Scheduling method for integrated energy system including electricity-to-gas conversion considering classified load demand response

Abstract

An embodiment of the present invention provides a method for dispatching an integrated energy system including power-to-gas conversion that takes into account the response to classified load demands, the method comprising: obtaining a wind power curve that takes into account the uncertainty of wind power output; clustering user load characteristics by an improved K-means clustering method to obtain different types of user load characteristics; determining the response to classified load demands based on the different types of user load characteristics; constructing an optimization model for an integrated energy system including power-to-gas conversion equipment that takes into account the response to classified load demands; solving the optimization model for the integrated energy system based on the wind power curve that takes into account the uncertainty of wind power output, different types of user load characteristics, and energy quotations provided by energy suppliers to obtain an optimized operation plan for the integrated energy system. The method for dispatching an integrated energy system including power-to-gas conversion that takes into account the response to classified load demands of the present invention can effectively utilize load-side flexibility resources and reduce the operating costs of the integrated energy system in the park.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of optimized scheduling of integrated energy systems, and in particular to a scheduling method for an integrated energy system including electricity-to-gas conversion taking into account classified load demand response. Background Art

[0002] As a natural extension of microgrids, the integrated energy system integrates multiple types of distributed energy and traditional energy such as cold, heat, electricity, and gas, and can achieve efficient cascade utilization of energy. It has become an important means of green, efficient and sustainable utilization of energy today. The park's integrated energy system is a multi-dimensional energy network system that uses distributed energy producers and consumers as nodes, electricity, cold and heat energy networks as links, and connects energy production, transmission, distribution, storage and use. Today, a single power system transaction can no longer meet the increasingly complex energy needs. The park's integrated energy system is increasingly participating in the multi-energy market, and the competition is becoming increasingly fierce.

[0003] As a multi-energy cogeneration system, it requires complex system coordination and energy regulation. Therefore, the operation and management of the park's integrated energy system is crucial to the realization of the system's comprehensive benefits. At present, the main purpose of the research on the operation and management of the park's integrated energy system is to maximize the system's benefits and value on the premise of matching the system's energy supply with user needs. Therefore, it is necessary to comprehensively consider energy and demand to achieve overall optimization. However, at present, the operation cost of the integrated energy system dispatch has not been fully considered for the operation and market competitiveness of the park's integrated energy system, resulting in a high operation cost for the integrated energy system dispatch. Summary of the invention

[0004] Based on the above technical problems, an embodiment of the present invention provides a scheduling method for an integrated energy system including electricity-to-gas conversion taking into account classified load demand response, so as to reduce the operating cost of the integrated energy system.

[0005] An embodiment of the present invention provides a method for dispatching an integrated energy system including electricity-to-gas taking into account classified load demand response, the method comprising:

[0006] Obtain wind power curves that take into account wind power output uncertainty;

[0007] The user load characteristics are clustered by the improved K-means clustering method to obtain different types of user load characteristics;

[0008] Determining classified load demand responses according to the load characteristics of the different types of users;

[0009] Constructing a comprehensive energy system optimization model including power-to-gas equipment considering the classified load demand response;

[0010] According to the wind power curve taking into account the uncertainty of wind power output, the load characteristics of different types of users, and the energy quotations provided by energy suppliers, the integrated energy system optimization model is solved to obtain an optimized operation plan for the integrated energy system.

[0011] Optionally, clustering the user load characteristics by an improved K-means clustering method to obtain different types of user load characteristics includes:

[0012] The DTW similarity distance index is used to replace the distance index in the original K-means clustering method to obtain the improved K-means clustering method;

[0013] The user load characteristics are processed by the improved K-means clustering method to obtain the similarity between the user load characteristics;

[0014] According to the similarity between the user load characteristics, the user load characteristics are classified to obtain different types of user loads with similar energy consumption characteristics, and the user load characteristics of the different types are determined.

[0015] Optionally, determining the classified load demand response according to the load characteristics of different types of users includes:

[0016] Determining corresponding time-of-use energy prices according to the load characteristics of different types of users;

[0017] Determining the participation of different types of users in demand response according to the time-of-use energy price;

[0018] The classified load demand response is determined according to the different types of user demand response participation, the price elasticity matrix and the corresponding time-of-use energy price.

[0019] Optionally, obtaining a wind power curve taking into account wind power output uncertainty includes:

[0020] The original scenario set of wind turbine power generation that obeys the probability distribution is generated by the Latin hypercube sampling method;

[0021] The original scene set is reduced by a synchronous back-substitution reduction method based on probability distance to obtain a streamlined scene set;

[0022] Performing a multiplication operation on the probability of occurrence of each scenario in the simplified scenario set and the corresponding wind turbine generator set power generation to obtain a product result;

[0023] The product result is determined as the wind power curve taking into account the uncertainty of wind power output.

[0024] Optionally, the user load characteristic is obtained by the following steps:

[0025] Obtain typical daily user load data;

[0026] The LSTM autoencoder is used to extract features of the typical day user load data to obtain the user load features.

[0027] Optionally, the step of constructing a comprehensive energy system optimization model including power-to-gas equipment that takes into account the classified load demand response includes:

[0028] Constructing an energy supplier optimization model with the goal of maximizing the revenue of the energy supplier;

[0029] Constructing an internal optimization model of the integrated energy system including power-to-gas equipment with the goal of minimizing the cost of the integrated energy system and considering the classified load demand response;

[0030] The integrated energy system optimization model is constructed according to the energy supplier optimization model and the internal optimization model of the integrated energy system.

