Scheduling operation method of new energy distribution and storage combined participation frequency modulation auxiliary service market

Abstract

The invention relates to the field of power systems, and particularly discloses a scheduling operation method for a new energy distribution and storage combined participation frequency modulation auxiliary service market, which comprises the following steps of: inputting output values of to-be-aggregated new energy power stations equipped with energy storage into a trained day-ahead output prediction model, generating a predicted output value of each day of each new energy power station, reducing a day scene by using a probability reduction method, and obtaining occurrence probabilities of a plurality of typical days; and based on the occurrence probability of the plurality of typical days and the predicted output value of the plurality of typical days of each new energy power station, optimizing the initialized plurality of alliance structures by using a particle swarm algorithm and taking the scheduling operation income of the alliance structure as a fitness value to obtain a target alliance structure, and outputting the capacity allocated to the frequency modulation auxiliary service market and the electric energy market by each new energy allocation storage in each alliance in the target alliance structure on each typical day. The energy storage utilization rate of new energy can be improved, and the power supply quality of the power grid can be improved.

Description

Technical Field

[0001] The present invention relates to the field of power systems, and in particular to a dispatching and operating method for a new energy distribution and storage system to jointly participate in a frequency regulation auxiliary service market. Background Art

[0002] With the development of new energy technologies, the installed capacity of new energy power plants has grown rapidly. However, the power generation output of new energy (such as wind power and photovoltaic power) is greatly affected by weather conditions and has the characteristics of intermittent and fluctuating, which brings huge challenges to the safe and stable operation of the power system. To this end, my country proposes to build a system-friendly new energy power station equipped with new energy storage, that is, new energy power plants are equipped with energy storage devices, so as to effectively smooth out the fluctuations of new energy power generation, improve the flexibility and regulation ability of the power system, and thus enhance the absorption capacity of new energy.

[0003] However, currently most energy storage services serve a single entity, and the equivalent utilization coefficient of new energy storage devices, the frequency of calling up energy storage devices, and the utilization rate are low. Summary of the invention

[0004] To this end, the present invention provides a dispatching and operating method for new energy distribution and storage to jointly participate in the frequency regulation auxiliary service market, in an effort to solve or at least alleviate the above problems.

[0005] According to one aspect of the present invention, there is provided a method for dispatching and operating a new energy distribution and storage system jointly participating in a frequency regulation auxiliary service market, comprising: inputting the output values ​​of a predetermined number of days before each day of each new energy power station to be aggregated into a trained day-ahead output prediction model to generate a predicted output value of each new energy power station on each day, wherein each new energy power station is equipped with energy storage; based on the predicted output values ​​of each new energy power station on each day, using a probability reduction method to reduce daily scenarios, and obtaining the probability of occurrence of multiple typical days and the predicted output values ​​of multiple typical days of each new energy power station; based on the probability of occurrence of multiple typical days and the predicted output values ​​of multiple typical days of each new energy power station, using a particle swarm algorithm, taking the dispatching and operating benefits of the alliance structure as the fitness value, optimizing the initialized multiple alliance structures, obtaining a target alliance structure, and outputting the dispatching and operating benefits of each new energy distribution and storage system in each alliance in the target alliance structure on each typical day. The dispatching and operation plan includes the capacity of each renewable energy distribution storage allocated to the frequency regulation auxiliary service market and the electric energy market on each typical day, and any alliance includes at least one renewable energy power station equipped with energy storage; wherein the dispatching and operation benefits of any alliance structure are obtained in the following manner: considering the penalty for renewable energy output deviation, taking the maximum dispatching and operation benefits of each alliance in the alliance structure as the goal, constructing a dispatching and operation model for each alliance to participate in the frequency regulation auxiliary service market; based on the occurrence probability of multiple typical days and the predicted output values ​​of each renewable energy power station on multiple typical days, solving the dispatching and operation model for each alliance to participate in the frequency regulation auxiliary service market, obtaining the dispatching and operation plan for each renewable energy distribution storage in each alliance on each typical day and the dispatching and operation benefits of each alliance under the corresponding dispatching and operation plan; based on the dispatching and operation benefits obtained for each alliance, obtaining the dispatching and operation benefits of the alliance structure.

[0006] Optionally, in the dispatching and operating method for the new energy distribution and storage to jointly participate in the frequency regulation auxiliary service market according to the present invention, for each alliance, its dispatching and operating model includes an objective function, and the objective function includes:

[0007]

[0008] Among them, R year represents the alliance's scheduling and operation benefits, ρ k represents the probability of occurrence of a typical day k, represents the secondary frequency modulation income of the alliance in period t on a typical day k, represents the alliance’s on-grid electricity sales revenue during period t on a typical day k, represents the output deviation penalty of the alliance in period t on a typical day k, represents the energy storage operation cost of the alliance in time period t on typical day k, K represents the total number of typical days, and T represents the total number of time periods.

[0009] Optionally, in the dispatching and operation method of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market according to the present invention, a particle swarm algorithm is used, and the dispatching and operation benefits of the alliance structure are used as the fitness value to optimize the initialized multiple alliance structures to obtain the target alliance structure, including: the number of initialized alliance structures, the speed and position of each alliance structure, and the number of iterations; based on the current position of each alliance structure, the fitness value of each alliance structure is obtained; according to the obtained fitness value of each alliance structure, the optimal individual position of each alliance structure and the optimal group position of the alliance structure group are determined, and it is detected whether the initialized number of iterations is reached or whether all new energy power stations become an alliance; if the iteration number is not reached and all new energy power stations do not become an alliance, the current number of iterations is increased by one, and the speed and position of each alliance structure are updated, and the above fitness value acquisition, optimal individual position and optimal group position determination, and current number of iterations and alliance Whether the alliance meets the requirements are repeatedly executed until the initialized number of iterations is reached or all new energy power stations become an alliance, and the current optimal group position is used as the final position of the alliance structure to obtain the target alliance structure.

[0010] Optionally, in the dispatching and operation method of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market according to the present invention, the position of each alliance structure is composed of the number of alliances under the alliance structure and the number of the alliance in which each new energy power station participates.

[0011] Optionally, in the dispatching and operation method for the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market according to the present invention, the day-ahead output prediction model adopts an Elman neural network model.

[0012] Optionally, in the dispatching and operation method of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market according to the present invention, the day-ahead output prediction model is trained based on the following method: the real output values ​​of a predetermined number of days before the prediction date in the training set are input into the pre-trained day-ahead output prediction model to obtain the predicted output value of the prediction day, where the prediction day is any day in the training set; based on the loss value between the real output value and the predicted output value of the prediction day, the parameters of the day-ahead output prediction model are updated until the loss value meets the predetermined conditions, the training is completed, and the trained day-ahead output prediction model is obtained.

[0013] Optionally, in the dispatching and operation method for the joint participation of new energy distribution and storage in the frequency regulation auxiliary service market according to the present invention, for each alliance, its dispatching and operation model also includes constraints, and the constraints include constraints on the declared power and discharge capacity of each new energy power station in the alliance, constraints on the charging and discharging capacity of each energy storage in the alliance, constraints on the charge state of each energy storage in the alliance, and capacity constraints on the participation of each new energy power station in the alliance in the electricity market.

[0014] Optionally, in the dispatching and operating method for the new energy distribution and storage to jointly participate in the frequency regulation auxiliary service market according to the present invention, the output deviation penalty of the alliance in the t period of the typical day k is obtained by the following formula:

[0015]

[0016] in, represents the deviation penalty ratio for participating in the electric energy market, represents the unit grid-connected electricity price in period t of a typical day k, It represents the electric energy deviation of energy storage i in period t of typical day k, or the electric energy deviation of the main body formed by energy storage i and its corresponding new energy power station in period t of typical day k, and I represents the total amount of energy storage.

[0017] According to another aspect of the present invention, there is provided a computing device, comprising: at least one processor; and a memory storing program instructions, wherein the program instructions are configured to be suitable for execution by at least one processor, and the program instructions include instructions for executing the dispatching and operation method of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market according to the present invention.

[0018] According to another aspect of the present invention, a readable storage medium storing program instructions is provided. When the program instructions are read and executed by a computing device, the computing device executes the scheduling and operation method of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market according to the present invention.

