An optimization method for wind power hybrid energy storage heating system considering capacity configuration

Abstract

The present application proposes an optimization method for a wind power hybrid energy storage heating system that takes capacity configuration into consideration, which belongs to the field of wind power hybrid energy storage technology, including: clustering and analyzing residential heat load data throughout the year by day using the k-means clustering analysis algorithm; obtaining wind power data in representative typical scenarios using a two-parameter Weibull model and scenario reduction technology; inputting hourly heat load distribution data on typical days and wind power data in representative typical scenarios into a two-layer optimization model for quantity configuration to iterate, and obtaining the optimal system capacity configuration plan; inputting the device capacity after particle swarm optimization into the lower model of the capacity configuration two-layer optimization model for iteration, and inputting the optimized hourly power and cost into the upper model of the capacity configuration two-layer optimization model for iteration. The present application takes into account the comprehensive benefits brought by the wind power storage heating project, and optimizes the capacity configuration in a supply-demand coordinated manner.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind power hybrid energy storage, and in particular relates to an optimization method for a wind power hybrid energy storage heating system taking capacity configuration into consideration. Background Art

[0002] The demand for heat load in northern residential heating season is huge, the amount of clean energy access is small, and the energy utilization efficiency is low. Wind power generation is one of the most widely used clean energy power generation methods because of its clean, low-carbon and abundant reserves. However, the randomness and volatility of wind power and the matching problem of building load have brought safety and stability problems to the energy system, and the problem of wind abandonment has become prominent. To solve this problem, installing supporting energy storage and heat storage devices in the renewable energy system is of great significance to improve the utilization rate of renewable energy and ensure the economic and stable operation of the system.

[0003] Domestic and foreign research on wind energy heating can be roughly divided into three categories: 1) using wind energy to take away low-temperature air, such as using wind energy to spray liquid heat exchange towers; 2) connecting wind power to heating equipment, such as electric boilers, split air conditioners, air source heat pumps, etc.; 3) wind power to power transmission equipment, such as circulating pumps, submersible pumps, etc. However, most of the existing research only takes economic cost as the optimization target, without considering the comprehensive benefits brought by wind power energy storage heating projects, and the mismatch between supply and demand leads to excessive initial investment, or does not take operating costs into account. Therefore, there is a need for a method to optimize the capacity configuration for supply and demand coordination. Summary of the invention

[0004] In view of the shortcomings of the prior art, this application proposes an optimization method for a wind power hybrid energy storage heating system considering capacity configuration, including:

[0005] Step S1: Obtaining the annual residential heat load data and wind speed data of the area to be optimized;

[0006] Step S2: For the annual residential heat load data of the area to be optimized, the k-means clustering analysis algorithm is used to cluster and analyze by day to obtain the hourly heat load distribution data of a typical day;

[0007] Step S3: Fit the wind speed data using a two-parameter Weibull model, sample the fitted wind speed data using the Monte Carlo method, obtain the wind power of a wind speed scenario each time, and reduce the wind power of the sampled wind speed scenario using a fast previous-generation scenario reduction technique based on probability distance, to obtain representative typical scenarios and wind power data in representative typical scenarios;

[0008] Step S4: Input the hourly heat load distribution data of a typical day and the wind power data in a representative typical scenario into a pre-established capacity configuration two-layer optimization model for iteration to obtain the optimal system capacity configuration plan; wherein, the upper model of the capacity configuration two-layer optimization model uses a particle swarm optimization algorithm to optimize the equipment capacity with the goal of maximizing the net benefit over the entire life cycle, and inputs the optimized equipment capacity into the lower model of the capacity configuration two-layer optimization model for iteration, and the lower model of the capacity configuration two-layer optimization model optimizes the hourly power of each device with the goal of minimizing the operating cost, and inputs the optimized hourly power and cost into the upper model of the capacity configuration two-layer optimization model for iteration.

[0009] The annual residential heat load data of the area to be optimized is clustered and analyzed by day using the k-means clustering analysis algorithm to obtain hourly heat load distribution data of a typical day, including:

[0010] Step S2.1: For the annual residential heat load data of the area to be optimized, the k-means clustering analysis algorithm is used to cluster them by day;

[0011] Step S2.2: Determine the optimal number of clusters according to the CH index, and the CH index calculation formula is as follows:

[0012]

[0013] Among them, CH(nk) is the CH index with nk as the cluster number, B nk is the distance between classes, W nk is the intra-class distance, and N is the number of samples of the annual residential heat load data in the area to be optimized;

[0014] Step S2.3: Based on the optimal number of clusters determined by the CH index, cluster analysis is performed on the annual residential heat load data of the optimized area to obtain hourly heat load distribution data of a typical day.

[0015] According to the optimal cluster number determined by the CH index, cluster analysis is performed on the annual residential heat load data of the optimized area to obtain hourly heat load distribution data of a typical day, including:

[0016] Step S2.3.1: Read all objects in the annual residential heat load data of the area to be optimized, select k objects as initial cluster centers according to the optimal number of clusters determined by the CH index, and calculate the Euclidean distance between the remaining objects and the initial cluster centers;

[0017] Step S2.3.2: Divide all objects in the annual residential heat load data of the area to be optimized into the class with the closest distance, recalculate the mean of all objects in the class, use the mean as the new cluster center of the class, iterate until the convergence condition is reached, obtain the classification result of the typical day, and determine the number of typical days;

[0018] Step S2.3.3: Take the average of the days of all typical days to obtain the hourly heat load distribution data of the typical days.

[0019] The two-parameter Weibull model is used to fit the wind speed data, and the Monte Carlo method is used to sample the fitted wind speed data. Each sampling obtains the wind power of a wind speed scenario, and the fast previous generation scenario reduction technology based on probability distance is used to reduce the wind power of the sampled wind speed scenario, so as to obtain representative typical scenarios and wind power data in the representative typical scenarios, including:

[0020] Step S3.1: Use the two-parameter Weibull model to fit the wind speed data. The wind speed probability density function after fitting is as follows:

[0021]

[0022] Where v is the wind speed, q is the shape parameter, c is the scale parameter, and f(v) is the wind speed probability density corresponding to wind speed v;

[0023] Step S3.2: The wind speed data is sampled based on the wind speed probability density function using the Monte Carlo method, a wind speed scenario is obtained each time a sample is taken, and the wind power of the wind speed scenario is calculated. The calculation formula is as follows:

[0024]

[0025] Where P is the wind power in the wind speed scenario, P r is the rated power, v is the wind speed, v r is the rated wind speed, v ci is the cut-in wind speed, v co To cut out the wind speed;

[0026] Step S3.3: Use the fast previous-generation scenario reduction technology based on probability distance to reduce the wind power of the sampled wind speed scenario, and obtain representative typical scenarios and wind power data in representative typical scenarios.