[0031] Optionally, the constructing of an energy supplier optimization model with the goal of maximizing the revenue of the energy supplier includes:

[0032] The energy supplier optimization model is constructed by taking the maximum energy sales revenue of the energy supplier as the first objective function and taking energy sales constraints and energy interaction constraints as the first constraint conditions.

[0033] Optionally, the constructing of an internal optimization model of an integrated energy system including power-to-gas equipment with the goal of minimizing the cost of the integrated energy system and taking into account the classified load demand response comprises:

[0034] The internal optimization model of the integrated energy system is constructed by taking the minimization of the operating cost of the integrated energy system as the second objective function and taking the energy balance constraint and the unit constraint as the second constraint conditions.

[0035] Optionally, the method further includes:

[0036] Determining the user's energy load demand according to the classified load demand response;

[0037] Determining the energy balance constraint according to the user energy load demand and equipment parameters;

[0038] Determining the maximum power generation power of the wind turbine generator set according to the wind power curve taking into account the uncertainty of wind power output;

[0039] The wind turbine group constraints are determined according to the maximum power generation power of the wind turbine group and the equipment parameters.

[0040] Optionally, solving the integrated energy system optimization model to obtain an optimized operation plan for the integrated energy system includes:

[0041] The comprehensive energy system optimization model is solved by integrating the multiverse optimization algorithm and the mixed integer linear programming method to obtain the optimized operation plan, which at least includes: a plan for purchasing energy from the energy supplier and a time-of-use pricing plan for the different types of loads.

[0042] In the method of the embodiment of the present invention, first, a wind power curve taking into account the uncertainty of wind power output is obtained; secondly, user load characteristics are clustered by an improved K-means clustering method to obtain different types of user load characteristics; then, classified load demand response is determined according to different types of user load characteristics; thirdly, an integrated energy system optimization model including power-to-gas equipment considering classified load demand response is constructed; finally, the integrated energy system optimization model is solved according to the wind power curve taking into account the uncertainty of wind power output, different types of user load characteristics, and energy quotations provided by energy suppliers, and an optimized operation plan of the integrated energy system is obtained. Through the method of the embodiment of the present method, the classified demand response and power-to-gas device are considered when scheduling the integrated energy system, and the energy utilization rate of the equipment is optimized by the classified load demand response, guiding users to optimize energy use plans and realize peak shaving and valley filling. The addition of power-to-gas improves the absorption of renewable energy and reduces the phenomenon of wind abandonment, which can better improve the economic efficiency of the operation of the park's integrated energy system, thereby effectively utilizing the load-side flexibility resources and reducing the operating cost of the park's integrated energy system. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0044] Figure 1 It is a flow chart of a method for dispatching an integrated energy system including power-to-gas conversion taking into account classified load demand response according to an embodiment of the present invention;

[0045] Figure 2 is a schematic diagram of electric load optimization after classified load response shown in one embodiment of the present invention;

[0046] Figure 3 is a wind power curve diagram taking into account the uncertainty of wind turbine output according to an embodiment of the present invention;

[0047] Figure 4 It is a flow chart showing a solution of an integrated energy system optimization model including power-to-gas equipment taking into account classified load demand response according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] Reference Figure 1 , Figure 1 FIG. 1 is a flow chart of a method for dispatching an integrated energy system including power-to-gas and taking into account classified load demand response according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0050] Step S11: Obtaining a wind power curve taking into account the uncertainty of wind power output.

[0051] In this embodiment, the uncertainty of wind turbine output in the park integrated energy system can be firstly determined based on historical data to determine a wind power curve that takes into account the uncertainty of wind power output. The wind power curve that takes into account the uncertainty of wind power output in this embodiment is the predicted power generation curve of the wind turbine in each period within 24 hours.

[0052] Step S12: clustering the user load characteristics by using an improved K-means clustering method to obtain different types of user load characteristics.

[0053] In this embodiment, the traditional original K-means clustering method can be improved, and the extracted user load characteristics can be clustered by the improved K-means clustering method, so as to obtain different types of user load characteristics, that is, the characteristics of different loads. Among them, the user load characteristics are the load characteristics extracted based on the historical user load data; in this embodiment, different types of energy consumption characteristics have their own certain load forms, so the user load characteristics are the load characteristics unique to each type of energy user.

[0054] Step S13: Determine the classified load demand response according to the load characteristics of the different types of users.

[0055] In this embodiment, after determining the user load characteristics corresponding to each type of load, a corresponding time-sharing demand response strategy can be formulated for the user load characteristics of each type of load and its expected load curve shape. Among them, the time-sharing demand response strategy can be a formulated energy price for each time period. In this way, the demand response of each type of load can be determined according to the time-sharing demand response strategy formulated for each type of load, thereby determining the classified load demand response. Among them, the classified load demand response in this embodiment refers to the load demand response generated by users with different energy characteristics for the formulated price.

[0056] Step S14: constructing a comprehensive energy system optimization model including power-to-gas equipment that takes into account the classified load demand response.

[0057] In this embodiment, an integrated energy system optimization model that takes into account classified load demand response and power-to-gas equipment can be constructed to determine the operation plan formulated by the integrated energy system to achieve optimization of the integrated energy system scheduling.

[0058] Step S15: Solve the integrated energy system optimization model based on the wind power curve taking into account the uncertainty of wind power output, the load characteristics of different types of users, and the energy quotations provided by energy suppliers to obtain an optimized operation plan for the integrated energy system.