[0019] According to the dispatching and operating method of the new energy storage and joint participation in the frequency regulation auxiliary service market of the present invention, the energy storage configured by the new energy is aggregated and simultaneously participates in the electric energy market and the frequency regulation auxiliary service market, and the new energy deviation penalty cost is considered when constructing the dispatching and operating model. In addition, by constructing an evolutionary cooperative game solving algorithm based on a particle swarm algorithm, a theory of alliance structure dominance is proposed to obtain a stable game structure of the cooperative game. Therefore, the present invention can not only improve the energy storage utilization rate of new energy, but also effectively participate in frequency regulation assistance, thereby improving the power supply quality of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] To achieve the above and related purposes, certain illustrative aspects are described herein in conjunction with the following description and accompanying drawings, which indicate various ways in which the principles disclosed herein can be practiced, and all aspects and their equivalents are intended to fall within the scope of the claimed subject matter. The above and other purposes, features and advantages of the present disclosure will become more apparent by reading the following detailed description in conjunction with the accompanying drawings. Throughout the present disclosure, the same reference numerals generally refer to the same parts or elements.

[0021] Figure 1A structural block diagram of a computing device 100 according to an embodiment of the present invention is shown;

[0022] Figure 2 A flowchart of a dispatching operation method 200 for new energy distribution and storage to jointly participate in a frequency regulation auxiliary service market according to an embodiment of the present invention is shown;

[0023] Figure 3 A schematic diagram showing a dispatching operation framework of a new energy distribution and storage jointly participating in a frequency regulation auxiliary service market according to an embodiment of the present invention;

[0024] Figure 4 A schematic diagram of a day-ahead output prediction model based on an Elman neural network model according to an embodiment of the present invention is shown;

[0025] Figure 5 A schematic diagram showing a flow chart of a method for dispatching and operating a new energy distribution and storage system jointly participating in a frequency regulation auxiliary service market according to yet another embodiment of the present invention;

[0026] Figure 6 A schematic diagram showing the market price of electric energy, the price of frequency regulation capacity and the price of frequency regulation mileage according to an embodiment of the present invention is shown;

[0027] Figure 7 A schematic diagram showing a frequency modulation call ratio every five minutes according to an embodiment of the present invention is shown;

[0028] Figure 8 A schematic diagram showing a new energy prediction error scenario according to an embodiment of the present invention is shown;

[0029] Fig. 9 A schematic diagram showing the revenue of different renewable energy power plants and renewable energy consumption rates under different schemes according to an embodiment of the present invention is shown;

[0030] Fig.10 A schematic diagram showing an alliance structure iteration result according to an embodiment of the present invention;

[0031] Fig.11 A schematic diagram showing relevant results obtained after allocating cooperative games according to Shapley values ​​according to an embodiment of the present invention;

[0032] Fig.12 A schematic diagram showing the capacity allocation results of energy storage on different typical days according to an embodiment of the present invention;

[0033] Fig.13 A schematic diagram showing comparison of results under different electric energy deviation penalty rates according to an embodiment of the present invention is shown;

[0034] Fig.14A schematic diagram showing alliance revenues under different allocation and storage ratios according to an embodiment of the present invention is shown;

[0035] Fig.15 A schematic diagram showing the prediction error absorption rate under different allocation and storage ratios according to an embodiment of the present invention is shown;

[0036] Fig.16 A schematic diagram showing comparison of results under different new energy change rate limits according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0037] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0038] The construction of system-friendly new energy power stations equipped with new energy storage is an important guarantee for improving the green electricity access system capabilities and building a new power system. However, a large number of friendly new energy power stations generally have prominent problems such as low utilization of storage (i.e. energy storage devices configured in new energy power stations / plants) and ineffective flexible adjustment capabilities.

[0039] In view of this, the present invention proposes that the utilization rate of energy storage configured in new energy can be improved by using new energy distribution storage to participate in frequency regulation, which can also improve the stability of the power system. In other words, new energy distribution storage participates in the electric energy market and the frequency regulation auxiliary service market at the same time. On this basis, the present invention also proposes that multiple new energy distribution storage can be aggregated into a new energy distribution storage cluster to participate in the electric energy market and the frequency regulation auxiliary service market (that is, multiple new energy distribution storage jointly participate in the electric energy market and the frequency regulation auxiliary service market), which can further improve the energy storage utilization rate of new energy and the stability of the power system.

[0040] Based on this, how to aggregate new energy distribution and storage and how to balance the capacity of new energy distribution and storage allocated to the electric energy market and the frequency regulation auxiliary service market has become a major problem. To this end, the present invention provides a scheduling operation method for new energy distribution and storage to jointly participate in the frequency regulation auxiliary service market.

[0041] The dispatching and operating method of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market of the present invention can be executed in a computing device. Figure 1A block diagram of the physical components (i.e., hardware) of a computing device 100 is shown. In a basic configuration, the computing device 100 includes at least one processing unit 102 and a system memory 104. According to one aspect, depending on the configuration and type of the computing device, the processing unit 102 can be implemented as a processor. The system memory 104 includes, but is not limited to, volatile storage (e.g., random access memory), non-volatile storage (e.g., read-only memory), flash memory, or any combination of such memories. According to one aspect, the system memory 104 includes an operating system 105 and a program module 106, and the program module 106 includes a joint operation module 120, and the joint operation module 120 is configured to execute the dispatching operation method 200 of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market of the present invention.

[0042] According to one aspect, operating system 105 is suitable for controlling the operation of computing device 100, for example. Furthermore, examples are practiced in conjunction with graphics libraries, other operating systems, or any other application programs, and are not limited to any particular application or system. Figure 1 This basic configuration is illustrated in FIG. 1 by those components within dashed line 108. According to one aspect, computing device 100 has additional features or functionality. For example, according to one aspect, computing device 100 includes additional data storage devices (removable and / or non-removable), such as magnetic disks, optical disks, or tapes. Such additional storage Figure 1 1 is illustrated by removable storage 109 and non-removable storage 110.

[0043] As stated above, according to one aspect, program modules are stored in the system memory 104. According to one aspect, the program modules may include one or more application programs, and the present invention does not limit the type of application programs, for example, the application programs may include: email and contact applications, word processing applications, spreadsheet applications, database applications, slide show applications, drawing or computer-aided applications, web browser applications, etc.

[0044] According to one aspect, the examples may be practiced on a circuit comprising discrete electronic components, a packaged or integrated electronic chip containing logic gates, a circuit utilizing a microprocessor, or a single chip containing electronic components or a microprocessor. Figure 1Each or many components shown in can be integrated into a system on a chip (SOC) on a single integrated circuit to practice examples. According to one aspect, such a SOC device may include one or more processing units, a graphics unit, a communication unit, a system virtualization unit, and various application functions, all of which are integrated (or "burned") into a chip substrate as a single integrated circuit. When operated via SOC, the functions described in this article can be operated via a dedicated logic integrated with other components of the computing device 100 on a single integrated circuit (chip). Embodiments of the present invention can also be practiced using other technologies capable of performing logical operations (such as AND, OR, and NOT), including but not limited to mechanical, optical, fluid, and quantum technologies. In addition, embodiments of the present invention can be practiced in a general-purpose computer or in any other circuit or system.

[0045] According to one aspect, the computing device 100 may also have one or more input devices 112, such as a keyboard, a mouse, a pen, a voice input device, a touch input device, etc. Output devices 114 may also be included, such as a display, a speaker, a printer, etc. The aforementioned devices are examples and other devices may also be used. The computing device 100 may include one or more communication connections 116 that allow communication with other computing devices 118. Examples of suitable communication connections 116 include, but are not limited to: RF transmitter, receiver and / or transceiver circuits; Universal Serial Bus (USB), parallel and / or serial ports.

[0046] The term computer-readable medium as used herein includes computer storage media. Computer storage media may include volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (e.g., computer-readable instructions, data structures, or program modules). System memory 104, removable storage 109, and non-removable storage 110 are all examples of computer storage media (i.e., memory storage). Computer storage media may include random access memory (RAM), read-only memory (ROM), electrically erasable read-only memory (EEPROM), flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical storage, cassettes, tapes, disk storage or other magnetic storage devices, or any other products that can be used to store information and can be accessed by computer device 100. According to one aspect, any such computer storage medium can be a part of computing device 100. Computer storage media do not include carrier waves or other propagated data signals.