[0027] The fast previous generation scenario reduction technology based on probability distance is used to reduce the wind power of the sampled wind speed scenario, and obtain representative typical scenarios and wind power data in the representative typical scenarios, including:

[0028] Step S3.3.1: Calculate the wind power P of any wind speed scenario in turn among all wind power sets P of the wind speed scenario set. i The Euclidean distance between the wind power of all other wind speed scenarios;

[0029] Step S3.3.2: Calculate the sum of the Euclidean distances between each wind speed scenario and the wind power of all other wind speed scenarios, find the wind speed scenario with the smallest sum of the Euclidean distances of wind power, and define it as a;

[0030] Step S3.3.3: Find the wind power of wind speed scenario b that is shortest in distance to the wind power of wind speed scenario a, replace the wind power of wind speed scenario a with the wind power of wind speed scenario b, add the wind speed probability density corresponding to the wind speed of a to the wind speed probability density corresponding to the wind speed of b, and form a new wind power set P' of the wind speed scenario set;

[0031] Step S3.3.4: Determine whether the current number of scenes is equal to the preset number; if not, repeat steps S3.3.1-S3.3.3; if equal, terminate scene reduction, take the current scene as the representative typical scene, and output the representative typical scene and the wind power data in the representative typical scene.

[0032] The particle swarm optimization algorithm uses an acceleration factor for optimization during search, and the calculation formula is as follows:

[0033] C1=C 1s +k(C 1e -C 1s ) / k max

[0034] C2=C 2s +k(C 2e -C 2s ) / k max

[0035] Among them, C1 is the individual acceleration factor, C2 is the social acceleration factor, and C 1e is the final value of the individual acceleration factor C1, C 2e is the final value of the social acceleration factor C2, C 1s is the initial value of the individual acceleration factor C1, C 2s is the initial value of the social acceleration factor C2, k max is the maximum number of iterations.

[0036] The upper model of the capacity configuration double-layer optimization model constructs the objective function based on heating revenue, electricity sales revenue, coal saving revenue, environmental revenue, peak load regulation revenue, equipment investment cost, operation cost, maintenance cost and decommissioning cost. The calculation formula is as follows:

[0037] f=Kh +K e +K c +K t +K ps

[0038] -(C Inv +C O +C M +C D )

[0039] Among them, f is the objective function value, K h is the heating benefit, K e is the revenue from electricity sales, K c For coal saving benefits, K t For environmental benefits, K ps is the peak load benefit, C Inv is the investment cost, C O is the operating cost, C M is the maintenance cost, C D for decommissioning costs;

[0040] The heating benefit is calculated as follows:

[0041]

[0042] Among them, C h is the heating unit price; P h-load is the heat load, T is the number of hours in a day, i.e. 24 hours;

[0043] The electricity sales revenue is calculated as follows:

[0044]

[0045] in, P is the price of electricity sold to the grid, e,t is the interactive power with the power grid during period t. If it is greater than zero, it indicates power purchase, and if it is less than zero, it indicates power sale;

[0046] The coal saving benefit is calculated as follows:

[0047]

[0048] Among them, P s is the total power generation of the system, P e,t is the interaction power with the grid during period t, μ cfu is the coal consumption rate of the coal-fired boiler, C Coal is the price of coal;

[0049] The environmental benefits are calculated as follows:

[0050]

[0051] Among them, P e,t is the interaction power with the power grid during period t; C t is the price per unit of CO2 emission;

[0052] The peak load benefit is calculated as follows:

[0053]

[0054] in, is the charging power of the battery during period t, is the heat storage power of the solid heat storage device during period t, C ps Subsidize the price of energy storage equipment for peak load regulation;

[0055] The investment cost is calculated as follows:

[0056]

[0057] Among them, x cap is the capacity of device i, C Inv,i is the unit capacity investment cost of equipment i, N i is the service life of device i, r is the discount rate, I is the total number of equipment;

[0058] The operating cost is calculated as follows:

[0059] C O =C s +C e

[0060]

[0061] Among them, C s,i is the unit operating cost of equipment i, is the power value of device i in period t, P is the price of electricity purchased from the power grid; e,t is the interaction power with the power grid during period t;

[0062] The maintenance cost is calculated as follows:

[0063]

[0064] Among them, x cap is the capacity of device i, C m,i is the unit capacity maintenance cost of equipment i, N i is the service life of device i, r is the discount rate.

[0065] The decommissioning cost is calculated as follows:

[0066]

[0067] Among them, a1 is the retired equipment disposal fee rate, a2 is the equipment residual value rate, x cap is the capacity of device i, C Inv,i is the unit capacity investment cost of equipment i, N i is the useful life of equipment i, and r is the discount rate.

[0068] The step of inputting the optimized equipment capacity into the lower model of the capacity configuration double-layer optimization model for iteration includes: inputting the hourly heat load distribution data of a typical day optimized by the particle swarm optimization algorithm and the wind power data in a representative typical scenario into the lower model of the capacity configuration double-layer optimization model to obtain a simplified operating cost, which is calculated as follows:

[0069] C' o =C′ s +C′ e

[0070]

[0071] Among them, d k is the number of days of typical day category k, k is the number of typical day categories, is the unit operating cost of equipment i under category k on a typical day, is the power value of device i in period t under typical day category k, P is the price of electricity purchased from the grid. e,t is the interaction power with the grid during period t, C' o is the simplified operating cost, C′ s The simplified operating cost of the electricity storage, heat storage and heating equipment is C e ' is the simplified electricity purchase cost, I is the total number of equipment, and K is the total number of categories on a typical day.

[0072] The pre-established capacity configuration two-layer optimization model constructs constraints including: electric power balance constraint, thermal power balance constraint, wind power output constraint, energy storage equipment constraint, and heating equipment constraint;

[0073] The electric power balance constraint is calculated as follows:

[0074]

[0075] Among them, P Wt is the wind power generation power, P B is the battery charging and discharging power, P e is the grid interaction power, Input power for heat pump heating;

[0076] The thermal power balance constraint is calculated as follows:

[0077]

[0078] in, is the heating output power of the heat pump, P Hs is the charging and discharging power of the solid heat storage device, P h-load is the heat load demand;

[0079] The wind power output constraint is calculated as follows:

[0080] 0≤P Wt (t)≤P Wt,max

[0081] Among them, P Wt is the wind power generation power, P Wt,max The maximum power of wind power generation.