[0059] In this embodiment, the input of the integrated energy system optimization model may include at least: a wind power curve that takes into account the uncertainty of wind power output (i.e., the power generation power of the wind turbine in each period within 24 hours), load characteristics of different types of users, and energy quotations provided by energy suppliers; the output of the integrated energy system optimization model may include at least: an optimized operation plan for the integrated energy system.

[0060] That is to say, this embodiment can solve the constructed integrated energy system optimization model containing power-to-gas equipment that takes into account the classified load demand response based on the wind power curve that takes into account the uncertainty of wind power output, the load characteristics of different types of users, and the energy quotations provided by energy suppliers, so as to obtain an optimized operation plan for the integrated energy system.

[0061] In this embodiment, not only is the energy utilization rate of the equipment optimized through classified load demand response, achieving peak shaving and valley filling of the load curve, but the addition of power-to-gas also improves the absorption of renewable energy and reduces wind curtailment, which can better improve the economic efficiency of the operation of the park's integrated energy system, thereby effectively utilizing load-side flexibility resources, improving the operational flexibility of the integrated energy system, and reducing the operating costs of the park's integrated energy system.

[0062] In combination with the above embodiments, in one implementation, the present invention further provides a method for dispatching an integrated energy system including electricity-to-gas taking into account classified load demand response. In the method, the above step S12 may specifically include steps S21 to S23:

[0063] Step S21: using the DTW similarity distance indicator to replace the distance indicator in the original K-means clustering method to obtain the improved K-means clustering method.

[0064] In this embodiment, when clustering user load characteristics, the traditional original K-means clustering method is no longer used, but the distance index in the original K-means clustering method is improved. In this embodiment, the DTW similarity distance index is used to replace the distance index in the original K-means clustering method. Specifically, the DTW similarity distance can be determined by the following formulas (1) to (4):

[0065]

[0066]

[0067]

[0068] DWT'(X',Y')=L'(n,m) (4)

[0069] Among them, formula (1) is the trend sequence of the original load sequence x, y, which can fully reflect the trend characteristic information of each time period; formula (2) is the Euclidean distance of the trend sequence; formula (3) is the cumulative cost matrix L' of the DTW distance after improving the bending times, which is used to calculate the sequence x i ′ and y j ′ trend characteristics. x 、r y and r max They represent the x sequence spacing, y sequence spacing and maximum spacing respectively. L(i-1,j) represents x i- ′1 and y j ′The subsequence accumulation distance when matching, that is, x i- ′1 and y j ′, L(i,j-1) represents the similarity between x i ′ and y j ' -1 The accumulated distance of subsequences during matching, L(i-1,j-1) represents x i- ′1 and y j ' -1 The subsequence accumulation distance during matching; Formula (4) is the improved DTW distance, which describes the trend characteristics of the sequence.

[0070] In this embodiment, after the DTW distance is determined, the DTW distance can be used to replace the distance in the original K-means clustering method, thereby obtaining an improved K-means clustering method.

[0071] Step S22: Processing the user load features by using the improved K-means clustering method to obtain the similarity between the user load features.

[0072] In this embodiment, the extracted user load features can be processed by the improved K-means clustering method to obtain the similarity between the user load features. Specifically, the similarity between the user load features can be determined by the following formula (5):

[0073]

[0074] Among them, in formula (5): D(X,Y) is the similarity between the X load sequence and the Y load sequence, α and λ are the numerical feature weight and trend feature weight respectively. When this distance is applied to sequence clustering analysis, it can not only compare the values ​​of the two sequences, but also well reflect the trend characteristics of the two. The improved K-means clustering method provided by this embodiment can effectively take into account the numerical characteristics and trend characteristics of the sequence, has a strong recognition ability for the trend feature details of the sequence, and has strong stability.

[0075] Step S23: classifying the user load characteristics according to the similarity between the user load characteristics, obtaining different types of user loads with similar energy consumption characteristics, and determining the user load characteristics of the different types.

[0076] In this embodiment, after obtaining the similarity between user load characteristics, the extracted user load characteristics can be classified according to the similarity between the user load characteristics to obtain different types of user loads, and each type of power load has similar energy characteristics. For example, the classification results of different types of electric loads and thermal loads can be obtained, such as clustering to obtain several types of power loads P1, P2, ..., P n and several types of heat loads H1, H2, ..., H n etc.

[0077] In this way, this embodiment can determine the user load characteristics of each type of user load based on the different types of user loads obtained. The user load characteristics may include user energy consumption characteristics, user energy consumption psychology, and the load curve shape expected by the user, thereby determining different types of user load characteristics.

[0078] In this embodiment, the DTW similarity distance is used to replace the distance in the original K-means clustering method, and the traditional original K-means clustering method is improved, so that when performing cluster analysis of different user loads, both the numerical values ​​of the two can be compared, and the trend characteristics of the two can be well reflected. That is, it can effectively take into account the numerical characteristics and trend characteristics of the sequence, has a strong recognition ability for the trend characteristic details of the sequence, and has strong stability.

[0079] In combination with the above embodiments, in one implementation, the present invention further provides a method for dispatching an integrated energy system including electricity-to-gas taking into account classified load demand response. In the method, the above step S13 may specifically include steps S31 to S33:

[0080] Step S31: Determine the corresponding time-of-use energy price according to the load characteristics of different types of users.