[0047] According to one aspect, communication media is implemented by computer readable instructions, data structures, program modules, or other data in a modulated data signal (e.g., a carrier wave or other transport mechanism), and includes any information delivery media. According to one aspect, the term "modulated data signal" describes a signal that has one or more characteristics set or changed in a manner that encodes information in the signal. By way of example and not limitation, communication media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, radio frequency (RF), infrared, and other wireless media.

[0048] Figure 2 A flowchart of a method 200 for dispatching and operating a new energy distribution and storage system to jointly participate in a frequency regulation auxiliary service market according to an embodiment of the present invention is shown. The method 200 is suitable for computing a system (e.g. Figure 1 Executed in the computing device 100 shown.

[0049] Here, the operating framework of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market of the present invention (or the operating framework of the new energy distribution and storage cluster participating in the frequency regulation auxiliary service market) is first explained.

[0050] Wind and solar power output is characterized by volatility, intermittency, and uncertainty. Connecting them to the grid alone often poses certain risks to grid dispatching and operation. Therefore, power supply-side energy storage can be used to track the day-ahead output plan of new energy sources to increase the quality of new energy output. However, there is currently a problem of low resource utilization in power supply-side energy storage. To address this, frequency modulation resources can be provided to the grid by aggregating new energy output and sharing energy storage resources. This can not only bring additional benefits to the power supply side and shorten the investment payback period, but also improve the grid margin and increase the safety of grid dispatching and operation.

[0051] Therefore, the present invention proposes a power supply side energy storage sharing mode, wherein the power supply side energy storage includes at least new energy storage, specifically, the new energy storage may include photovoltaic power generation storage and wind power generation storage, and of course, in some embodiments, other storage may also be included, which is not limited by the present invention.

[0052] Next, combine Figure 3The operating framework of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market constructed by the present invention is explained. Photovoltaic power generation distribution and storage and wind power distribution and storage give priority to meeting the output of photovoltaic power plants and wind power plants-renewable energy power plants (REPP) when their output is less than the day-ahead plan. When the self-storage cannot meet the planned output of the new energy power station-energy storage, capacity can be purchased from other energy storage to avoid assessment deviations. In addition, new energy self-storage power stations can form alliances with independent energy storage, and the remaining capacity can jointly participate in frequency regulation auxiliary services to increase revenue. Among them, the alliance is explained here. In the present invention, in addition to new energy power stations equipped with energy storage, an alliance may also include independent energy storage. That is to say, an alliance includes at least one new energy power station equipped with energy storage, which is composed of a new energy power station and distribution storage.

[0053] The above is the operating framework of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market of the present invention. Next, the dispatching and operating method of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market of the present invention is described. Figure 2 As shown, the method 200 begins at 210 .

[0054] In 210, the output values ​​of the previous predetermined number of days of each of the new energy power stations to be aggregated are input into the trained day-ahead output prediction model to generate the predicted output values ​​of each of the new energy power stations to be aggregated on each day. Each of the new energy power stations to be aggregated is equipped with energy storage.

[0055] Among them, the prediction of new energy day-ahead output is divided into two methods: physical and statistical. The physical method predicts the weather by solving fluid mechanics and thermodynamics equations, combined with unit characteristic modeling, which is suitable for short-term to medium-term predictions, but requires complex data and high-precision weather forecasts. The statistical law is based on historical data and simplifies the processing. It is suitable for short-term or ultra-short-term predictions, but the long-term prediction error is large. Based on this, and taking into account the characteristics of the Elman neural network that memorizes historical data, the present invention proposes that the day-ahead output of new energy can be predicted based on the Elman neural network, which can improve the accuracy, efficiency and stability of the day-ahead output prediction of new energy. Therefore, according to one embodiment of the present invention, the day-ahead output prediction model adopts the Elman neural network model, and the structure diagram shows Figure 4 .

[0056] In addition, considering that the output of new energy power stations has a certain seasonality, the output data of new energy power stations (specifically, new energy power stations equipped with energy storage) for three months can be selected according to the seasonal principle as the training set (i.e., training samples) of each season to train the day-ahead output prediction model of each season (i.e., each season corresponds to a day-ahead output prediction model). Further, in some embodiments, the day-ahead output prediction model of any season can be trained in the following manner.

[0057] The actual output value of the predetermined number of days before the predicted day of any new energy power station in the training set of the season is input into the pre-trained day-ahead output prediction model to obtain the predicted output value of the predicted day of the new energy power station. Among them, the predicted day is any day in the training set, and further, there is any day of the predetermined number of days before in the training set. Then, based on the loss value between the actual output value and the predicted output value of the predicted day of the new energy power station, the parameters of the day-ahead output prediction model are updated (that is, the parameters of the day-ahead output prediction model are adjusted). Repeat the above steps until the loss value meets the predetermined conditions, the training is completed, and the trained day-ahead output prediction model of the season is obtained. In this way, for any new energy power station on any day subsequently, the predicted output value of the day can be obtained by inputting the output value of the predetermined number of days before the day into the trained day-ahead output prediction model of the season to which the day belongs.

[0058] In some embodiments, the value of the predetermined number may be three, which is of course only an example and is not limited to this in the present invention. Figure 4 The day-ahead output prediction model based on the Elman neural network model is shown to further illustrate the above learning process.

[0059] The mathematical expression between every two layers of the day-ahead output prediction model is as follows.

[0060] p d =g(ω 3 , {x,b2})

[0061] x=f(ω 1 x c +ω 2 {p d-1 ,p d-2 ,p d-3 ,b1})

[0062] x c (k) = x(k-1)

[0063] In the formula, p dis a T-dimensional output vector, representing the predicted value of renewable energy output on day d, T means that day d is divided into T time periods, x is an n-dimensional intermediate layer vector, x(k) and x(k-1) represent the kth component and (k-1)th component of x respectively, b2 is a constant term from the intermediate layer to the output layer, {p d-1 ,p d-2 ,p d-3 ,b1} is a (3*T+1)-dimensional input vector, p d-1 、p d-2 、p d-3 They represent the new energy output on (d-1), (d-2), and (d-3) days respectively, b1 is the constant term from the input layer to the middle layer, and x c is the m-dimensional receiving layer state vector, x c (k) represents x c The kth component of 3 is the connection weight from the hidden layer to the output layer, ω 2 is the connection weight from the input layer to the intermediate layer, ω 1 is the connection weight from the receiving layer to the middle layer, g(*) is the transfer function of the output neuron, which is a linear combination of the output of the middle layer, that is, g(ω 3 ,{x,b2})=ω 3 ·{x,b2},ω 3 is an n+1 dimensional vector, {x, b2} is an n+1 dimensional vector, f(*) is the transfer function of the middle layer neuron, and this embodiment adopts the Sigmoid function.

[0064] The learning index function may adopt an error square sum function, as shown in the following formula: In some embodiments, a gradient descent method may be used for solving.

[0065]

[0066] Among them, p d (t) They represent the predicted output value and the actual output value of the new energy at the tth period on day d, respectively. E(ω) is the error square sum function, which is used to evaluate the error between the predicted output value and the actual output value.

[0067] After the above training, a trained day-ahead output prediction model is obtained, which can then be used to generate the predicted output values ​​of each new energy power station to be aggregated on each day. Specifically, in some embodiments, the trained day-ahead output prediction model can be used to generate the predicted output values ​​of each new energy power station to be aggregated on each day of the year. The predicted output values ​​of each day can specifically include the predicted output values ​​of each time period of the day.

[0068] At this point, the predicted output values ​​of each new energy power station to be aggregated on each day of the year are obtained. Among them, considering that wind and solar power generation are greatly affected by environmental factors, its output has a certain uncertainty, but the number of daily output scenes in a whole year is too large. Therefore, in some embodiments, the probability rapid reduction method can be further used to reduce the scenes, as follows.

[0069] In 220, based on the predicted output values ​​of each new energy power station on each day, the daily scene is reduced using the probability reduction method to obtain the occurrence probabilities of multiple typical days and the predicted output values ​​of multiple typical days of each new energy power station. According to an embodiment of the present invention, reducing the daily scene using the probability reduction method may specifically include the following steps.

[0070] The first step is to calculate the Euclidean distance between every two scenarios, which can be calculated by the following formula: It should be noted that in this embodiment, the daily output is taken as a scenario, and it is assumed that the probability of each scenario occurring is the same.

