Demand response method and device for distributed load resources

Abstract

The invention discloses a demand response method and device for distributed load resources. The method comprises the following steps: acquiring a plurality of pieces of historical response data of a plurality of distributed load resources; determining a posterior random forest model according to the plurality of historical response data; when demand response invitation information from the power grid is received, multiple pieces of current operation data of the multiple distributed load resources are obtained; determining a plurality of probability density functions according to the plurality of current operation data and a posterior random forest model; and determining a target declaration amount according to the demand response invitation information and the plurality of probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declaration amount. The probability density function is determined according to the current operation data and the posterior random forest model to dynamically represent the response success rate of the distributed load resources, so that the target declaration amount is obtained, the calculation cost of the declaration amount is reduced, and the accuracy of the declaration amount is improved. Therefore, the response success rate and economic benefits of the virtual power plant participating in the demand response invitation are improved.

Description

Technical Field

[0001] This application relates to the field of information processing technologies, and particularly relates to a demand response method and device for distributed load resources. Background Art

[0002] When a virtual power plant participates in a demand response invitation from the power grid, it is necessary to determine the declared quantity that takes into account both the response capabilities of distributed load resources and the maximization of benefits for response. In the prior art, methods such as elastic price matrix modeling and linear efficiency assumption are used to determine the declared quantity, which have high requirements for data samples. Due to the limitations of data privacy protection, it is difficult to obtain comprehensive parameter information for accurate modeling and calculation, resulting in high calculation costs and uncertainties for the obtained declared quantity. This makes it difficult for distributed load resources to successfully complete the declared quantity, leading to a low response success rate and poor economic benefits. Therefore, how to improve the response success rate and economic benefits of virtual power plants participating in demand response invitations has become a technical problem that needs to be further solved. Summary of the Invention

[0003] This application proposes a demand response method and device for distributed load resources to solve the problems of low success rate of virtual power plants participating in demand response invitations and poor economic benefits, and to improve the response success rate and economic benefits of virtual power plants participating in demand response invitations.

[0004] In a first aspect, an embodiment of this application provides a demand response method for distributed load resources, which is applied to a server in a virtual power plant. The method includes:

[0005] Obtain multiple historical response data of multiple distributed load resources, where a single historical response data is used to characterize the historical response situation of the corresponding distributed load resource;

[0006] Determine a posteriori random forest models according to the multiple historical response data, where the a posteriori random forest models are used to indicate the predicted response situations corresponding to the multiple distributed load resources;

[0007] When receiving demand response invitation information from the power grid, obtain multiple current operation data of the multiple distributed load resources, where a single current operation data is used to characterize the current operation situation of the corresponding distributed load resource;

[0008] Determine multiple probability density functions according to the multiple current operation data and the a posteriori random forest models, where a single probability density function is used to indicate the response success rate of the corresponding distributed load resource;

[0009] Determine a target declared quantity according to the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declared quantity.

[0010] In a second aspect, an embodiment of the present application provides a demand response device for distributed load resources, which is applied to a server in a virtual power plant. The device includes:

[0011] A first receiving unit, configured to obtain multiple historical response data of multiple distributed load resources, where a single historical response data is used to characterize the historical response situation of the corresponding distributed load resource;

[0012] A first processing unit, configured to determine a posteriori random forest models according to the multiple historical response data, where the posteriori random forest models are used to indicate the predicted response situations of the multiple distributed load resources; when receiving demand response invitation information from the power grid, obtain multiple current operation data of the multiple distributed load resources, where a single current operation data is used to characterize the current operation situation of the corresponding distributed load resource; determine multiple probability density functions according to the multiple current operation data and the posteriori random forest models, where a single probability density function is used to indicate the response success rate of the corresponding distributed load resource; and determine a target declaration quantity according to the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declaration quantity.

[0013] In a third aspect, an embodiment of the present application provides a server, including a processor, a memory, and one or more programs. The one or more programs are stored in the memory and are configured to be executed by the processor. The programs include instructions for performing the steps in the method according to any one of the first aspects.

[0014] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which computer programs / instructions are stored. When the computer programs / instructions are executed by a processor, the steps of the method according to any one of the first aspects are implemented.