[0082] The energy storage device constraint is calculated as follows:

[0083]

[0084] Among them, P i in (t) represents the electrical / thermal storage power at time t, Indicates the maximum electrical / thermal storage power, P i out (t) represents the discharge / heat release power at time t, represents the maximum discharge / heat release power, X and Y are 0-1 variables of the energy storage device’s charge and discharge state; the energy storage device’s state of charge SOC constraint is as follows:

[0085] S SOCmin ≤S SOC (t)≤S SOCmax

[0086] Among them, S SOC (t) is the battery charge state at time t, S SOCmin is the upper limit of the state of charge, S SOCmax It is the lower limit of the charge state;

[0087] The heating equipment constraint is calculated as follows:

[0088] Heating power constraints of water source heat pump:

[0089]

[0090] Among them, η Hp is the heating efficiency of the heat pump, represents the input power of the heat pump at time t, PHp,max Indicates the maximum heating power of the heat pump.

[0091] The lower model of the capacity configuration double-layer optimization model optimizes the hourly power of each device with the goal of minimizing the operating cost, and inputs the optimized hourly power and cost into the upper model of the capacity configuration double-layer optimization model for iteration, including: inputting the hourly heat load distribution data of a typical day and the wind power data in a representative typical scenario into the lower model of the capacity configuration double-layer optimization model, using the CPLEX solver for operation optimization, obtaining the optimized hourly power and cost, and inputting the optimized hourly power and cost into the upper model of the capacity configuration double-layer optimization model for iteration.

[0092] Beneficial technical effects of the present invention:

[0093] This application proposes an optimization method for a wind power hybrid energy storage heating system considering capacity configuration, and establishes a two-layer optimization model for capacity configuration. First, the k-means clustering analysis algorithm is used to cluster the annual heat load data of northern residential buildings by day, and the number of days of typical days and the hourly heat load distribution data of typical days are obtained; secondly, the Monte Carlo method is used to sample the probability distribution model of wind speed based on the two-parameter Weibull model to obtain the predicted wind speed scenario according to the wind speed data, and the wind speed-wind power conversion formula is used to generate the wind power output scenario. Then, a representative typical scenario is obtained by using the fast previous generation scenario reduction technology based on probability distance; further, the hourly heat load distribution data and wind power data of typical days are input into the two-layer capacity configuration optimization model; the model is solved by the particle swarm optimization algorithm; the upper model optimizes the equipment capacity of the system with the goal of maximizing the net profit of the entire life cycle, and the lower model optimizes the hourly power of each device in the system with the goal of minimizing the operating cost; finally, through the two-way coupling between the upper and lower layers and nested iteration, the optimal system capacity configuration plan is obtained. This application takes into account the comprehensive benefits brought by the wind power energy storage heating project, and considers the excessive initial investment caused by the mismatch between supply and demand, and considers the operating costs, and optimizes the capacity configuration in a coordinated manner with supply and demand. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] Figure 1 A flow chart of an optimization method for a wind power hybrid energy storage heating system considering capacity configuration according to an embodiment of the present application;

[0095] Figure 2 A schematic diagram of the double-layer optimization model process of an embodiment of the present application;

[0096] Figure 3 The hourly heat load curve of the residence in the embodiment of the present application throughout the year;

[0097] Figure 4 Typical daily load curve of a residence in the embodiment of the present application;

[0098] Figure 5 Typical wind power output scenario 1 of the embodiment of the present application;

[0099] Figure 6 Typical wind power output scenario 2 of the embodiment of the present application;

[0100] Figure 7 Typical daily power balance of the embodiment of the present application;

[0101] Figure 8 Typical day-to-day thermal power balance of the embodiment of the present application;

[0102] Fig. 9 Typical daily two-electric power balance of the embodiment of the present application;

[0103] Fig.10 Typical day two heat power balance of the embodiment of the present application. DETAILED DESCRIPTION

[0104] The specific implementation of the present application is further described in detail below in conjunction with the drawings and examples.

[0105] Example:

[0106] This embodiment proposes an optimization method for a wind power hybrid energy storage heating system considering capacity configuration, such as Figure 1 As shown, including:

[0107] Step S1: Obtaining the annual residential heat load data and wind speed data of the area to be optimized;

[0108] In this embodiment, the annual heat load data of the northern residence and the wind speed data of the northern residence are obtained.

[0109] Step S2: For the annual residential heat load data of the area to be optimized, a k-means clustering analysis algorithm (i.e., a k-means clustering algorithm) is used to cluster and analyze the data by day to obtain the hourly heat load distribution data of a typical day;

[0110] In this embodiment, in order to improve the optimization efficiency and consider the capacity configuration problem, the k-means clustering analysis algorithm is used to cluster and analyze the annual heat load data of northern residential buildings by day to obtain the number of typical days and the heat load distribution data and characteristics of typical days by hour.

[0111] Step S3: Fit the wind speed data using a two-parameter Weibull model, sample the fitted wind speed data using the Monte Carlo method, obtain the wind power of a wind speed scenario each time, and reduce the wind power of the sampled wind speed scenario using a fast previous-generation scenario reduction technique based on probability distance, to obtain representative typical scenarios and wind power data in representative typical scenarios;

[0112] Step S4: Input the hourly heat load distribution data of a typical day and the wind power data in a representative typical scenario into a pre-established capacity configuration two-layer optimization model for iteration to obtain the optimal system capacity configuration plan; wherein, the upper model of the capacity configuration two-layer optimization model uses a particle swarm optimization algorithm to optimize the equipment capacity with the goal of maximizing the net benefit over the entire life cycle, and inputs the optimized equipment capacity into the lower model of the capacity configuration two-layer optimization model for iteration, and the lower model of the capacity configuration two-layer optimization model optimizes the hourly power of each device with the goal of minimizing the operating cost, and inputs the optimized hourly power and cost into the upper model of the capacity configuration two-layer optimization model for iteration.

[0113] In step S2, the annual residential heat load data of the area to be optimized is clustered and analyzed by day using the k-means clustering analysis algorithm to obtain hourly heat load distribution data of a typical day, including:

[0114] Step S2.1: For the annual residential heat load data of the area to be optimized, the k-means clustering analysis algorithm is used to cluster them by day;

[0115] The cluster analysis of this embodiment uses the k-means algorithm to classify the load characteristics of northern residences. Assume that there are m days of sample data in the heating period, n sample data per day, and the data set is represented by an m×n matrix X to represent the heat load of the residence.