[0081] In this embodiment, after obtaining the load characteristics of different types of users, the corresponding time-of-use energy prices can be determined according to the load characteristics of different types of users. Specifically, the corresponding time-of-use energy prices can be formulated for the user load characteristics of each type of load and its expected load curve shape. For example, for various types of user loads belonging to electric loads, corresponding time-of-use electricity prices can be formulated; for various types of user loads belonging to thermal loads, corresponding time-of-use thermal prices can be formulated, thereby forming a time-of-use electricity price and time-of-use thermal price demand response strategy to guide the load to change the energy consumption plan.

[0082] Step S32: Determine the participation degree of different types of user demand response according to the time-sharing energy price.

[0083] In this embodiment, the participation of different types of users in demand response can be determined based on the formulated time-of-use energy price, and the user demand response participation reflects the participation of users in demand. Among them, this embodiment takes into account the influence of users' energy consumption psychology, and the effect of demand response is uncertain. Therefore, the participation of each type of user in demand response can be determined based on the time-of-use energy price of each time period corresponding to each type of user, as well as the critical incentive energy price and saturation incentive point energy price of each type of user. Taking the electricity load and electricity price as an example, the user demand response participation can be determined by formula (6):

[0084]

[0085] Among them, in formula (6): δ N Demand response participation for the Nth category of users; is the time-of-use electricity price for the Nth category of users during period t; represents the critical incentive electricity price for the Nth category of users; Represents the saturation incentive electricity price for the Nth type of user.

[0086] It should be noted that the method for determining the user demand response participation degree based on heat load and heat price is the same as or similar to the method for determining the user demand response participation degree based on electricity load and electricity price.

[0087] Step S33: Determine the classified load demand response according to the different types of user demand response participation, price elasticity matrix and the corresponding time-of-use energy price.

[0088] In this embodiment, the classified load demand response can be determined according to the different types of user demand response participation, the price elasticity matrix and the time-of-use energy prices corresponding to the different types of power loads. The classified load demand response reflects the user's participation in the load demand response. For example, the classified load demand response can be determined by a demand response model that takes into account the user's demand response participation.

[0089] Among them, this embodiment can model the demand response model according to different types of user demand response participation, price elasticity matrix and corresponding time-of-use energy price. For example, the price elasticity matrix is ​​used to model the response of the user's electric load to the electricity price and the response of the user's thermal load to the heat price. The following takes the modeling of the user's electric load response to the electricity price as an example, and considers the load quantity of the user's demand response parameter to the price response model as shown in formula (7). It should be noted that the establishment method for the response modeling of the user's thermal load and heat price is the same or similar to the establishment method for the response modeling of the electric load and the electricity price.

[0090]

[0091] In formula (7): is the original load of the Nth type of user demand response in period t; The load change in period t obtained by adjusting the power consumption of controllable loads after the price demand response of the Nth type of user; is the change in electricity price in period t after the Nth type of user responds to demand; is the electricity price of the user in time period t before the response; E is the electricity quantity and price elasticity coefficient matrix; δ N The demand response participation rate of the Nth type of user.

[0092] In this way, according to the participation of different types of user demand response, price elasticity matrix and corresponding time-of-use energy price, the classified load demand response of different types of user loads can be determined, so that the user load demand can be determined according to the classified load demand response, providing a basis for the subsequent scheduling of the integrated energy system.

[0093] In this embodiment, by analyzing the load characteristics unique to each type of power user, clustering and differentiating the user loads, and responding to different types of time-of-use electricity price demands, the load demand response potential can be better stimulated. According to the load characteristics of different types of energy users, time-of-use energy price demand response is carried out. By adjusting the time-of-use energy price, users are accurately guided to use energy at off-peak times, which can more effectively reduce the difference between peak and low energy consumption periods, flatten the load curve, and help improve the economic efficiency of the park's integrated energy system. Figure 2 As shown, Figure 2 is a schematic diagram of electric load optimization after classified load response according to an embodiment of the present invention. Figure 2 In the figure, the load curve after response is flatter than the original load curve, which reflects the time-of-use electricity price demand response. By adjusting the time-of-use electricity price and accurately guiding users to use electricity at off-peak times, the gap in electricity consumption between peak and off-peak periods can be effectively reduced, which helps to improve the economic efficiency of the park's comprehensive energy system operation.

[0094] In combination with the above embodiments, in one implementation, the present invention further provides a method for dispatching an integrated energy system including electricity-to-gas taking into account classified load demand response. In the method, the above step S11 may specifically include steps S41 to S44:

[0095] Step S41: Generate an original scenario set of wind turbine generating power that obeys probability distribution by using Latin hypercube sampling method.

[0096] In this embodiment, a large number of original scene sets of wind turbine power generation (output) that obey probability distribution can be generated by Latin hypercube sampling (LHS). Specifically, the expected short-term predicted power value μ of the wind turbine can be obtained based on historical data. w Assuming that the wind turbine power error satisfies the normal distribution of N(0, σ2), a large number of original scene sets of wind turbine output that obey the probability distribution can be generated by the Latin hypercube sampling method. Among them, one scene of this embodiment is the output (power generation) of the wind turbine in the integrated energy system 24 hours a day.

[0097] Step S42: The original scene set is reduced by a synchronous back-substitution reduction method based on probability distance to obtain a simplified scene set.

