cleaning, normalization and feature extraction preprocessing on the collected data; Step S2-2: Taking maximizing the utilization rate of wind and solar power as the objective function, the particle swarm algorithm is used to optimize the connection weights and bias terms of the BP neural network; The optimization objective function is: Where Pwind(w,t) and Psolar(s,t) represent the power generation of the wind turbine and photovoltaic generator respectively when the wind force is w and the light intensity is s; F(w,t) represents the predicted wind energy value when the wind force is w at time t; G(s,t) represents the predicted light energy value when the light intensity is s at time t; S2-3: Establishing the operation cost optimization model of multi-energy complementary system: minC tot =C p +C ma +C op (33) The corresponding power balance constraint is: P load +P ELE,out +P Bc +P FC,in =P be +P CHP +P ELE,in +P Bd +P FC,out +P PV +P wind (34) P HL +ε Tc P Tc =P CHP,h +P GB +ε Td P Td (35) Where P Load , P HL , P be , P CHP,h , P GB , P CHP , P ELE,in , P ELE,out , P Bc , P Bd , P FC,in , P FC,out , P Tc , P Td , P PV , P PW They represent the total load power demand of the multi-energy complementary system, the total heat load demand of the system, the electric power purchased from the power grid, the thermal power output of the cogeneration unit, the heating power of the gas boiler, the power generation power of the cogeneration unit, the electric power obtained by the electrolyzer, the electric power used by the electrolyzer to produce hydrogen, the charging power of the BESS, the discharging power of the BESS, the electric power obtained by the fuel cell, the electric power generated by the fuel cell, the thermal storage power of the TES, the heat release power of the TES, the photovoltaic power generation power, and the wind turbine power generation power; ε Tc , ε Td are binary variables for heat storage and heat release of TES, respectively; in: C ma =∑ i C ma,i S cap,i (37) In the formula: i is 1, which means photovoltaic unit; i is 2, which means wind turbine unit; i is 3, which means battery energy storage system; i is 4, which means thermal energy storage system; i is 5, which means cogeneration unit; i is 6, which means electrolyzer ELC; i is 7, which means hydrogen storage tank; i is 8, which means fuel cell; i is 9, which means gas boiler; i is 10, which means absorption refrigerator; i is 11, which means voltage refrigerator; r is the base discount rate; l i is the life cycle of equipment i; C p,i is the unit investment cost of equipment i; S p,i is the unit capacity of equipment i; C ma,i is the annual fixed maintenance cost of equipment i, S cap,i is the unit maintenance cost of equipment i; Build an operating cost model: C op =C ele +C bg +C en +C ab +C ob +C oe +C ot (38) In the formula, C ele is the transaction cost with the power grid, C bg is the gas purchase cost, C en is the environmental cost of the gas unit, C ab Penalty costs for abandoning solar and wind power, C ob is the charge and discharge loss and maintenance cost of EES, C oe is the heat storage and charging and discharging heat loss cost of HES, C ot is the heat storage and charging and discharging heat loss cost of TES; be,t , se,t are the time-of-use electricity purchase price and the time-of-use electricity sales price traded with the power grid at time t, P be,t , P se,t are the purchased power and sold power at time t respectively; f g is the purchase price of natural gas; P CHP,t , P GB,t are the output power of the cogeneration unit and the gas boiler unit at time t respectively; η GB is the working efficiency of the gas boiler unit; η g Environmental costs of gas boiler units and power grid electricity purchase; w , pv are the penalty factors for wind and solar power abandonment respectively; ΔP aw,t , ΔP aw,t are the amount of wind and solar power abandoned at time t respectively.

Images