[0116]

[0117] Step S2.2: Determine the optimal number of clusters according to the CH index, and the CH index calculation formula is as follows:

[0118]

[0119] Among them, CH(k) is the CH index with nk as the cluster number, B nk is the distance between classes, W nk is the intra-class distance, and N is the number of samples of the annual residential heat load data in the area to be optimized;

[0120] In this embodiment, since the k-means algorithm is sensitive to the selection of the initial cluster centroid and may fall into a local optimum, the Calinski-Haeabasz (CH) index is selected to determine the optimal number of clusters k, which is essentially the ratio of the distance between classes to the distance within classes. The higher the CH value, the better the clustering effect.

[0121] Step S2.3: Based on the optimal number of clusters determined by the CH index, cluster analysis is performed on the annual residential heat load data of the optimized area to obtain hourly heat load distribution data of a typical day, including:

[0122] Step S2.3.1: Read all objects in the annual residential heat load data of the area to be optimized, select k objects as initial cluster centers according to the optimal number of clusters determined by the CH index, and calculate the Euclidean distance between the remaining objects and the initial cluster centers;

[0123] Step S2.3.2: Divide all objects in the annual residential heat load data of the area to be optimized into the class with the closest distance, recalculate the mean of all objects in the class, use the mean as the new cluster center of the class, iterate until the convergence condition is reached, obtain the classification result of the typical day, and determine the number of typical days;

[0124] Step S2.3.3: Take the average of the days of all typical days to obtain the hourly heat load distribution data of the typical days.

[0125] In step S3, the two-parameter Weibull model is used to fit the wind speed data, and the fitted wind speed data is sampled using the Monte Carlo method. Each sampling obtains the wind power of a wind speed scenario, and the fast previous generation scenario reduction technology based on probability distance is used to reduce the wind power of the sampled wind speed scenario to obtain representative typical scenarios and wind power data in the representative typical scenarios, including:

[0126] Step S3.1: Use the two-parameter Weibull model to fit the wind speed data. The fitted wind speed probability density function is as follows:

[0127]

[0128] Where v is the wind speed, k is the shape parameter, c is the scale parameter, and f(v) is the wind speed probability density corresponding to wind speed v;

[0129] The two-parameter Weibull distribution model has a good fitting effect on wind speed distribution and has good adaptability to wind speed series of different time lengths (months / years).

[0130] Step S3.2: The wind speed data is sampled based on the wind speed probability density function using the Monte Carlo method, a wind speed scenario is obtained each time a sample is taken, and the wind power of the wind speed scenario is calculated. The calculation formula is as follows:

[0131]

[0132] Where P is the wind power in the wind speed scenario, P r is the rated power, v is the wind speed, v r is the rated wind speed, v ci is the cut-in wind speed, v co To cut out the wind speed;

[0133] Step S3.3: Use the fast previous scenario reduction technology based on probability distance to reduce the wind power of the sampled wind speed scenario, and obtain representative typical scenarios and wind power data in representative typical scenarios. The wind power data in representative typical scenarios is used as input data for the two-layer capacity configuration optimization model to reduce the complexity of the model and improve the optimization efficiency.

[0134] Based on the calculated wind power P, the wind power scenario is reduced, and the fast previous scenario reduction technology of probability distance is used to obtain representative typical scenarios. The probability distance uses Euclidean distance, and the calculation formula is as follows:

[0135]

[0136] Where d is the Euclidean distance; P i is the wind speed power of the i-th scenario; P j is the wind speed power of j scenarios.

[0137] The fast previous generation scenario reduction technology based on probability distance is used to reduce the wind power of the sampled wind speed scenario, and obtain representative typical scenarios and wind power data in the representative typical scenarios, including:

[0138] Step S3.3.1: Calculate the wind power P of any wind speed scenario in turn among all wind power sets P of the wind speed scenario set. i The Euclidean distance between the wind power of all other wind speed scenarios;

[0139] Step S3.3.2: Calculate the sum of the Euclidean distances between each wind speed scenario and the wind power of all other wind speed scenarios, find the wind speed scenario with the smallest sum of the Euclidean distances of wind power, and define it as a;

[0140] It can be understood that: for example, there are 100 wind speed scenarios in the wind speed scenario set, and there are 99 wind speed scenarios in addition to a. The Euclidean distances between the wind power of these 99 wind speed scenarios and the wind power of a are added. Each scenario is calculated once, and the scenario with the smallest sum of Euclidean distances is selected as a.

[0141] Step S3.3.3: Find the wind power of wind speed scenario b that is shortest in distance to the wind power of wind speed scenario a, replace the wind power of wind speed scenario a with the wind power of wind speed scenario b, add the wind speed probability density corresponding to the wind speed of a to the wind speed probability density corresponding to the wind speed of b, and form a new wind power set P' of the wind speed scenario set;

[0142] Step S3.3.4: Determine whether the current number of scenes is equal to the preset number; if not, repeat steps S3.3.1-S3.3.3; if equal, terminate scene reduction, take the current scene as the representative typical scene, and output the representative typical scene and the wind power data in the representative typical scene.

[0143] In step S4, a two-layer optimization model for capacity configuration is established. The model is solved by a particle swarm optimization algorithm. The upper layer model optimizes the equipment capacity of the system with the goal of maximizing the net benefit over the entire life cycle, and the lower layer model optimizes the hourly power of each device in the system with the goal of minimizing the operating cost.

[0144] Among them, the particle swarm optimization algorithm updates itself by iterating the individual extreme value of the particles and the global extreme value of the population to find the best solution. The particle speed and position update formula are as follows:

[0145]

[0146] Among them, ω is the inertia weight factor, c1 and c2 are learning factors, r1 and r2 are random numbers, k is the number of iterations, V i k is the velocity vector of particle i at the kth iteration, is the position vector of particle i at the kth iteration, P i k is the historical optimal position of particle i in the kth iteration, is the historical optimal position of the population at the kth iteration.

[0147] In the later stage of the search of the particle swarm algorithm, the positions of particles in the population lack diversity. Theoretically, this optimization algorithm based on swarm search needs to detect every point in the population to find the global optimal point, but this will result in a huge workload and is difficult to implement. Therefore, let the particles fly across the entire search space as much as possible at the beginning of the search to achieve diversity and avoid premature convergence to local extremes. To this end, you can choose to use a relatively large acceleration factor C1 and a relatively small C2 in the initial search, so that the particles can be more freely dispersed in the search space, thereby increasing the diversity of particles; as the number of iterations increases, let C1 decrease linearly and C2 increase linearly, then as the number of iterations increases, C1 becomes smaller and C2 increases, which enhances the ability of particles to converge to the global optimal.