[0015] In a fifth aspect, an embodiment of the present application provides a computer program product, including computer programs / instructions. When the computer programs / instructions are executed by a processor, some or all of the steps of the method according to any one of the first aspects of the embodiments of the present application are implemented.

[0016] It can be seen that in this application, multiple historical response data of multiple distributed load resources are obtained, and a single historical response data is used to characterize the historical response situation of the corresponding distributed load resource; a posteriori random forest model is determined according to the multiple historical response data, and the a posteriori random forest model is used to indicate the predicted response situation corresponding to the multiple distributed load resources; when receiving demand response invitation information from the power grid, multiple current operation data of the multiple distributed load resources are obtained, and a single current operation data is used to characterize the current operation situation of the corresponding distributed load resource; multiple probability density functions are determined according to the multiple current operation data and the a posteriori random forest model, and a single probability density function is used to indicate the response success rate of the corresponding distributed load resource; a target declaration quantity is determined according to the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declaration quantity. In this way, the probability density function is determined according to the current operation data and the a posteriori random forest model to dynamically characterize the response success rate of the distributed load resource, so as to obtain the target declaration quantity. This not only eliminates the need for precise prior assumptions and reduces the dependence on data samples, but also can quantify the uncertainty of the distributed load resources aggregated by the virtual power plant, significantly improving the prediction accuracy of the response ability of the distributed load resources, reducing the calculation cost of the declaration quantity, improving the accuracy of the declaration quantity, and thus improving the response success rate and economic benefits of the virtual power plant participating in the demand response invitation. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0018] Figure 1 is a schematic structural diagram of a virtual power plant provided by an embodiment of the present application;

[0019] Figure 2 is a schematic structural diagram of a power system provided by an embodiment of the present application;

[0020] Figure 3 is a schematic structural diagram of a server in a virtual power plant provided by an embodiment of the present application;

[0021] Figure 4 is a schematic flow chart of a demand response method for distributed load resources provided by an embodiment of the present application;

[0022] Figure 5 is a schematic flow chart of another demand response method for distributed load resources provided by an embodiment of the present application;

[0023] Figure 6 is a schematic flowchart of another demand response method for distributed load resources provided by an embodiment of the present application;

[0024] Figure 7 is a schematic flowchart of yet another demand response method for distributed load resources provided by an embodiment of the present application;

[0025] Figure 8 is a schematic diagram of a scenario of a demand response method for distributed load resources provided by an embodiment of the present application;

[0026] Figure 9 is a block diagram of the functional units of a demand response device for distributed load resources provided by an embodiment of the present application;

[0027] Figure 10 is a block diagram of the functional units of another demand response device for distributed load resources provided by an embodiment of the present application;

[0028] Figure 11 is a block diagram of the structure of a server provided by an embodiment of the present application. Detailed implementation manners

[0029] In order to enable those skilled in the art to better understand the solution of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without making creative efforts belong to the scope of protection of the present application.

[0030] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products or devices.

[0031] Referring to "embodiment" herein means that a specific feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0032] The "and / or" in the embodiments of the present application describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent the following three situations: A exists alone; A and B exist simultaneously; B exists alone. Here, A and B may be singular or plural.

[0033] In the embodiments of the present application, the symbol " / " may indicate that the front and rear associated objects have an "or" relationship. Additionally, the symbol " / " may also represent a division sign, that is, perform a division operation. For example, A / B may represent A divided by B.

[0034] The "at least one (piece)" or its similar expression in the embodiments of the present application refers to any combination of these items, including any combination of single item (piece) or plural items (pieces), referring to one or more, and multiple referring to two or more. For example, at least one (piece) of a, b, or c may represent the following seven situations: a, b, c, a and b, a and c, b and c, a, b, and c. Here, each of a, b, and c may be an element or a set containing one or more elements.

[0035] The "equal to" in the embodiments of the present application can be used in combination with greater than, applicable to the technical solutions adopted when greater than, or can also be used in combination with less than, applicable to the technical solutions adopted when less than. When equal to is used in combination with greater than, it is not used in combination with less than; when equal to is used in combination with less than, it is not used in combination with greater than.

[0036] To better understand the solutions of the embodiments of the present application, the terminal devices, related concepts, and backgrounds that may be involved in the embodiments of the present application will be introduced below.