[0071]

[0072] Where, d u,v represents the Euclidean distance between scene u and scene v, g u (t), g v (t) represent the predicted output values ​​of the tth period of scenario u and scenario v respectively (i.e., the predicted output values ​​of the tth period of day u and day v). u (t), g v In some embodiments, the predicted output values ​​of all new energy power station scenarios u and v in the tth period can be combined to obtain g u (t) and g v (t). Furthermore, the specific combination method can be summation, that is, summing the predicted output values ​​of all new energy power station scenarios u in the rth period as g u (t), the predicted output values ​​of all renewable energy power station scenarios v in the tth period are summed up as g v (r). Of course, this is only an example, and the present invention is not limited to this. In a specific embodiment, those skilled in the art can set it according to actual needs.

[0073] In the second step, for each current scene, the sum of its Euclidean distances with other scenes is obtained, and the scene with the smallest sum of Euclidean distances with other scenes is determined as a candidate scene, which can be specifically expressed as the following formula.

[0074]

[0075] In the formula, sd(u) represents the sum of the Euclidean distances between scene u and the remaining scenes, V represents the total number of remaining scenes, and u * The scene with the smallest sum of Euclidean distances from other current scenes is represented as the candidate scene.

[0076] The third step is to obtain the scene with the smallest Euclidean distance to the candidate scene from all current scenes, which can be specifically expressed as the following formula.

[0077]

[0078] Where, d u*,v Represents the candidate scene u * The Euclidean distance from scene v, v * It represents the scene with the smallest Euclidean distance to the candidate scene.

[0079] The fourth step is to add the probability of occurrence of the scene with the smallest Euclidean distance to the candidate scene to the probability of the candidate scene, and delete the scene with the smallest Euclidean distance to the candidate scene. * The probability of occurrence is updated to candidate scenario u * The probability of occurrence and scenario v * The sum of the probabilities of the scenario v * delete.

[0080] Step 5: Check whether the current number of remaining scenes has reached the preset number of remaining scenes. For example, in some embodiments, the preset number of remaining scenes is NRE, then check whether the current number of remaining scenes has reached NRE.

[0081] In step 6, if the current number of remaining scenarios does not reach the preset number of remaining scenarios, repeat the above steps 1 to 5 until the current number of remaining scenarios reaches the preset number of remaining scenarios, and obtain multiple typical scenarios and the probability of occurrence of each typical scenario. Taking the above example, NRE typical scenarios and the probability of occurrence of the NRE typical scenarios are obtained.

[0082] In the above 210, the predicted output values ​​of each new energy power station to be aggregated on each day of the year have been obtained, so the predicted output values ​​of each typical day of each new energy power station can be directly obtained. So far, the occurrence probabilities of multiple typical days and the predicted output values ​​of multiple typical days of each new energy power station are obtained.

[0083] Then, enter 230, based on the occurrence probability of multiple typical days and the predicted output values ​​of multiple typical days of each renewable energy power station, use the particle swarm algorithm, take the dispatching and operating benefits of the alliance structure as the fitness value, optimize the initialized multiple alliance structures, obtain the target alliance structure, and output the dispatching and operating plan of each renewable energy distribution storage in each alliance in the target alliance structure on each typical day.

[0084] It should be noted that any alliance structure includes at least one alliance. Any alliance includes at least one new energy power station equipped with energy storage. The dispatching operation plan of each new energy distribution storage on each typical day includes the capacity of each new energy distribution storage allocated to the frequency regulation auxiliary service market and the electric energy market on each typical day. In some embodiments, it can be specifically the capacity of each new energy distribution storage allocated to the frequency regulation auxiliary service market and the capacity allocated to the electric energy market in each period of each typical day.

[0085] Next, the method of obtaining the dispatching and operating income of any alliance structure is explained as follows.

[0086] The first step is to consider the penalty for new energy output deviation, take the maximization of the dispatching and operation benefits of each alliance in the alliance structure as the goal, and build a dispatching and operation model for each alliance participating in the frequency regulation ancillary service market.

[0087] Taking into account the limited energy storage resources, in order to balance the capacity reserved for REPP and the capacity participating in the frequency regulation ancillary service market, and to minimize the operating cost of REPP-energy storage, according to an embodiment of the present invention, for each alliance, a scheduling operation model of the alliance (or a day-ahead scheduling operation model of the alliance) can be constructed with the goal of maximizing the scheduling operation benefit of the alliance.

[0088] Specifically, the alliance's scheduling operation model includes the following objective functions.

[0089]

[0090] Among them, R year represents the alliance's scheduling and operation benefits, ρ k represents the probability of occurrence of a typical day k, represents the secondary frequency modulation income of the alliance in period t on a typical day k, represents the alliance’s on-grid electricity sales revenue during period t on a typical day k, represents the output deviation penalty of the alliance in period t on a typical day k, represents the energy storage operation cost of the alliance in time period t on typical day k, K represents the total number of typical days, and T represents the total number of time periods.

[0091] The following is an explanation of secondary frequency regulation revenue, grid-connected electricity sales revenue (also known as electricity market participation revenue), energy storage operating costs, and output deviation penalty costs.

[0092] 1) Secondary frequency modulation income

[0093] According to market rules, the frequency regulation service compensation fee obtained by the winning AGC unit includes frequency regulation capacity compensation and frequency regulation mileage compensation. Since the specific clearing amount of frequency regulation mileage can only be known in the intraday stage, this embodiment uses the historical capacity-mileage ratio to estimate the possible clearing amount of frequency regulation mileage in the day-ahead stage. Based on this, in some embodiments, the secondary frequency regulation income can be specifically obtained by the following formula.

[0094]

[0095] in, represents the unit capacity compensation price of the alliance participating in secondary frequency regulation during period t on typical day k (i.e., the unit capacity compensation price of the energy storage in the alliance participating in secondary frequency regulation during period t on typical day k), represents the capacity of the i-th energy storage (which can be a supporting storage or an independent energy storage) in the alliance to participate in secondary frequency regulation during the t period of a typical day k, represents the unit mileage compensation price for the alliance to participate in secondary frequency modulation during the t period of a typical day k, represents the frequency regulation mileage call rate, and I represents the total amount of energy storage in the alliance, specifically, the total amount of supporting storage and independent energy storage in the alliance.

[0096] 2) Income from on-grid electricity sales

[0097] The income from the sale of electricity on the grid can also be called the income from the alliance's participation in the electricity market or the income from the amount of electricity on the grid. Considering the total on-grid power of the alliance, the income from the sale of electricity on the grid can be obtained through the following formula.

[0098]

[0099] in, represents the unit grid-connected electricity price for period t on a typical day k, g j,k,t represents the actual grid-connected electricity of the j-th renewable energy power station REPP in the alliance during period t on a typical day k (i.e., the actual grid-connected electricity reported), and J represents the total number of renewable energy power stations REPP in the alliance.

[0100] 3) Energy storage operating costs

[0101] In order to maximize the operating income of the REPP-energy storage alliance, low-cost energy storage can be preferentially called to meet the tracking plan and frequency regulation capacity of REPP. Therefore, this embodiment incorporates the energy storage operating cost into the objective function. The energy storage operating cost can be specifically obtained by the following formula.

[0102]

[0103] in, represents the unit capacity operation and maintenance cost of the i-th energy storage in the alliance, represents the capacity of the i-th energy storage in the alliance to participate in secondary frequency regulation during period t of typical day k, It represents the capacity reserved by the i-th energy storage in the alliance during period t of typical day k to ensure the tracking plan of REPP planned output.

[0104] 4) Output deviation penalty cost

[0105] In order to balance the relationship between capacity allocation benefits and deviation penalties and maximize benefits, it is necessary to measure the deviation penalty cost. At this time, not only can the corresponding income not be obtained, but also the corresponding penalty must be deducted. Therefore, the deviation penalty can be obtained specifically through the following formula.

[0106]

[0107] in, represents the deviation penalty ratio for participating in the electric energy market, It represents the energy deviation of the i-th energy storage in the alliance during period t of a typical day k. The energy deviation is the deviation between the actual output of the cluster and the reported grid-connected output. Energy storage can reserve capacity to track the output of renewable energy to compensate for the deviation.