[0098] In this embodiment, after obtaining the original scene set, the original scene set can be reduced by a synchronous back-generation reduction method based on probability distance to obtain a simplified scene set to reduce the complexity of calculation. Specifically, the first number of scenes arranged from high to low probability of scene occurrence can be determined from the original scene set by the synchronous back-generation reduction method as simplified scenes, a simplified scene set is generated, and multiple simplified scenes with corresponding probabilities are derived. Among them, the first number can be arbitrarily set according to experience and needs, for example, the first number can be 10, and this embodiment does not impose any specific restrictions on this.

[0099] Step S43: performing multiplication operation on the probability of occurrence of each scenario in the simplified scenario set and the corresponding wind turbine generator set power generation to obtain a product result.

[0100] In this embodiment, after the simplified scene set is determined, the probability of occurrence of each scene in the simplified scene set can be determined (for example, 10 simplified scene sets and their corresponding probabilities P of wind turbine output are obtained). n , where n = 1, 2, ..., 10) and the wind turbine generation power corresponding to each scenario, so as to multiply and sum the probability of each scenario in the simplified scenario set and the wind turbine generation power corresponding to each scenario to obtain the product result. For example, the product result can be: F = P1 x p1 + P2 x p2 ... + P 10 xp 10 , where P n is the wind turbine power generation corresponding to each simplified scenario, P n is the probability of occurrence of each simplified scenario, n=1,2,…,10.

[0101] Among them, in view of the uncertainty of wind turbine output, the wind turbine power generation in each time period under each scenario may be different, so the wind turbine power generation corresponding to each simplified scenario can be a 1x24 matrix, and the 24 values ​​in the matrix correspond to the wind turbine power generation in each time period within 24 hours under the simplified scenario, and these 24 values ​​can be partially the same or different.

[0102] Step S44: determining the multiplication result as the wind power curve taking into account the uncertainty of wind power output.

[0103] In this embodiment, after obtaining the product result, the product result can be determined as a wind power curve that takes into account the uncertainty of wind power output. Figure 3 As shown, Figure 3 The wind power curve diagram taking into account the uncertainty of wind turbine output is shown in one embodiment of the present invention.

[0104] In combination with the above embodiments, in an optional embodiment, the present invention further proposes a method for dispatching an integrated energy system including electricity-to-gas taking into account classified load demand response. In the method, the user load characteristics are obtained by the following steps:

[0105] Step S51: Obtain typical daily user load data;

[0106] Step S52: using an LSTM autoencoder to extract features from the typical day user load data to obtain the user load features.

[0107] In this embodiment, the user load data of a typical day can be obtained, and the typical day user load data of this embodiment is historical load data. Then, a pre-trained LSTM autoencoder is used to extract features of the typical day user load data, thereby obtaining the user load features output by the LSTM autoencoder.

[0108] Among them, the training process of the LSTM autoencoder can be as follows: use the user load data as the model input, and use the mean square error loss function as the loss function to train the LSTM autoencoder until the mean square error loss function is small enough to end the training of the LSTM autoencoder; then input the test set sequence into the trained LSTM autoencoder, test the model, and use the output of the model as the extracted load feature.

[0109] For example, 10,000 daily load data of commercial and industrial users in a certain area are used to train the LSTMAE network. Based on the trained LSTMAE network model, feature extraction and subsequent clustering analysis are performed on the load data set of 1,200 commercial and industrial users in the area. Then, after using LSTMAE to extract features from the load curve and obtain a reduced-dimensionality feature sequence, the improved K-means clustering method is used to cluster the reduced-dimensionality feature sequence corresponding to the daily user load sequence to obtain the classification results of different types of user loads.

[0110] In this embodiment, a deep learning load classification method is used to extract the features of typical daily loads, and cluster them into several types of loads with similar load characteristics, and formulate corresponding time-sharing demand response strategies based on the energy consumption characteristics of each type of load. Among them, the LSTM autoencoder can effectively extract the deep features of the load curve and realize the dimensionality reduction of the high-dimensional load curve, with good dimensionality reduction accuracy and robustness.

[0111] In combination with the above embodiments, in one implementation, the present invention further provides a method for dispatching an integrated energy system including electricity-to-gas taking into account classified load demand response. In the method, the above step S14 may specifically include steps S61 to S63:

[0112] Step S61: constructing an energy supplier optimization model with the goal of maximizing the profit of the energy supplier.

[0113] In this embodiment, in order to construct the comprehensive energy system optimization model, an energy supplier optimization model with the goal of maximizing the profit of energy suppliers can be constructed first. The purpose of the energy supplier optimization model is to maximize the profit of energy suppliers as much as possible.

[0114] Step S62: constructing an internal optimization model of the integrated energy system including the power-to-gas equipment with the goal of minimizing the cost of the integrated energy system and taking into account the classified load demand response.

[0115] In this embodiment, in addition to constructing an energy supplier optimization model, it is also necessary to construct an internal optimization model of the integrated energy system containing power-to-gas equipment that takes into account the classified load demand response and aims to minimize the cost of the integrated energy system. The purpose of the internal optimization model of the integrated energy system is to minimize the cost of the integrated energy system.

[0116] Step S63: constructing the integrated energy system optimization model according to the energy supplier optimization model and the internal optimization model of the integrated energy system.

[0117] In this embodiment, after constructing the energy supplier optimization model and the internal optimization model of the integrated energy system, an integrated energy system optimization model can be constructed based on the energy supplier optimization model and the internal optimization model of the integrated energy system. The purpose of the integrated energy system optimization model is to maximize the benefits of energy suppliers as much as possible on the basis of minimizing the costs of the integrated energy system.