[0148] The particle swarm optimization algorithm uses an acceleration factor for optimization during search, and the calculation formula is as follows:

[0149] C1=C 1s +k(C 1e -C 1s ) / k max

[0150] C2=C 2s +k(C 2e -C 2s ) / k max (8)

[0151] Among them, C1 is the individual acceleration factor, C2 is the social acceleration factor, and C1 is greater than C2 in the early stage of iteration. As the number of iterations increases, C2 is greater than C1, and C 1e is the final value of the individual acceleration factor C1, C 2e is the final value of the social acceleration factor C2, C 1s is the initial value of the individual acceleration factor C1, C 2s is the initial value of the social acceleration factor C2, k max is the maximum number of iterations.

[0152] The upper model of the capacity configuration double-layer optimization model constructs the objective function based on heating revenue, electricity sales revenue, coal saving revenue, environmental revenue, peak load regulation revenue, equipment investment cost, operation cost, maintenance cost and decommissioning cost. The calculation formula is as follows:

[0153]

[0154] Among them, f is the objective function value, K h is the heating benefit, K e is the revenue from electricity sales, K c For coal saving benefits, K t For environmental benefits, K psis the peak load benefit, C Inv is the investment cost, C O is the operating cost, C M is the maintenance cost, C D for decommissioning costs;

[0155] Among them, the heating revenue, the system meets the heating demand of northern residential buildings, and charges users according to the amount of heat supplied, the calculation formula is as follows:

[0156]

[0157] Among them, C h is the heating unit price; P h-load is the heat load, T is the number of hours in a day, i.e. 24 hours;

[0158] The revenue from electricity sales is that when the output of wind turbines and batteries cannot be absorbed, the electricity is sold to the power grid to obtain revenue. The calculation formula is as follows:

[0159]

[0160] in, P is the price of electricity sold to the grid, e,t is the interactive power with the power grid during period t. If it is greater than zero, it indicates power purchase, and if it is less than zero, it indicates power sale;

[0161] The coal-saving benefits mentioned above refer to the benefits of reducing coal consumption after clean heating replaces coal-fired boilers. The calculation formula is as follows:

[0162]

[0163] Among them, P s is the total power generation of the system, P e,t is the interaction power with the grid during period t, μ cfu is the coal consumption rate of the coal-fired boiler, C Coal is the price of coal;

[0164] The environmental benefits mentioned above refer to the benefits of reducing carbon dioxide emissions after replacing coal-fired boilers. At the same time, the pollution generated by coal-fired power generation when purchasing electricity from the power grid must be deducted. The calculation formula is as follows:

[0165]

[0166] Among them, P e,t is the interaction power with the power grid during period t; C t is the price per unit of CO2 emission;

[0167] The peak-shaving benefits mentioned above refer to the subsidies given according to local policies when the hybrid energy storage system participates in peak-shaving. The calculation formula is as follows:

[0168]

[0169] in, is the charging power of the battery during period t, is the heat storage power of the solid heat storage device during period t, C ps Subsidize the price of energy storage equipment for peak load regulation;

[0170] The investment cost, the annual equal investment cost is the total system investment cost equally distributed to each year in the useful life, and the calculation formula is as follows:

[0171]

[0172] Among them, x cap is the capacity of device i, C Inv,i is the unit capacity investment cost of equipment i, N i is the service life of device i, r is the discount rate, I is the total number of equipment;

[0173] The operating cost, system operating cost includes the operating cost of electricity storage, heat storage and heating equipment C s and electricity purchase cost C e In order to ensure the stability of energy supply during operation, the system needs to purchase electricity from the power grid when the output is insufficient. The calculation formula is as follows:

[0174] C O =C s +C e (16)

[0175]

[0176] Among them, C s,i is the unit operating cost of equipment i, is the power value of device i in period t, P is the price of electricity purchased from the power grid; e,t is the interaction power with the power grid during period t;

[0177] The maintenance cost, system maintenance cost is the labor and maintenance costs incurred to ensure the normal operation of each device in the system, and the calculation formula is as follows:

[0178]

[0179] Among them, x cap is the capacity of device i, C m,i is the unit capacity maintenance cost of equipment i, N i is the service life of device i, r is the discount rate.

[0180] The decommissioning cost, system decommissioning cost refers to the cost of handling the equipment after it reaches the end of its service life. However, the equipment has a certain residual value after retirement, so it should be the processing cost minus the residual value of the equipment. The calculation formula is as follows:

[0181]

[0182] Among them, a1 is the retired equipment disposal fee rate, a2 is the equipment residual value rate, x cap is the capacity of device i, C Inv,i is the unit capacity investment cost of equipment i, N i is the service life of device i, r is the discount rate.

[0183] The step of inputting the optimized equipment capacity into the lower model of the capacity configuration double-layer optimization model for iteration includes: inputting the hourly heat load distribution data of a typical day optimized by the particle swarm optimization algorithm and the wind power data in a representative typical scenario into the lower model of the capacity configuration double-layer optimization model to obtain a simplified operating cost, which is calculated as follows:

[0184] C' o =C' s +C' e (twenty one)

[0185]

[0186] Among them, d k is the number of days of typical day category k, k is the number of typical day categories, is the unit operating cost of equipment i under category k on a typical day, is the power value of device i in period t under typical day category k, P is the price of electricity purchased from the grid. e,t is the interaction power with the grid during period t, C' o For the simplified operating cost, C' s The simplified operating cost of the electricity storage, heat storage and heating equipment is C e ' is the simplified electricity purchase cost, I is the total number of equipment, and K is the total number of categories on a typical day. The results of the lower-level optimization model are affected by the upper-level decision variables. By optimizing the hourly power of system equipment, the operating cost of the system is minimized. The load data of a typical day and the wind power scenario after curtailment are used to represent the whole year and input into the lower-level model.

[0187] The pre-established capacity configuration two-layer optimization model constructs constraints including: electric power balance constraint, thermal power balance constraint, wind power output constraint, energy storage equipment constraint, and heating equipment constraint;

[0188] The electric power balance constraint is calculated as follows:

[0189]

[0190] Among them, P Wt is the wind power generation power, P B is the battery charging and discharging power, P e is the grid interaction power, Input power for heat pump heating;

[0191] The thermal power balance constraint is calculated as follows:

[0192]

[0193] in, is the heating output power of the heat pump, P Hs is the charging and discharging power of the solid heat storage device, P h-load is the heat load demand; the wind power output constraint is calculated as follows:

[0194] 0≤P Wt (t)≤P Wt,max (26)

[0195] Among them, P Wt is the wind power generation power, P Wt,max The maximum power of wind power generation.