[0037] (1) Virtual Power Plant: (Virtual Power Plant, VPP) An energy management system that uniformly coordinates and controls distributed power generation, demand-side response, and energy storage resources, interacts with the power grid through intelligent communication technology, and participates in power market transactions. The virtual power plant needs to obtain a large amount of user power consumption data from the load aggregator, including real-time data of power consumption load, power consumption habits, adjustable potential, etc., in order to perform overall power dispatching and optimization decisions, and achieve unified coordination and control of distributed power sources, energy storage, and load resources.

[0038] (2) Load Aggregator: An institution or platform managed by or cooperating with the virtual power plant, whose main function is to centrally manage and dispatch the power demands of distributed load resources to achieve flexible regulation and response to the power system.

[0039] (3) Demand response invitation: When the power grid needs to adjust electricity consumption due to tight or excessive supply, it issues a request instruction to entities such as virtual power plants to regulate the balance of power supply and demand, which is used to balance power supply and demand and ensure the stability of the power grid. When entities such as virtual power plants meet the corresponding response requirements, they can obtain corresponding economic subsidies or rewards.

[0040] In traditional power systems, large-scale thermal power units have long dominated. However, as clean energy gradually replaces traditional thermal power, the proportion of thermal power units in the power system is continuously decreasing. Although this transformation is of great significance for reducing carbon emissions, it also brings new challenges: the peak load regulation ability of the power system is weakened. Thermal power units have played a key role in balancing power supply and demand and coping with peak loads for a long time due to their stable output and flexible regulation characteristics. The decrease in their proportion makes the power system significantly insufficient in regulation ability when facing peak electricity consumption. Take fast-charging electric vehicle chargers as an example. Their electricity consumption has the characteristics of randomness and high power. Once such uncertain high-power large-load electricity-consuming facilities are used intensively, it is extremely easy to cause an imbalance in the real-time supply and demand of the system, seriously affecting the stable operation of the power grid.

[0041] To address these challenges, demand response through aggregating adjustable load resources by virtual power plants to participate in power grid dispatching has become a crucial technical means. A virtual power plant is not a power plant in the traditional sense. Instead, through advanced information technology and intelligent control technology, it integrates dispersed energy resources such as distributed power sources, energy storage systems, and adjustable loads to achieve unified and coordinated control. It can real-time monitor the operating status and load demand of the power grid, and flexibly adjust the output and electricity consumption of each resource according to the actual situation, effectively balancing power supply and demand, and enhancing the stability and reliability of the power grid. Virtual power plants play an irreplaceable role in peak shaving and valley filling, optimizing power resource allocation, and improving the flexibility of the power system, providing strong support for the stable operation of the new power system and the realization of environmental protection goals.

[0042] When the power grid issues a demand response invitation, the virtual power plant promises to the power grid the electricity quantity to be supplied in this response, that is, the declared quantity, so as to reasonably allocate the declared electricity quantity to distributed load resources, and finally obtain the actual response quantity feedback by the power grid. In the actual scenario, the actual response quantity of the virtual power plant needs to meet the requirements of the power grid to obtain remuneration. Specifically, in some regions (such as Shenzhen), the power grid requires that the actual response quantity be 80%-130% of the declared quantity. If the actual response quantity is less than 80%, it is regarded as a response failure and the remuneration is zero; if the actual response quantity exceeds 130%, the settlement is made at most according to 130%. Therefore, for the virtual power plant, it is crucial to predict in advance the probability of successful response of load resources to determine the accurate declared quantity, which directly affects whether enough load resources can be aggregated to complete at least 80% of the declared quantity, so as to ensure the success of the response. Therefore, the virtual power plant formulates a demand response declaration strategy considering the uncertainty of energy-consuming loads and different risk preferences.

[0043] Currently, when the virtual power plant participates in the demand response invitation of the power grid, it is necessary to determine the declared quantity that takes into account both the response ability of distributed load resources and the maximization of benefits for response. In the prior art, methods such as elastic price matrix modeling and linear efficiency hypothesis are used to determine the declared quantity, which have high requirements for data samples. Due to the limitation of data privacy protection, it is difficult to obtain comprehensive parameter information for accurate modeling and calculation, and the calculation cost and uncertainty of the obtained declared quantity are high, making it difficult for distributed load resources to successfully complete the declared quantity, resulting in a low response success rate and poor economic benefits. Therefore, how to improve the response success rate and economic benefits of the virtual power plant participating in the demand response invitation has become a technical problem that needs to be further solved.