[0108] Furthermore, the dispatching and operation model of the alliance also includes constraints corresponding to the above objective function, including constraints on the declared power and discharge capacity of each new energy power station in the alliance, constraints on the charging and discharging capacity of each energy storage in the alliance, constraints on the state of charge of each energy storage in the alliance, and capacity constraints on the participation of each new energy power station in the alliance in the electricity market, which are explained in detail below.

[0109] 1) Constraints on the reported power and discharge volume of each new energy power station in the alliance: that is, constraints on the reported power and discharge volume in the electric energy market, as follows.

[0110]

[0111] Among them, g j,k,t represents the actual grid-connected power generation of the jth renewable energy power station REPP in the alliance during the t period of a typical day k; represents the actual discharge of the jth renewable energy power station REPP in the alliance during the t period of a typical day k; It represents the tracking output of the i-th energy storage (i.e. the energy storage equipped by the j-th new energy power station) in the alliance during the t period of a typical day k when participating in the electric energy market. When its value is positive, the actual output of the new energy is less than the predicted value. At this time, the energy storage discharge power is The charging power is 0. When the value is negative, the actual output of new energy is greater than the predicted value. At this time, the energy storage charging power is The discharge power is 0, so and The relationship can be specifically expressed as the following formula.

[0112]

[0113]

[0114] in, They represent the charging power and discharging power of the i-th energy storage in the alliance at time t on a typical day k, Respectively Take the negative and take the positive.

[0115] 2) Charging and discharging capacity constraints of each energy storage in the alliance: Considering the safe operation of the power grid, when declaring energy storage, the capacity should meet its physical constraints, as follows.

[0116]

[0117] in, represents the charging capacity of the i-th energy storage in the alliance at time t on a typical day k, represents the frequency regulation capacity of the i-th energy storage in the alliance during the t period of a typical day k, represents the capacity of the i-th energy storage in the alliance to reduce the output of renewable energy in the t period of a typical day k, represents the maximum charging capacity of the i-th energy storage in the alliance, represents the discharge capacity of the i-th energy storage in the alliance during period t on a typical day k, represents the upward frequency regulation capacity of the i-th energy storage in the alliance at period t on a typical day k, represents the capacity of the i-th energy storage in the alliance to increase the output of renewable energy in the t period of a typical day k, represents the maximum discharge capacity of the i-th energy storage in the alliance.

[0118] 3) State of charge constraints of each energy storage in the alliance: In actual operation, the state of charge of the energy storage specifically satisfies the following constraints.

[0119]

[0120] SOC i,min ≤SOC i,k,t ≤SOC i,max

[0121] Among them, SOC i,k,t , SOC i,k,t-1 They represent the state of charge of the i-th energy storage in the alliance at period t and period (t-1) on a typical day k, respectively. They represent the charging efficiency and discharging efficiency of the i-th energy storage in the alliance, represents the frequency regulation mileage of the i-th energy storage in the alliance during the t period of a typical day k, E i,max represents the capacity of the i-th energy storage in the alliance, Δt represents the duration of a period, SOC i,min , SOC i,max They represent the minimum charge and maximum charge of the i-th energy storage in the alliance respectively.

[0122] 4) Capacity constraints of each renewable energy power station in the alliance participating in the electricity market: This can also be called the capacity constraints of the renewable energy output tracking plan, in which the reserve capacity of energy storage for the uncertainty of REPP output should meet the predicted error range, and the regulation of energy storage on the joint output of REPP-energy storage includes downward regulation and upward regulation. Downward regulation means that when the actual output of REPP is higher than the predicted output, the energy storage performs charging action, and upward regulation means that when the actual output of REPP is lower than the predicted output, the energy storage performs discharging action. The allocation of energy storage capacity according to time periods is too complicated, so the capacity of the renewable energy output tracking plan can be set as a certain proportion of the predicted output of renewable energy, as shown in the following formula.

[0123]

[0124]

[0125] in, represents the predicted power generation (i.e., predicted output value) of the j-th renewable energy power station REPP in the alliance during the t period of a typical day k, and They represent the upward capacity allocation coefficient and downward capacity allocation coefficient of the j-th renewable energy power station REPP in the alliance, They respectively represent the capacity of increasing and decreasing the output of renewable energy of the i-th energy storage (i.e., the energy storage equipped by the j-th renewable energy power station) in the alliance during period t of a typical day k.

[0126] At this point, a dispatching operation model for each alliance in the alliance structure to participate in the frequency regulation ancillary service market has been constructed, and the next step is to proceed to the second step.

[0127] In the second step, based on the occurrence probability of multiple typical days and the predicted output values ​​of each new energy power station on multiple typical days, the dispatching and operation models of the constructed alliances participating in the frequency regulation auxiliary service market are solved to obtain the dispatching and operation plans of each new energy distribution and storage in each alliance on each typical day and the dispatching and operation benefits of each alliance under the corresponding dispatching and operation plans.

[0128] According to an embodiment of the present invention, when solving the dispatching and operation model of each alliance participating in the frequency regulation auxiliary service market, the actual grid-connected power output values ​​(also referred to as actual grid-connected power) of each new energy power station on multiple typical days need to be input into the model. Among them, considering that some of the predicted output values ​​of multiple typical days of each new energy power station may not meet the technical regulations for wind farm access to the power system, this embodiment further adjusts the actual reported power of the new energy station.

[0129] Specifically, the national standard technical regulations for wind farm access to the power system and the technical regulations for photovoltaic power station access to the power system stipulate the active power variation limit of new energy power stations. Based on this, in some embodiments, a decision-making model for the actual grid-connected capacity reported by new energy power stations can be constructed with the goal of minimizing the distance between the grid-connected output and the predicted output to make decisions on the actual grid-connected capacity of new energy power stations.

[0130] According to one embodiment of the present invention, the decision model for the actual grid-connected capacity reported by a new energy power station includes the following objective function.

[0131]

[0132] Among them, f represents the distance between the online output and the predicted output, g j,k,t represents the actual grid-connected power generation of the jth renewable energy power station REPP during the t period of a typical day k, It represents the predicted power generation of the j-th renewable energy power station REPP in the t period of typical day k.

[0133] Furthermore, the decision model for the actual grid-connected capacity reported by the new energy power station also includes constraints corresponding to the objective function, specifically including constraints on the output change rate of the new energy power station, and constraints on the daily grid-connected capacity and daily predicted capacity, as follows.

[0134] Constraints on the output change rate of new energy power stations: The output change rate of new energy power stations cannot exceed the limit, as shown in the following formula.

[0135]

[0136] Among them, g j,k,t+10min represents the actual grid-connected power of the j-th renewable energy power station REPP during the period (t+10min) on a typical day k, Represents the installed capacity of the j-th renewable energy power station REPP.

[0137] Constraints on daily online access and daily forecast: In order to make the most of new energy, the online access and forecast can be set to be the same within a day, as follows.

[0138]

[0139] By solving the decision model for the actual grid-connected capacity application of the above-mentioned new energy power station, the actual grid-connected capacity of the new energy power station can be obtained. Subsequently, by solving the dispatching and operation model of each alliance participating in the frequency regulation auxiliary service market, the dispatching and operation plan of each new energy storage in each alliance on each typical day and the dispatching and operation income of each alliance under the corresponding dispatching and operation plan can be obtained. The solution of the model can be completed using the relevant solver, which will not be repeated here.

[0140] The third step is to obtain the dispatching and operating benefits of the alliance structure based on the dispatching and operating benefits of each alliance. Specifically, the dispatching and operating benefits of each alliance are added together, and the sum obtained is the dispatching and operating benefits of the alliance structure.

[0141] The above is the way to obtain the dispatching operation benefits of any alliance structure. Next, how to use the particle swarm algorithm to optimize the alliance structure to obtain the target alliance structure (i.e., the most stable alliance structure) is explained as follows.

[0142] Step 1: Initialize the number of alliance structures, the speed and position of each alliance structure, and the number of iterations. In some embodiments, the initial inertia weight and the inertia weight when iterating to the maximum number of generations are also set.

[0143] Among them, the position of each alliance structure can be composed of the number of alliances under the alliance structure and the number of the alliances participated by each new energy power station. Further, the position of each alliance structure can be expressed as a vector, and the elements in the vector are the number of alliances under the alliance structure and the number of the alliances participated by each new energy power station.