[0196] The energy storage device constraint is calculated as follows:

[0197]

[0198] Among them, P i in (t) represents the electrical / thermal storage power at time t, Indicates the maximum electrical / heat storage power, P i out (t) represents the discharge / heat release power at time t, represents the maximum discharge / heat release power, X and Y are 0-1 variables of the energy storage device’s charge and discharge state; the energy storage device’s state of charge SOC constraint is as follows:

[0199] S SOCmin ≤S SOC (t)≤S SOCmax (28)

[0200] Among them, S SOC (t) is the battery charge state at time t, S SOCmin is the upper limit of the state of charge, S SOCmax It is the lower limit of the charge state;

[0201] The heating equipment constraint is calculated as follows:

[0202] Heating power constraints of water source heat pump:

[0203]

[0204] Among them, η Hp is the heating efficiency of the heat pump, represents the input power of the heat pump at time t, P Hp,max Indicates the maximum heating power of the heat pump.

[0205] In step S4, the data processed in steps S2 and S3 are input into the two-layer capacity configuration optimization model established in step 4. Through bidirectional coupling between the upper and lower layers and nested iteration, the optimal system capacity configuration scheme is obtained as follows: Figure 2 As shown. The upper optimization configuration model passes the capacity plans of each device initialized by the particle swarm algorithm to the lower optimization operation model. The lower model inputs the heat load data obtained in step S2 and the wind power data obtained in step S3, and uses the CPLEX solver (CPLEX solver, IBM ILOG Concert Technology, is an operations research optimization solver provided by IBM ILOG) for operation optimization, and passes the optimized operation power and cost results of each device to the upper layer; then, the upper model corrects the system net benefit objective function value according to the feedback results of the lower layer, and generates a new equipment capacity plan, and repeats the above process; finally, through the cyclic coupling iteration between the two-layer models, the optimal system capacity configuration plan is obtained.

[0206] A residential complex in Liaoning was selected as the research object. The heating period of this residential complex is 151 days (November 1 to March 31), and the total construction area is 14515m 2 The annual cumulative heat load is 1073625.65kW·h, and the annual cumulative heat load index is 73.97kW·h / m 2 The maximum heat load of the whole year is 969.49kW. The hourly heat load curve of the whole year is as follows Figure 3 shown.

[0207] It is calculated that when the number of clusters k of the residential cluster is 2, the CH index is the largest, that is, the optimal number of clusters is 2. The heat load clustering results are shown in Table 1. The typical daily load curve is as follows: Figure 4 Its data is used as the input data of the two-layer capacity configuration optimization model.

[0208] Table 1 Heat load clustering results

[0209]

[0210] Combining Table 1 and Figure 4The clustering results show that during the heating period, the number of typical days of heat load in the residential complex is the largest, totaling 84 days, with a large heat load, mainly concentrated in the three months from December to February when the average temperature is below 0℃. The number of days in the second category is relatively small, totaling 67 days, with a small heat load, mainly concentrated in November and March when the temperature is higher in winter. The heat load on a typical day generally reaches its highest at around 7 am and is the lowest at around 13 pm. The heat load increases with time at night. This is because the outdoor temperature is low at night in winter and people are all indoors, while the outdoor temperature is high during the day, the indoor rate of people is low, and the heat disturbance of people and equipment is small, so the heat load is low.

[0211] Based on the wind speed data measured every 15 minutes at a wind tower in a wind farm in northern Shenyang, the wind speed resources in the area are evaluated. According to the wind speed data, the Monte Carlo method is used to sample the probability distribution model of wind speed based on the two-parameter Weibull model to obtain the predicted wind speed scenario, and the wind speed-wind power conversion formula is used to generate the wind power output scenario. Then, a representative typical scenario is obtained using the rapid previous generation scenario reduction technology based on probability distance. The present invention is set to generate 1,000 scenarios, the number of scenario reductions is 2, and the typical wind power output scenario 1 is as follows. Figure 5 As shown in Figure 2, typical wind power output scenario 2 is as follows Figure 6 The data is used as the input data of the two-layer capacity configuration optimization model.

[0212] The present invention optimizes the capacity configuration of the system's battery, solid heat storage unit and water source heat pump, and the PSO algorithm parameters are set as follows: the learning factor is C1=2; C2=1, and the maximum inertia weight coefficient ω max =0.9, minimum inertia weight coefficient ω min =0.4, the number of iterations k is 100, and the number of particle populations is 50.

[0213] The technical and economic parameters of each equipment are shown in Table 2. The retired equipment treatment fee rate a1 is 3.5%, the equipment residual value rate a2 is 3%, the discount rate r is 6.7%, and the heating unit price C h is 0.181 yuan / (kW·h), and the coal consumption rate of the coal-fired boiler is μ cf u is 0.153kg / (kW·h), coal price C Coal The unit emission price of CO2 is 0.5 yuan / kg. t The electricity price from the power grid is 0.365 yuan / (kW·h). The price of electricity sold to the grid is 0.519 yuan / (kW·h). The unit subsidy price of energy storage equipment participating in peak load regulation is C ps The heat load data, wind power data and all parameters are brought into the capacity configuration double-layer optimization model for solution.

[0214] To verify the effectiveness of the method proposed in the present invention and analyze the economic benefits brought by hybrid energy storage to the system, three schemes are proposed for comparative analysis. 1) Scheme 1: using the method of the present invention to configure batteries + solid thermal storage, 2) Scheme 2: only configure batteries, 3) Scheme 3: only configure solid thermal storage, the capacity configuration optimization of each scheme is shown in Table 3.

[0215] Table 2: Technical and economic parameters of each equipment

[0216]

[0217] Table 3: Capacity configuration optimization of each solution

[0218]

[0219]

[0220] As shown in Table 3, the net income of the hybrid energy storage scheme proposed in the present invention is the highest, which is about 22.08% higher than that of Scheme 2 and about 14.10% higher than that of Scheme 3. It also has the highest environmental income and performs better in reducing carbon emissions. The configuration capacity of the water source heat pump is the smallest, which reduces the investment in heating equipment. Scheme 2 is configured with a larger capacity battery, so the electricity sales income and peak-shaving income are slightly higher than those of Scheme 1, but its various costs are the highest among the three schemes, and the total cost is 45% higher than that of Scheme 1, of which the operating cost is 149.4% higher than that of Scheme 1, and the investment cost is 15.9% higher than that of Scheme 1. The various costs of Scheme 3 are the lowest among the three schemes, but the total income is 13.7% lower than that of Scheme 1 and 10.7% lower than that of Scheme 2. Its various incomes are lower than those of the other two schemes. In summary, the hybrid energy storage system of the method proposed in the present invention has the highest comprehensive economic benefit.