[0118] In this embodiment, an external bargaining model between the park integrated energy system and the energy supplier can be constructed based on the constructed energy supplier optimization model and the internal optimization model of the integrated energy system, that is, the integrated energy system optimization model, so that the park integrated energy system can optimize the operation plan according to the electricity and heat price signals, formulate energy purchase plans and submit them to energy suppliers, so as to maximize the benefits of energy suppliers as much as possible on the basis of minimizing the cost of the integrated energy system.

[0119] In combination with the above embodiments, in one implementation, the present invention further provides a method for dispatching an integrated energy system including electricity-to-gas taking into account classified load demand response. In the method, the above step S61 may specifically include step S71:

[0120] Step S71: Taking the maximum energy sales revenue of the energy supplier as the first objective function, and taking energy sales constraints and energy interaction constraints as first constraints, constructing an optimization model for the energy supplier.

[0121] In this embodiment, when constructing the energy supplier optimization model, the energy supplier's energy sales revenue can be maximized as the first objective function, and the energy sales constraint and energy interaction constraint can be used as the first constraint conditions to construct the energy supplier optimization model. Specifically, the expression of the first objective function can be as follows:

[0122]

[0123]

[0124] In formulas (8) and (9): F ESN is the comprehensive income of energy providers; T is the dispatch period; R SEN (t) revenue for energy suppliers; is the electricity price, heat price and gas price; C SEN (t) is the cost of energy supplier; ρ caol is the price of coal; Coal consumption by energy suppliers; SEN , b SEN , c SEN is the production cost coefficient of the energy supplier; g SEN (t) is the total amount of energy sold; They are the amount of electricity, heat and gas purchased for the park's integrated energy system.

[0125] In combination with the above embodiments, in one implementation, the present invention further provides a method for dispatching an integrated energy system including electricity-to-gas taking into account classified load demand response. In the method, the above step S62 may specifically include step S81:

[0126] Step S81: Taking the minimization of the operation cost of the integrated energy system as the second objective function, and taking the energy balance constraint and the unit constraint as the second constraint conditions, an internal optimization model of the integrated energy system is constructed.

[0127] In this embodiment, when constructing the internal optimization model of the integrated energy system, the minimum operating cost of the integrated energy system can be used as the second objective function, and corresponding time-of-use electricity price demand response strategies can be formulated for different types of loads to achieve peak load shaving and valley filling of the load curve, optimize the energy purchase plan, and use energy balance constraints and unit constraints as the second constraint conditions to construct the internal optimization model of the integrated energy system. Specifically, the expression of the second objective function can be shown in the following formula (10):

[0128]

[0129] In formula (10): C PIES is the comprehensive energy system cost of the park; T is the scheduling period.

[0130]

[0131] In formula (11), Ω e C is a collection of devices; OM,i (t) is the operation and maintenance cost of equipment i in period t; P i (t) is the operating power of device i during period t; c i,om is the operating cost coefficient of equipment i; WPP is wind turbine; GT is gas turbine; CH is heat exchanger; EHP is electric heat pump; P2G is power to gas device; HSD is heat storage tank; ESD is electric energy storage device; GSD is gas storage tank.

[0132] In combination with the above embodiments, in one implementation, the present invention further provides a method for dispatching an integrated energy system including electricity-to-gas taking into account classified load demand response. The method may further include steps S91 to S94:

[0133] Step S91: Determine the user's energy load demand according to the classified load demand response.

[0134] After determining the classified load demand response, this embodiment can determine the energy load demand of each type of user, such as the user's electrical load demand or the user's thermal load demand, according to the classified load response demand.

[0135] Step S92: Determine the energy balance constraint according to the user's energy load demand and equipment parameters.

[0136] In this embodiment, the energy balance constraint of the internal optimization model of the integrated energy system can be determined based on the determined user energy load demand and the equipment parameters in the park. Specifically, the energy balance constraint of the park can be shown in the following formula (12):

[0137]

[0138] In formula (12): Q E (t) is the user's electricity load demand; Q WWP (t) is the output of wind turbine; Q GT (t) is the power generated by the gas turbine; Q ESD (t) is the power provided by the electric energy storage device; Q H (t) is the user's heat load demand; Q EHP (t) is the operating power of the electric heat pump; Q CH (t) is the heat exchanger power; Q HSD (t) is the power of the heat storage tank; Q G (t) is the natural gas consumption of the park system; Q G,P2G(t) Natural gas flow provided to the P2G device (power-to-gas device); Q E,P2G (t) is the operating power of the P2G device; Q GSD (t) The natural gas flow rate provided to the gas storage tank unit.

[0139] Step S93: determining the maximum power generation of the wind turbine according to the wind power curve taking into account the uncertainty of wind power output.

[0140] In this embodiment, the maximum power generation power of the wind turbine generator set may be determined based on the acquired wind power curve taking into account the uncertainty of the wind turbine generator set output.

[0141] Step S94: determining the wind turbine constraints according to the maximum power generation of the wind turbine and the equipment parameters.

[0142] In this embodiment, the unit constraints of the internal optimization model of the integrated energy system can be determined based on the determined maximum power generation of the wind turbine, i.e., the maximum output of the wind turbine, and the equipment parameters. The unit constraints can be shown in the following formulas (13)-(15):

[0143]

[0144] In formula (13), is the maximum output of the wind turbine; is the maximum power generation capacity of the gas turbine; is the maximum operating power of the heat exchanger; It is the maximum operating power of the electric heat pump.