[0144] Specifically, the number of new energy power stations to be aggregated is M, then the position x of the nth alliance structure particle is n It is represented by a (M+1)-dimensional vector, where the first element x n,t (1) represents the number of alliances under the nth alliance structure at the tth iteration. The subsequent element x n,t (2) x n,t (3)…x n,t (M+1) indicates the number of each new energy power station participating in the alliance.

[0145] In addition, the location of each alliance structure satisfies the following two constraints.

[0146]

[0147] 1≤x n,t (2),x n,t (3),…,x n,t (M+1)≤x n,t (1)

[0148] The speed of the alliance structure is:

[0149] v n,t+1 =v n,t +c1r1[p n,best -x n,t ]+c2r2[p best -x n,t ]

[0150] Among them, v n,t+1 is the speed of the nth particle at the (t+1)th iteration, c1 represents the individual learning factor, r1 and r2 represent uniformly distributed random numbers in the range [0,1], c2 represents the group learning factor, and p n,best represents the optimal individual extreme value of the nth particle itself, p best represents the global optimal extreme value of the particle swarm, x n,t Represents the position of the nth particle at the tth iteration.

[0151] Step 2: Based on the current position of each alliance structure, obtain the fitness value of each alliance structure.

[0152] The above has explained the method of obtaining the dispatching and operating benefits of each alliance structure, and this embodiment uses the dispatching and operating benefits of the alliance structure as the fitness value. Therefore, after obtaining the predicted output values ​​of multiple typical days of each new energy power station to be aggregated and the probability of occurrence of each typical day, the fitness value of each alliance structure can be obtained based on the current position of each alliance structure.

[0153] Step three: According to the obtained fitness values ​​of each alliance structure, determine the optimal individual position of each alliance structure and the optimal group position of the alliance structure group, and detect whether the current number of initialization iterations has been reached or whether all the new energy power stations to be aggregated have become an alliance.

[0154] Specifically, for each alliance structure, the optimal individual position of each alliance structure is determined by comparing its current fitness value with its historical optimal fitness value, and then the optimal group position is determined based on the final individual position.

[0155] If the number of initialization iterations has not been reached and all the new energy power stations to be aggregated have not formed an alliance, then proceed to step 4.

[0156] Step 4: Add one to the current number of iterations, update the speed and position of each alliance structure, and continue to execute steps 2 to 4 until the initial number of iterations is reached or all new energy power stations to be aggregated become an alliance. Stop the iteration, take the current optimal group position as the final position of the alliance structure, and obtain the target alliance structure.

[0157] Among them, the speed and position of the alliance structure are updated according to the particle speed update formula and particle position update formula of the particle swarm algorithm, which will not be repeated here. In addition, when obtaining the fitness value of each alliance structure, the scheduling and operation plan of each new energy storage in each alliance in the alliance structure on a typical day has been obtained. Therefore, after obtaining the target alliance structure, the scheduling and operation plan of each new energy storage in each alliance in the target alliance structure on a typical day can be obtained.

[0158] It can be seen that the dispatching and operation optimization of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market is divided into two stages. The first stage needs to consider the stability of the alliance to optimize the alliance structure. The goal of the second stage is to further optimize the allocation of energy storage capacity on the basis of the corresponding alliance structure to determine the energy storage capacity required for error tracking of new energy and the capacity for participating in frequency regulation auxiliary services, so as to maximize the alliance's benefits. For a better understanding, the present invention also provides an embodiment of the dispatching and operation method of the new energy distribution and storage jointly participating in the frequency regulation auxiliary service market. Figure 5 Describe the specific process.

[0159] (1) In data processing:

[0160] 1) Use the trained day-ahead output prediction model to predict the day-ahead output of renewable energy;

[0161] 2) Use the probability reduction method to generate annual typical scenarios and their probability of occurrence.

[0162] (2) During data input:

[0163] The alliance data, typical day wind and solar output and forecast data are input into the solution model. The alliance data includes exogenous parameters related to the alliance optimization operation, such as energy storage capacity, installed capacity of new energy, etc.

[0164] (3) Solving the two-stage model:

[0165] In order to solve such a two-stage model, this embodiment sets up an evolutionary cooperative game solving algorithm based on a particle swarm algorithm, which is specifically as follows.

[0166] 1) Initialize the velocity and position of the alliance structure particles.

[0167] First, set the parameters required by the algorithm, including the number of iterations G and the number of particles N l , initial inertia weight ω int , the inertia weight ω when the iteration reaches the maximum number of generations end , the number of entities M.

[0168] The nth (n∈N l ) The position x of the alliance structure particle n,tIt can be represented as a (M+1)-dimensional vector, where x n,t (1) represents the number of alliances under this alliance structure, [x n,t (2),x n,t (3),…,x n,t (M+1)] represents the number of each subject participating in the alliance. The particle position needs to satisfy the following two constraints.

[0169]

[0170] 1≤x n,t (2),x n,t (3),…,x n,t (M+1)≤x n,t (1)

[0171] The nth (n∈N l ) The position v of the alliance structure particle n It can also be represented as a (M+1)-dimensional vector.

[0172] 2) After defining the position and speed of the particles, it is necessary to enter the second stage of the optimal solution of the alliance operation scheduling. After solving it using the relevant solver, the optimal benefit of the corresponding alliance structure particle is obtained. Then the fitness of the alliance structure particle is calculated. Among them, R n (i) represents the income of the ith alliance in the alliance structure. In order to prove the rationality of this fitness function, the advantages of the alliance under different income conditions are described below.

[0173] when There must be at least one subject that is allocated less benefits in the m-alliance structure than in the n-alliance structure.

[0174] ① When only one subject A is allocated less, and the benefits of other subjects in the two alliance structures remain unchanged, A is more inclined to choose the n-alliance structure, so n>m.

[0175] ② When a part of the subject B is allocated less than n in the m-alliance structure, and another part of the subject C is allocated more than n in the m-alliance structure, sum(|R - (B)|)>sum(R + (C)), that is, the smaller part is greater than the larger part, and B will be more motivated to dissolve alliance m and form alliance n. Therefore, n>m.

[0176] 3) Calculate the individual optimum and group optimum, and update the position and velocity of the particles.

[0177] 4) When the alliance reaches the maximum number of iterations or all entities become a coalition, the iteration stops, otherwise continue with 2)-4).

[0178] The above is the specific process of the operation method of the new energy distribution and storage jointly participating in the frequency regulation and electric energy market of this embodiment. For further understanding, the following is an example of four entities equipped with energy storage, namely, wind power plant 1, photovoltaic power plant 1, wind power plant 2, and photovoltaic power plant 2. The initial alliance structure particles only contain two alliance structures of [2,1,2,1,2] (indicating that two alliances are formed, the first entity is in the first alliance, the second entity is in the second alliance, the third entity is in the first alliance, and the fourth entity is in the second alliance) and [3,3,2,3,1]. After optimizing the alliance decisions respectively, the benefits of the first and second alliances of the first alliance structure are 10 and 20 respectively, and the benefits of the 1st, 2nd, and 3rd alliances of the second alliance structure are 2, 3, and 5 respectively. The alliance structure fitness is 30 and 10 respectively, and the first alliance is more stable. At this time, the [2,1,2,1,2] particle shares its position, that is, the particles of the next generation tend to be closer to [2,1,2,1,2]. After multiple iterations, the solution becomes stable and the alliance structure can be output.

[0179] In addition, this embodiment also verifies the effectiveness of the proposed operation strategy through simulation examples of 5 typical 24-hour periods a day ago, as follows.

[0180] 1. Basic parameters

[0181] The market price of electric energy, frequency regulation capacity and mileage refer to the PJM market price, such as Figure 6 shown.

[0182] The frequency modulation mileage call rate uses the ancillary service market operation data disclosed by the relevant PJM market, among which the frequency modulation call rate every five minutes is obtained by randomly sampling the frequency modulation call every five minutes, see Figure 7 When its value is negative, it indicates downward frequency modulation, and when its value is positive, it indicates upward frequency modulation. The sum of the calling ratio multiplied by the absolute value of the frequency modulation capacity is the frequency modulation mileage calling rate for this hour.