[0221] Typical day operation optimization results:

[0222] The lower optimization operation model is solved using the CPLEX solver. The electrical power and thermal power balances of typical day 1 and typical day 2 after optimization are as follows: Figure 7-10 shown.

[0223] Depend on Figure 7-10As shown in Table 4, the net benefits of the system on different typical days vary greatly. The net benefit on the first typical day is about 108.7% higher than that on the second typical day. Heating and environmental benefits are the main reasons for the difference. This is mainly because the heat load on the first typical day is large, so the heating benefit and the environmental benefit brought by reducing carbon emissions after using the clean heating system of the present invention to replace coal-fired boilers are higher. The average clean energy utilization rate of the system is 99.27%. Through the coordination of wind turbines, batteries, solid heat storage devices, water source heat pumps and other equipment in the system, as well as power trading with the power grid, the balance of electric heat supply and demand in the system can be achieved, wind power consumption and load demand can be guaranteed, and clean heating of northern residences can be achieved.

[0224] Table 4 Typical day system optimization operation results

[0225]

[0226] The various embodiments in the present application are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0227] The protection scope of the present application is not limited to the above-mentioned embodiments. Obviously, those skilled in the art can make various changes and modifications to the present disclosure without departing from the scope and spirit of the present disclosure. If these changes and modifications fall within the scope of the claims of the present disclosure and their equivalents, the intention of the present disclosure also includes these changes and modifications.

Claims

1. An optimization method for a wind power hybrid energy storage heating system considering capacity configuration, characterized in that: include: Step S1: Obtaining the annual residential heat load data and wind speed data of the area to be optimized; Step S2: For the annual residential heat load data of the area to be optimized, the k-means clustering analysis algorithm is used to cluster and analyze by day to obtain the hourly heat load distribution data of a typical day; Step S3: Fit the wind speed data using a two-parameter Weibull model, sample the fitted wind speed data using the Monte Carlo method, obtain the wind power of a wind speed scenario each time, and reduce the wind power of the sampled wind speed scenario using a fast previous scenario reduction technique based on probability distance, to obtain representative typical scenarios and wind power data in representative typical scenarios, including: Step S3.1: Use the two-parameter Weibull model to fit the wind speed data. The fitted wind speed probability density function is as follows: Where v is the wind speed, q is the shape parameter, c is the scale parameter, and f(v) is the wind speed probability density corresponding to the wind speed v; Step S3.2: The wind speed data is sampled based on the wind speed probability density function using the Monte Carlo method, a wind speed scenario is obtained each time a sample is taken, and the wind power of the wind speed scenario is calculated. The calculation formula is as follows: Where P is the wind power in the wind speed scenario, P r is the rated power, v is the wind speed, v r is the rated wind speed, v ci is the cut-in wind speed, v co To cut out the wind speed; Step S3.3: Use the fast previous generation scenario reduction technology based on probability distance to reduce the wind power of the sampled wind speed scenario, and obtain representative typical scenarios and wind power data in representative typical scenarios, including: Step S3.3.1: in all wind power sets P of the wind speed scenario set, calculate the Euclidean distance between the wind power of any wind speed scenario and the wind power of all other wind speed scenarios in turn; Step S3.3.2: Calculate the sum of the Euclidean distances between each wind speed scenario and the wind power of all other wind speed scenarios, find the wind speed scenario with the smallest sum of the Euclidean distances of wind power, and define it as a; Step S3.3.3: Find the wind power of wind speed scenario b that has the shortest distance to the wind power of wind speed scenario a, replace the wind power of wind speed scenario a with the wind power of wind speed scenario b, add the wind speed probability density corresponding to the wind speed of a to the wind speed probability density corresponding to the wind speed of b, and form a new wind power set of wind speed scenario sets; Step S3.3.4: Determine whether the current number of scenarios is equal to the preset number; if not, repeat steps S3.3.1-S3.3.3; if equal, terminate scenario reduction, take the current scenario as a representative typical scenario, and output representative typical scenarios and wind power data in representative typical scenarios; Step S4: Input the hourly heat load distribution data of a typical day and the wind power data in representative typical scenarios into the pre-established capacity configuration double-layer optimization model for iteration to obtain the optimal system capacity configuration plan; wherein, the upper model of the capacity configuration double-layer optimization model uses a particle swarm optimization algorithm to optimize the equipment capacity with the goal of maximizing the net benefit over the entire life cycle, and inputs the optimized equipment capacity into the lower model of the capacity configuration double-layer optimization model for iteration, and the lower model of the capacity configuration double-layer optimization model optimizes the hourly power of each device with the goal of minimizing the operating cost, and inputs the optimized hourly power and cost into the upper model of the capacity configuration double-layer optimization model for iteration; The upper model of the capacity configuration double-layer optimization model constructs the objective function based on heating revenue, electricity sales revenue, coal saving revenue, environmental revenue, peak load regulation revenue, equipment investment cost, operation cost, maintenance cost and decommissioning cost. The calculation formula is as follows: f=K h +K e +K c +K t +K ps -(C Inv +C O +C M +C D ) Among them, f is the objective function value, K h is the heating benefit, K e is the revenue from electricity sales, K c For coal saving benefits, K t For environmental benefits, K ps is the peak load benefit, C Inv is the investment cost, C O is the operating cost, C M is the maintenance cost, C D for decommissioning costs; The heating benefit is calculated as follows: Among them, C h is the heating unit price; P h-load is the heat load, T is the number of hours in a day, i.e. 24 hours; The electricity sales revenue is calculated as follows: in, P is the price of electricity sold to the grid. e,t is the interactive power with the power grid during period t. If it is greater than zero, it indicates power purchase, and if it is less than zero, it indicates power sale; The coal saving benefit is calculated as follows: Among them, P s is the total power generation of the system, P e,t is the interaction power with the grid during period t, μ cfu is the coal consumption rate of the coal-fired boiler, C Coal is the price of coal; The environmental benefits are calculated as follows: Among them, P e,t is the interaction power with the power grid during period t; C t is the price per unit of CO2 emission; The peak load benefit is calculated as follows: in, is the charging power of the battery during period t, is the heat storage power of the solid heat storage device during period t, C ps Subsidize the price of energy storage equipment for peak load regulation; The investment cost is calculated as follows: Among them, x cap is the capacity of device i, C Inv,i is the unit capacity investment cost of equipment i, N i is the useful life of equipment i, r is the discount rate, and I is the total number of equipment; The operating cost is calculated as follows: C O =C s +C e Among them, C s,i is the unit operating cost of equipment i, is the power value of device i in period t, P is the price of electricity purchased from the power grid; e,t is the interaction power with the power grid during period t; The maintenance cost is calculated as follows: Among them, x cap is the capacity of device i, C m,i is the unit capacity maintenance cost of equipment i, N i is the useful life of equipment i, r is the discount rate; The decommissioning cost is calculated as follows: Among them, a1 is the retired equipment disposal fee rate, a2 is the equipment residual value rate, x cap is the capacity of device i, C Inv,i is the unit capacity investment cost of equipment i, N i is the useful life of equipment i, r is the discount rate; The step of inputting the optimized equipment capacity into the lower model of the capacity configuration double-layer optimization model for iteration includes: inputting the hourly heat load distribution data of a typical day optimized by the particle swarm optimization algorithm and the wind power data in a representative typical scenario into the lower model of the capacity configuration double-layer optimization model to obtain a simplified operating cost, which is calculated as follows: C' o =C′ s +C′ e Among them, d k is the number of days of typical day category k, k is the number of typical day categories, is the unit operating cost of equipment i under category k on a typical day, is the power value of device i in period t under typical day category k, P is the price of electricity purchased from the power grid. e,t is the interaction power with the grid during period t, C' o is the simplified operating cost, C′ s C′ is the simplified operating cost of the electricity storage, heat storage and heating equipment. e The simplified electricity purchase cost is shown in Figure 2. I is the total number of equipment and K is the total number of categories on a typical day.