[0145]

[0146] In formula (14), is the conversion efficiency of electricity to gas; Hg is the calorific value of natural gas; is the maximum operating power of the P2G device; is the minimum operating power of the P2G device.

[0147]

[0148] In formula (15), E x (t) is the energy storage state of the energy storage device x; E x,min is the minimum energy storage state of the energy storage device x; E x,max is the maximum energy storage state of the energy storage device x; x,d (t), λ x,c (t) are the charging and discharging states of the energy storage device x, respectively; Q x,d,max , Q x,c,maxare the maximum energy storage and release powers of energy storage device x; ESD, HSD, and GSD are electric energy storage, heat storage tank, and gas storage tank respectively; E x (0) is the energy storage state of the energy storage device at time x 0; E x (24) is the energy storage state of the energy storage device at time x 24.

[0149] It can be understood that the unit constraints may include at least: wind turbine constraints, gas turbine constraints, heat exchanger constraints, electric heat pump constraints, power-to-gas constraints and energy storage device constraints; wherein the above formula (13) respectively represents the wind turbine constraints, gas turbine constraints, heat exchanger constraints and electric heat pump constraints; the above formula (14) represents the power-to-gas constraints; and the above formula (15) represents the energy storage device constraints.

[0150] In combination with the above embodiments, in one implementation, the present invention further provides a method for dispatching an integrated energy system including electricity-to-gas taking into account classified load demand response. In the method, the step S15 of "solving the integrated energy system optimization model to obtain an optimized operation plan of the integrated energy system" may specifically include step S101:

[0151] Step S101: Integrate the multiverse optimization algorithm and the mixed integer linear programming method to solve the integrated energy system optimization model and obtain the optimized operation plan.

[0152] In this embodiment, the comprehensive energy system optimization model can be solved based on the wind power curve that takes into account the uncertainty of wind power output, the load characteristics of different types of users, and the energy quotations provided by energy suppliers, and the multiverse optimization algorithm and mixed integer linear programming method are integrated to obtain the optimal operation plan of the park comprehensive energy system. Among them, the optimal operation plan output by the comprehensive energy system optimization model can at least include: the energy supplier's energy purchase plan, and the time-of-use pricing plan for different types of loads. And, in one embodiment, the input of the comprehensive energy system optimization model can also include: various parameters of various equipment in the park.

[0153] In this embodiment, a multiverse optimization algorithm is used in combination with a mixed integer linear programming method to solve the optimization model of the park's integrated energy system. The law of objects with high expansion rates moving in space along wormholes during the random creation of the universe is utilized, and the optimal search is achieved by randomly transmitting objects between white holes and black holes. The algorithm has the characteristics of fast computing speed, strong convergence, and suitability for high-dimensional computing.

[0154] In an alternative embodiment, please refer to Figure 4 , Figure 4is a flow chart showing a solution of an integrated energy system optimization model including power-to-gas equipment considering classified load demand response according to an embodiment of the present invention. Figure 4 It can also be understood as a flow chart of a method for dispatching an integrated energy system including electricity-to-gas conversion that takes into account classified load demand response.

[0155] like Figure 4 As shown, after the process starts, data is initialized first, and the first iteration is performed, i=1; then the initial population of energy supplier quotation strategies is generated; then the energy supplier sends quotation strategy information to the park integrated energy system; then the park integrated energy system is optimized and operated, that is, solved through the integrated energy system optimization model, and the energy supplier receives the operation result: the lower-level energy purchase plan. Among them, in the lower-level framework, it can be that the integrated energy system (integrated energy system optimization model) receives the quotation strategy information, calls the solver to solve the optimization result of the integrated energy system (integrated energy system optimization model), and feeds back the obtained energy purchase plan to the energy supplier.

[0156] After the energy supplier receives the lower-level energy purchase plan, the energy supplier's profit can be calculated based on the energy purchase plan, and then it is determined whether the current number of iterations reaches the maximum number of iterations or whether the minimum error accuracy is met; if not, the roulette selection mechanism is executed, WEP and TDR are updated, and a new quotation strategy population is obtained, so as to perform the next iteration, so that i=i+1, and the step of "sending quotation strategy information to the park integrated energy system" is repeated to continue the execution process; if yes, the integrated energy system directly outputs the strategy information and operation results, and the process ends.

[0157] In an optional implementation, in order to verify the impact of power-to-gas and park user demand response (classified load demand response) on the optimized operation of the park's integrated energy system, four scenarios are set for analysis: Scenario 1: Power-to-gas and demand response are not considered; Scenario 2: Power-to-gas is considered, but demand response is not considered; Scenario 3: Demand response is considered, but power-to-gas is not considered; Scenario 4: Demand response and power-to-gas are considered.

[0158] The utility functions of the integrated energy system of the park under the above four scenarios are calculated respectively. The calculation results are shown in Table 1.

[0159] Table 1

[0160]

[0161] From the analysis of Table 1 above, it can be seen that by optimizing the energy utilization rate of equipment through classified load demand response and guiding users to optimize energy consumption plans, peak shaving and valley filling can be achieved; and by adding power-to-gas, the consumption of renewable energy is improved, the phenomenon of wind abandonment is reduced, and the economic efficiency of the park system operation can be better improved. When the classified load demand response and power-to-gas are considered during the operation of the park's comprehensive energy system, the linkage effect within the park is maximized, and the total operating cost is 30,202.7 yuan. Compared with scenario 1, the total operating cost of the park system under scenarios 2, 3 and 4 is reduced by 924.3 yuan, 1,267.1 yuan and 2,039.93 yuan respectively. By analyzing the unique load characteristics of each type of power user, the user load is clustered and differentiated, and the time-of-use electricity price demand response is performed for different categories, as shown in Tables 2 and 3, which better stimulates the potential of load demand response. The power-to-gas device further deepens the coupling between electricity and improves the operational flexibility of the system.