[0044] To solve the above problems, the embodiments of the present application provide a demand response method and device for distributed load resources. The method determines the probability density function according to the current operation data and the posterior random forest model to dynamically characterize the response success rate of distributed load resources, so as to obtain the target declared quantity. It not only does not require precise prior assumptions and reduces the dependence on data samples, but also can quantify the uncertainty of the distributed load resources aggregated by the virtual power plant, significantly improving the prediction accuracy of the response ability of distributed load resources, reducing the calculation cost of the declared quantity, improving the accuracy of the declared quantity, and thus improving the response success rate and economic benefits of the virtual power plant participating in the demand response invitation.

[0045] Please refer to Figure 1 , Figure 1 which is a schematic structural diagram of a virtual power plant provided by the embodiments of the present application. As Figure 1 shown, the virtual power plant 100 includes a terminal device 110 and a server 120. The terminal device 110 is communicatively connected to the server 120, and the server 120 can be a single server or a server group composed of multiple servers.

[0046] In the daily use of the virtual power plant 100, the server 120 obtains multiple historical response data of multiple distributed load resources, and a single historical response data is used to characterize the historical response situation of the corresponding distributed load resource; a posteriori random forest model is determined according to the multiple historical response data, and the a posteriori random forest model is used to indicate the predicted response situation corresponding to the multiple distributed load resources; when receiving demand response invitation information from the power grid, multiple current operation data of the multiple distributed load resources are obtained, and a single current operation data is used to characterize the current operation situation of the corresponding distributed load resource; multiple probability density functions are determined according to the multiple current operation data and the a posteriori random forest model, and a single probability density function is used to indicate the response success rate of the corresponding distributed load resource; a target declaration quantity is determined according to the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declaration quantity.

[0047] Please refer to Figure 2 , Figure 2 which is a schematic structural diagram of a power system provided by an embodiment of the present application. As Figure 2 shown, the power system 200 includes the virtual power plant 100, multiple distributed load resources 210, and the power grid 220. The server 120 in the virtual power plant 100 is communicatively connected to the multiple distributed load resources 210, and the server 120 in the virtual power plant 100 is communicatively connected to the power grid 220.

[0048] Please refer to Figure 3 , Figure 3 which is a schematic structural diagram of a server in a virtual power plant provided by an embodiment of the present application. As Figure 3As shown in the figure, the server 120 includes a processor 310 and a memory 320, and the processor 310 is communicatively connected to the memory 320. Among them, one or more programs are stored in the memory 320, and the one or more programs are configured to be executed by the processor 310. The functions of the one or more programs are to obtain multiple historical response data of multiple distributed load resources, and a single historical response data is used to characterize the historical response situation of the corresponding distributed load resource; determine a posteriori random forest models according to the multiple historical response data, and the a posteriori random forest models are used to indicate the predicted response situations corresponding to the multiple distributed load resources; when receiving demand response invitation information from the power grid, obtain multiple current operation data of the multiple distributed load resources, and a single current operation data is used to characterize the current operation situation of the corresponding distributed load resource; determine multiple probability density functions according to the multiple current operation data and the a posteriori random forest models, and a single probability density function is used to indicate the response success rate of the corresponding distributed load resource; determine a target declaration volume according to the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declaration volume.

[0049] The following introduces a demand response method for distributed load resources provided by an embodiment of the present application.

[0050] Please refer to Figure 4 , Figure 4 which is a schematic flowchart of a demand response method for distributed load resources provided by an embodiment of the present application, and is applied to the server 120 in the virtual power plant 100 as shown in Figure 1 The virtual power plant 100 includes a terminal device 110 and a server 120. The terminal device 110 is communicatively connected to the server 120. The server 120 can be a single server or a server group composed of multiple servers. As shown in Figure 4 the figure, the method includes the following steps:

[0051] Step S410, obtain multiple historical response data of multiple distributed load resources.

[0052] Among them, a single historical response data is used to characterize the historical response situation of the corresponding distributed load resource.

[0053] Among them, the multiple historical response data can be directly collected by the server for all the historical response data of the distributed load resources, or can be collected by the load aggregator managed by the virtual power plant or cooperating with the virtual power plant for all the historical response data of the distributed load resources and then sent to the server of the virtual power plant.