[0183] According to the power supply side energy storage installed capacity data of a certain province, this embodiment sets the power supply side energy storage participating in the aggregation to include a lithium iron phosphate battery, two compressed air energy storages, and a full-solid flow energy storage. The capacities of the four energy storages are 510, 300, 110, and 120MWh, respectively, and the maximum charging and discharging powers are 255, 150, 55, and 60MW, respectively. The new energy output refers to the wind and solar output of PJM in the past year. The annual utilization hours of a certain province are used. The annual utilization hours of wind power are 2500 hours, and the annual utilization hours of photovoltaic power are 1588.5. The installed capacity is calculated, and then the energy storage installed capacity ratio is used to scale the new energy output proportionally. Finally, the new energy sites are set to be two wind power plants and two photovoltaic plants, with installed capacities of 1700, 1000, 200, and 400MW, respectively. The storage capacity and installed capacity are shown in Table 1 below.

[0184] Table 1

[0185]

[0186] After using the ELMAN NN neural network to predict the annual wind and solar power output, the daily prediction error was reduced based on the distance rapid reduction method to obtain 5 typical daily scenarios, whose scenario occurrence probabilities are 0.003, 0.622, 0.2202, 0.1518, and 0.003, respectively. The data are as follows Figure 8 shown. Figure 8 The moments 1 to 24, 25 to 48, 49 to 72, and 73 to 96 are the errors of wind power plant 1, photovoltaic power plant 1, wind power plant 2, and photovoltaic power plant 2 in this scenario respectively.

[0187] 2. Analysis of the effectiveness of the solution

[0188] In order to verify that the solution proposed in this embodiment can increase the revenue of new energy storage and distribution without causing system instability, this implementation also sets up a solution comparison as shown in Table 2 below.

[0189] Table 2

[0190]

[0191]

[0192] The benefits of the above schemes are compared as follows: Fig. 9 As shown in the figure, it can be seen that the benefits of Scheme 2 are improved to a certain extent compared with Scheme 1. This is because the prediction errors of different renewable energy power plants have certain time complementarity. Scheme 3 is improved to a certain extent compared with Scheme 2 because energy storage reduces the deviation penalty caused by errors. Scheme 4 is greatly improved compared with Scheme 3, indicating that the participation of renewable energy power plants in the electricity energy market and frequency regulation auxiliary service market has good economic performance.

[0193] Fig. 9 The red line is the error absorption rate under different schemes, where the new energy absorption rate indicates the proportion of the forecast error reduced by the scheme to the total error under this scheme. It can be seen from the figure that the absorption rate of scheme 2 has increased, which is because the forecast errors of different new energy power fields have a certain time complementarity. The error absorption rate of scheme 3 has been greatly improved due to the addition of energy storage. In the case of increased income in scheme 4, the error absorption rate has not been greatly reduced. This is because the alliance will use the idle time of energy storage to participate in frequency regulation auxiliary services. For example, it will not participate in the electricity market during the night time of photovoltaic power generation, and energy storage will not allocate corresponding capacity to deal with the forecast deviation of new energy. At this time, this part of capacity will be used to participate in the frequency regulation auxiliary service market in scheme 4.

[0194] 3. Results Analysis

[0195] 1) Alliance structure stability analysis and benefit distribution

[0196] Fig.10 The result of the alliance structure iteration is shown in the figure. After 20 iterations, the result is stable. The stable alliance structure is {wind 1, light 1, wind 2, light 2}, that is, all entities are allied together. In order to verify the stability of the alliance, the exhaustive method is used to calculate the possible benefits of all alliances as shown in Table 3 below.

[0197] Table 3

[0198]

[0199] After calculating the above data, it can be seen that the alliance {wind 1, light 1, wind 2, light 2} meets the conditions for the stability of the allocation strategy. The alliance constructed in this embodiment is stable and can ensure the existence of this alliance. The reason for forming this alliance is that there is a certain time complementarity between the prediction errors of different power plants. At the same time, aggregating energy storage can ensure that energy storage meets the demand for new energy while having more capacity for frequency regulation auxiliary services.

[0200] The relevant results obtained after allocating the cooperative game according to the Shapley value are as follows Fig.11 As shown. According to the blue line in the figure, as the installed capacity of alliance members increases, their benefits also increase accordingly. The unit energy storage installed power benefit represents the unit installed power benefit of different entities minus the benefits of new energy power generation. The result after Min-Max normalization is shown in the figure. It can be seen that the unit energy storage installed power benefit is inversely proportional to the new energy error under the unit energy storage installed capacity. This shows that when the model considers the penalty for electric energy deviation, the alliance will allocate more energy storage capacity to compensate for the prediction error in order to avoid deviation penalty costs. At the same time, this also prevents storage from participating in the corresponding frequency regulation auxiliary services. However, when the model does not consider the deviation penalty, it should be considered in the allocation stage, otherwise it will cause unfairness.

[0201] 2) Analysis of capacity allocation results

[0202] The capacity of the energy storage system is limited. How can we meet the output tracking requirements of new energy units while using the remaining capacity to participate in the frequency regulation market? Fig.12 The results of energy storage capacity allocation under different typical days. Under this scheme, the absorption rate of new energy reaches 94.67%, and it still has the capacity to participate in the frequency regulation auxiliary service market, which shows that it is feasible for new energy to share energy storage to participate in the frequency regulation auxiliary service market. This is because: (1) When new energy participates in the electricity market together, their prediction errors will offset each other, which to a certain extent avoids the simultaneous charging and discharging of different energy storage. (2) The output of renewable energy is low in some periods, and the required energy storage capacity is also low. At this time, the energy storage capacity can be used to participate in the frequency regulation auxiliary service market. For example, if the photovoltaic unit does not generate electricity at night, there will be no corresponding output deviation. In this case, its energy storage can be used to participate in the frequency regulation market. (3) The use of neural networks to predict the output of new energy is relatively accurate most of the time, but only some scenarios are less accurate, and more energy storage is needed to track the output of new energy, such as typical day 4.

[0203] 4. Sensitivity analysis

[0204] 1) Sensitivity analysis of power energy deviation penalty rate

[0205] Since the electric energy deviation penalty rate has a certain impact on the allocation of energy storage capacity, when the deviation penalty rate is too low, the capacity of energy storage allocated to new energy will be too small, and a larger portion of the energy storage capacity will be used to participate in frequency regulation auxiliary services, which will affect the safe and stable operation of the power system; when the deviation penalty rate is too high, the profits of market entities will drop significantly, but the improvement on error absorption will not be obvious. Therefore, this embodiment conducts a sensitivity analysis on the electric energy deviation penalty rate, aiming to select a suitable electric energy deviation penalty rate range. Fig.13 This is a sensitivity analysis of the electricity energy deviation penalty rate. It can be seen from the figure that when the electricity energy deviation penalty rate changes from 0.2 to 1.2, the new energy consumption rate increases significantly, but the alliance revenue decreases not significantly; when the electricity energy deviation penalty rate changes from 1.2 to 1.4, the new energy consumption rate does not increase significantly. Therefore, the electricity energy deviation penalty rate should be selected in the range of 0.4-1.2.

[0206] 2) Sensitivity analysis of storage ratio of new energy power plants

[0207] Fig.14 and Fig.15 The alliance benefits and forecast error absorption rate under different allocation and storage ratios are shown respectively. Fig.14 and Fig.15It can be seen that when the allocation and storage ratio increases from a smaller position to a certain amount, the alliance revenue and the forecast error absorption rate increase sharply at the same time, which shows that the increase in alliance revenue mainly depends on reducing the penalty cost brought by the forecast error; when the allocation and storage ratio increases to a certain proportion, its change rate is basically stable. At the same time, it can be seen that the rate of change of alliance revenue and forecast error absorption rate with the change of the proportion of allocation and storage power to installed capacity is greater than the rate of change with the change of energy storage duration, which shows that both frequency regulation and new energy allocation and storage require power-type energy storage rather than capacity-type energy storage.