[0162] Table 2 Optimization of time-of-use electricity price for classified loads

[0163]

[0164] Table 3 Optimization of time-of-use heat price for classified loads

[0165]

[0166] In addition, it should be noted that Table 2 and Table 3 can also be examples of time-of-use pricing schemes for different types of loads output by the integrated energy system optimization model.

[0167] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0168] It should be noted that, for the sake of simplicity, the method embodiments are described as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0169] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0170] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0171] The above is a detailed introduction to the scheduling method of an integrated energy system including electricity-to-gas conversion taking into account the classified load demand response provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for general technical personnel in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for dispatching an integrated energy system including electricity-to-gas taking into account classified load demand response, characterized in that: The method comprises: Obtain wind power curves that take into account wind power output uncertainty; The user load characteristics are clustered by the improved K-means clustering method to obtain different types of user load characteristics; Determining classified load demand responses according to the load characteristics of the different types of users; Constructing a comprehensive energy system optimization model including power-to-gas equipment considering the classified load demand response; Solving the comprehensive energy system optimization model according to the wind power curve taking into account the uncertainty of wind power output, the load characteristics of different types of users, and the energy quotations provided by energy suppliers to obtain an optimized operation plan for the comprehensive energy system; The construction of a comprehensive energy system optimization model including power-to-gas equipment taking into account the classified load demand response comprises: Constructing an energy supplier optimization model with the goal of maximizing the revenue of the energy supplier; Constructing an internal optimization model of the integrated energy system including power-to-gas equipment with the goal of minimizing the cost of the integrated energy system and considering the classified load demand response; Constructing the integrated energy system optimization model according to the energy supplier optimization model and the integrated energy system internal optimization model; The step of solving the integrated energy system optimization model to obtain an optimized operation plan for the integrated energy system includes: The comprehensive energy system optimization model is solved by integrating the multiverse optimization algorithm and the mixed integer linear programming method to obtain the optimized operation plan, which at least includes: a plan for purchasing energy from the energy supplier and a time-of-use pricing plan for different types of loads; The step of obtaining a wind power curve taking into account the uncertainty of wind power output includes: The original scenario set of wind turbine power generation that obeys the probability distribution is generated by the Latin hypercube sampling method; The original scene set is reduced by a synchronous back-substitution reduction method based on probability distance to obtain a streamlined scene set; Performing a multiplication operation on the probability of occurrence of each scenario in the simplified scenario set and the corresponding wind turbine generator set power generation to obtain a product result; The product result is determined as the wind power curve taking into account the uncertainty of wind power output.

2. The method for dispatching an integrated energy system including power-to-gas taking into account classified load demand response according to claim 1, characterized in that: The improved K-means clustering method is used to cluster the user load characteristics to obtain different types of user load characteristics, including: The DTW similarity distance index is used to replace the distance index in the original K-means clustering method to obtain the improved K-means clustering method; The user load characteristics are processed by the improved K-means clustering method to obtain the similarity between the user load characteristics; According to the similarity between the user load characteristics, the user load characteristics are classified to obtain different types of user loads with similar energy consumption characteristics, and the user load characteristics of the different types are determined.

3. The method for dispatching an integrated energy system including power-to-gas taking into account classified load demand response according to claim 1, characterized in that: The step of determining the classified load demand response according to the load characteristics of different types of users includes: Determining corresponding time-of-use energy prices according to the load characteristics of different types of users; Determining the participation of different types of users in demand response according to the time-of-use energy price; The classified load demand response is determined according to the different types of user demand response participation, the price elasticity matrix and the corresponding time-of-use energy price.

4. The method for dispatching an integrated energy system including power-to-gas taking into account classified load demand response according to claim 1, characterized in that: The user load characteristics are obtained by the following steps: Obtain typical daily user load data; The LSTM autoencoder is used to extract features of the typical day user load data to obtain the user load features.

5. The method for dispatching an integrated energy system including power-to-gas taking into account classified load demand response according to claim 1, characterized in that: The energy supplier optimization model is constructed with the goal of maximizing the benefits of the energy supplier, including: The energy supplier optimization model is constructed by taking the maximum energy sales revenue of the energy supplier as the first objective function and taking energy sales constraints and energy interaction constraints as the first constraint conditions.

6. The method for dispatching an integrated energy system including power-to-gas taking into account classified load demand response according to claim 1, characterized in that: The construction of an internal optimization model of the integrated energy system including the power-to-gas equipment with the goal of minimizing the cost of the integrated energy system and considering the classified load demand response comprises: The internal optimization model of the integrated energy system is constructed by taking the minimization of the operating cost of the integrated energy system as the second objective function and taking the energy balance constraint and the unit constraint as the second constraint conditions.

7. The method for dispatching an integrated energy system including power-to-gas taking into account classified load demand response according to claim 6, characterized in that: The method further comprises: Determining the user's energy load demand according to the classified load demand response; Determining the energy balance constraint according to the user energy load demand and equipment parameters; Determining the maximum power generation of the wind turbine according to the wind power curve taking into account the uncertainty of wind power output; The wind turbine group constraints are determined according to the maximum power generation power of the wind turbine group and the equipment parameters.

Images