2. The optimization method of the wind power hybrid energy storage heating system considering capacity configuration according to claim 1 is characterized in that: The annual residential heat load data of the area to be optimized is clustered and analyzed by day using the k-means clustering analysis algorithm to obtain hourly heat load distribution data of a typical day, including: Step S2.1: For the annual residential heat load data of the area to be optimized, the k-means clustering analysis algorithm is used to cluster them by day; Step S2.2: Determine the optimal number of clusters according to the CH index, and the CH index calculation formula is as follows: Among them, CH(nk) is the CH index with nk as the cluster number, B nk is the distance between classes, W nk is the intra-class distance, and N is the number of samples of the annual residential heat load data in the area to be optimized; Step S2.3: Based on the optimal number of clusters determined by the CH index, cluster analysis is performed on the annual residential heat load data of the optimized area to obtain hourly heat load distribution data of a typical day.

3. The optimization method of the wind power hybrid energy storage heating system considering capacity configuration according to claim 2 is characterized in that: According to the optimal cluster number determined by the CH index, cluster analysis is performed on the annual residential heat load data of the optimized area to obtain hourly heat load distribution data of a typical day, including: Step S2.3.1: Read all objects in the annual residential heat load data of the area to be optimized, select k objects as initial cluster centers according to the optimal number of clusters determined by the CH index, and calculate the Euclidean distance between the remaining objects and the initial cluster centers; Step S2.3.2: Divide all objects in the annual residential heat load data of the area to be optimized into the class with the closest distance, recalculate the mean of all objects in the class, use the mean as the new cluster center of the class, iterate until the convergence condition is reached, obtain the classification result of the typical day, and determine the number of typical days; Step S2.3.3: Take the average of the days of all typical days to obtain the hourly heat load distribution data of the typical days.

4. The optimization method of the wind power hybrid energy storage heating system considering capacity configuration according to claim 1 is characterized in that: The particle swarm optimization algorithm uses an acceleration factor for optimization during search, and the calculation formula is as follows: C1=C 1s +k(C 1e -C 1s ) / k max C2=C 2s +k(C 2e -C 2s ) / k max Among them, C1 is the individual acceleration factor, C2 is the social acceleration factor, and C 1e is the final value of the individual acceleration factor C1, C 2e is the final value of the social acceleration factor C2, C 1s is the initial value of the individual acceleration factor C1, C 2s is the initial value of the social acceleration factor C2, k max is the maximum number of iterations.

5. The optimization method of the wind power hybrid energy storage heating system considering capacity configuration according to claim 1 is characterized in that: The pre-established capacity configuration two-layer optimization model constructs constraints including: electric power balance constraint, thermal power balance constraint, wind power output constraint, energy storage equipment constraint, and heating equipment constraint; The electric power balance constraint is calculated as follows: Among them, P Wt is the wind power generation power, P B is the battery charging and discharging power, P e is the grid interaction power, Input power for heat pump heating; The thermal power balance constraint is calculated as follows: in, is the heating output power of the heat pump, P Hs is the charging and discharging power of the solid heat storage device, P h-load is the heat load demand; The wind power output constraint is calculated as follows: 0≤P Wt (t)≤P Wt,max Among them, P Wt is the wind power generation power, P Wt,max is the maximum power of wind power generation; The energy storage device constraint is calculated as follows: Among them, P i in (t) represents the electrical / thermal storage power at time t, Indicates the maximum electrical / heat storage power, P i out (t) represents the discharge / heat release power at time t, represents the maximum discharge / heat release power, X and Y are 0-1 variables of the energy storage device’s charge and discharge state; the energy storage device’s state of charge SOC constraint is as follows: S SOCmin ≤S SOC (t)≤S SOCmax Among them, S SOC (t) is the battery charge state at time t, S SOCmin is the upper limit of the state of charge, S SOCmax It is the lower limit of the charge state; The heating equipment constraint is calculated as follows: Heating power constraints of water source heat pump: Among them, η Hp is the heating efficiency of the heat pump, represents the input power of the heat pump at time t, P Hp,max Indicates the maximum heating power of the heat pump.

6. The optimization method of the wind power hybrid energy storage heating system considering capacity configuration according to claim 1 is characterized in that: The lower model of the capacity configuration double-layer optimization model optimizes the hourly power of each device with the goal of minimizing the operating cost, and inputs the optimized hourly power and cost into the upper model of the capacity configuration double-layer optimization model for iteration, including: inputting the hourly heat load distribution data of a typical day and the wind power data in a representative typical scenario into the lower model of the capacity configuration double-layer optimization model, using the CPLEX solver for operation optimization, obtaining the optimized hourly power and cost, and inputting the optimized hourly power and cost into the upper model of the capacity configuration double-layer optimization model for iteration.

Images