[0208] 3) Sensitivity analysis of new energy change rate

[0209] Since the renewable energy change rate limit has a certain impact on the allocation of energy storage capacity, when the renewable energy change rate limit is low, it will make it difficult to absorb some renewable energy even if they are equipped with energy storage of specified capacity; when the renewable energy change rate limit is too high, the changes in renewable energy participating in the electric energy market will fluctuate too much, which will affect the safe and stable operation of the power system. Therefore, this embodiment conducts a sensitivity analysis on the renewable energy change rate limit, aiming to select a suitable renewable energy change rate limit. Fig.16 This is a sensitivity analysis of the limit of the change rate of new energy. It can be seen from the figure that with the increase of the limit of the change rate of new energy, the alliance's benefits show an upward trend. This shows that under a higher limit of the change rate of new energy, most of the new energy output in the alliance can meet its limit requirements with the help of energy storage, so as to make more effective use of new energy and thus bring higher economic benefits. It can be seen that when the fluctuation limit is 40, the energy storage equipped according to regulations can basically realize the consumption of new energy, but the establishment of this policy must not only consider the consumption of new energy, but also the power system consumption capacity under different proportions of new energy installed capacity.

[0210] In summary, the present invention proposes a cluster mode of energy storage for aggregation-friendly new energy power stations, constructs a two-stage scheduling optimization model for cluster frequency regulation auxiliary services, and gives full play to the flexible regulation role of energy storage resources. First, a scheduling framework for the joint participation of new energy storage clusters in the frequency regulation auxiliary service market is constructed; secondly, considering constraints such as the upper limit of energy storage capacity and the limit of fluctuation in new energy output, a scheduling model for the collaborative participation of new energy and energy storage in the electric energy and auxiliary service market is established with the goal of maximizing cluster benefits under deviation assessment. Finally, an evolutionary cooperative game solving algorithm for the particle swarm algorithm is proposed based on the fitness value of the alliance structure to obtain a stable alliance structure. The present invention can not only improve the energy storage utilization rate of new energy, but also enable it to effectively participate in frequency regulation assistance, thereby improving the quality of power supply to the power grid.

[0211] The various techniques described herein may be implemented in combination with hardware or software, or a combination thereof. Thus, the method and apparatus of the present invention, or some aspects or parts of the method and apparatus of the present invention may be in the form of program codes (i.e., instructions) embedded in a tangible medium, such as a removable hard disk, a USB flash drive, a floppy disk, a CD-ROM, or any other machine-readable storage medium, wherein when the program is loaded into a machine such as a computer and executed by the machine, the machine becomes a device for practicing the present invention.

[0212] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0213] It should be understood that in order to streamline the present disclosure and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this method of disclosure should not be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim.

[0214] In addition, unless otherwise specified, the use of ordinal numbers "first", "second", "third", etc. to describe common objects merely indicates that different instances of similar objects are involved, and is not intended to imply that the objects so described must have a given order in time, space, ranking or in any other manner.

[0215] Although the present invention has been described according to a limited number of embodiments, it will be apparent to those skilled in the art, with the benefit of the above description, that other embodiments may be envisioned within the scope of the invention thus described. In addition, it should be noted that the language used in this specification is selected primarily for readability and teaching purposes, rather than for explaining or defining the subject matter of the present invention. Therefore, many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the appended claims. The disclosure of the present invention is illustrative, not restrictive, with respect to the scope of the present invention, which is defined by the appended claims.

Claims

1. A dispatching and operating method for new energy distribution and storage to jointly participate in the frequency regulation auxiliary service market, the method comprising: Inputting the output values ​​of the previous predetermined number of days of each new energy power station to be aggregated into the trained day-ahead output prediction model to generate the predicted output value of each new energy power station on each day, wherein the new energy power station is equipped with energy storage; Based on the predicted output values ​​of each renewable energy power station on each day, the daily scenarios are reduced using the probability reduction method to obtain the occurrence probabilities of multiple typical days and the predicted output values ​​of each renewable energy power station on multiple typical days; Based on the occurrence probabilities of multiple typical days and the predicted output values ​​of multiple typical days of each new energy power station, a particle swarm algorithm is used to optimize the initialized multiple alliance structures with the dispatching and operating benefits of the alliance structure as the fitness value, and a target alliance structure is obtained. The dispatching and operating plans of each new energy distribution storage in each alliance in the target alliance structure on each typical day are output, and the dispatching and operating plans include the capacity of each new energy distribution storage allocated to the frequency regulation auxiliary service market and the electric energy market on each typical day, and any alliance includes at least one new energy power station equipped with energy storage; Among them, the dispatching and operating benefits of any alliance structure are obtained in the following ways: Taking into account the penalty for deviation in output of renewable energy, and taking the maximum dispatching and operating benefits of each alliance in the alliance structure as the goal, a dispatching and operating model for each alliance participating in the frequency regulation auxiliary service market is constructed; Based on the probability of occurrence of multiple typical days and the predicted output values ​​of each renewable energy power station on multiple typical days, the dispatching and operation model of each alliance participating in the frequency regulation auxiliary service market is solved to obtain the dispatching and operation plan of each renewable energy storage in each alliance on each typical day and the dispatching and operation income of each alliance under the corresponding dispatching and operation plan; Based on the dispatching and operating benefits of each alliance obtained, the dispatching and operating benefits of the alliance structure are obtained.

2. The method of claim 1, wherein: For each alliance, its scheduling operation model includes an objective function, which includes: Among them, R year represents the alliance's scheduling and operation benefits, ρ k represents the probability of occurrence of a typical day k, represents the secondary frequency modulation income of the alliance in period t on a typical day k, represents the alliance’s on-grid electricity sales revenue during period t on a typical day k, represents the output deviation penalty of the alliance in period r on a typical day k, represents the energy storage operation cost of the alliance in time period t on typical day k, K represents the total number of typical days, and T represents the total number of time periods.

3. The method according to claim 1 or 2, wherein: The particle swarm algorithm is used to optimize the initialized multiple alliance structures with the scheduling and operation benefits of the alliance structure as the fitness value to obtain the target alliance structure, including: Initialize the number of alliance structures, the speed and position of each alliance structure, and the number of iterations; Based on the current position of each alliance structure, obtain the fitness value of each alliance structure; According to the fitness values ​​of each alliance structure obtained, the optimal individual position of each alliance structure and the optimal group position of the alliance structure group are determined, and it is detected whether the initialization number of iterations has been reached or whether all new energy power stations have become an alliance; If the current number of iterations has not been reached and all the new energy power stations have not formed an alliance, the current number of iterations will be increased by one, and the speed and position of each alliance structure will be updated, and the above fitness value acquisition, optimal individual position and optimal group position determination, and current number of iterations and alliance compliance detection steps will be repeated until the initialization number of iterations is reached or all the new energy power stations form an alliance, and the current optimal group position will be used as the final position of the alliance structure to obtain the target alliance structure.

4. The method of claim 3, wherein: The position of each alliance structure is composed of the number of alliances under the alliance structure and the number of the alliances in which each new energy power station participates.

5. The method according to any one of claims 1 to 4, wherein: The day-ahead output prediction model adopts the Elman neural network model.

6. The method according to any one of claims 1 to 5, wherein: The day-ahead output prediction model is trained based on the following method: Input the actual output values ​​of a predetermined number of days before the prediction date in the training set into the pre-trained day-ahead output prediction model to obtain the predicted output value of the prediction day, where the prediction day is any day in the training set; Based on the loss value between the actual output value and the predicted output value of the predicted day, the parameters of the day-ahead output prediction model are updated until the loss value meets a predetermined condition, and the training is completed to obtain a trained day-ahead output prediction model.

7. The method of claim 2, wherein: For each alliance, its dispatching and operation model also includes constraints, including constraints on the declared power and discharge capacity of each new energy power station in the alliance, constraints on the charging and discharging capacity of each energy storage in the alliance, constraints on the charge state of each energy storage in the alliance, and capacity constraints on the participation of each new energy power station in the alliance in the electricity market.

8. The method of claim 2, wherein: The output deviation penalty of the alliance in the t period of a typical day k is obtained by the following formula: in, represents the deviation penalty ratio for participating in the electric energy market, represents the unit grid-connected electricity price in period t of a typical day k, represents the electric energy deviation of energy storage i in period t of typical day k, and I represents the total amount of energy storage.

9. A computing device comprising: at least one processor; as well as A memory storing program instructions, wherein the program instructions are configured to be executed by the at least one processor, and the program instructions include instructions for executing the method according to any one of claims 1 to 8.

10. A readable storage medium storing program instructions, when the program instructions are read and executed by a computing device, the computing device executes the method according to any one of claims 1 to 8.

Images