[0054] Among them, the single historical response data includes the historical actual response amount, historical target declaration amount, historical actual power consumption, historical ambient temperature, historical ambient humidity, historical electricity price, and load nature corresponding to the distributed load resources for each historical demand response invitation. The load nature is label information used to indicate the electricity consumption type of the corresponding distributed load resources, and the electricity consumption types include industrial electricity, commercial electricity, and residential electricity.

[0055] Among them, the multiple historical response data corresponds one-to-one with the multiple distributed load resources.

[0056] Step S420, determine a posteriori random forest model according to the multiple historical response data.

[0057] Among them, the a posteriori random forest model is used to indicate the predicted response situation corresponding to the multiple distributed load resources.

[0058] Among them, the a posteriori random forest model includes multiple decision trees.

[0059] In a possible embodiment, the determining the a posteriori random forest model according to the multiple historical response data includes: determining training data according to the multiple historical response data, where the training data is time series data; obtaining a preset number of decision trees, number of feature selections, and number of feature dimensions. The number of decision trees is used to indicate the number of decision trees in the multiple decision trees, the number of feature selections is used to indicate the number of features to be selected for constructing a single decision tree, and the number of feature dimensions is used to indicate the number of features to be extracted from the training data; determining multiple features according to the number of feature dimensions and the training data; and constructing the multiple decision trees according to the number of feature selections and the multiple features.

[0060] Among them, the determining the training data according to the multiple historical response data can specifically be: performing data cleaning processing and data standardization processing on the multiple historical response data. The data cleaning processing is used to correct data and fill in missing data, and the data standardization processing is used to improve data quality and adapt to algorithm requirements.

[0061] Among them, the data cleaning processing includes at least one of the following steps: performing predictive filling on the multiple historical response data according to a preset data filling model to fill in missing data; deleting duplicate data in the multiple historical response data to remove redundant data; and deleting or correcting noise data and error data in the multiple historical response data to remove redundant data or correct error data.

[0062] Among them, the data standardization process includes at least one of the following steps: encoding categorical variables for the multiple historical response data to adapt to the algorithm requirements; balancing the quantity differences of data in different categories for the multiple historical response data; performing feature selection on the multiple historical response data to reduce the feature dimension.

[0063] Among them, the number of decision trees, the number of feature selections, and the number of feature dimensions can be set manually, or can be selected by using model selection methods such as grid search and random search in combination with cross-validation to select the ones that can make the trained posterior random forest model perform optimally.

[0064] Among them, determining multiple features according to the number of feature dimensions and the training data; constructing the multiple decision trees according to the number of feature selections and the multiple features can specifically be: extracting features from the training data according to the number of feature dimensions to obtain the multiple features; selecting features from the multiple features according to the number of feature selections to construct a tree structure to obtain the multiple decision trees. For example: when the number of decision trees is set to N, the number of feature selections is 12, and the number of feature dimensions is 24, the multiple features extracted from the training data include 24 features. When constructing each decision tree, 12 features are randomly selected from the 24 features, and the decision tree learning algorithm is used to recursively perform node splitting on the training data according to the randomly selected 12 features to obtain multiple feature points and splitting points, thereby constructing a structure, and finally obtaining a posterior random forest model containing N decision trees.

[0065] Among them, please refer to Figure 5 , Figure 5 is a schematic flowchart of another demand response method for distributed load resources provided by an embodiment of the present application. As Figure 5 shown, a demand response method for distributed load resources provided by an embodiment of the present application includes the following steps:

[0066] Step S421, determining training data according to the multiple historical response data.

[0067] Among them, the training data is time series data;

[0068] Step S422, obtaining a preset number of decision trees, a number of feature selections, and a number of feature dimensions.

[0069] Among them, the number of decision trees is used to indicate the number of decision trees in the multiple decision trees, the number of feature selections is used to indicate the number of features to be selected for constructing a single decision tree, and the number of feature dimensions is used to indicate the number of features to be extracted from the training data;

[0070] Step S423: Determine multiple features according to the number of feature dimensions and the training data.

[0071] Step S424: Construct the multiple decision trees according to the number of feature selections and the multiple features.

[0072] It can be seen that in this example, the posterior random forest model is trained according to the historical operation data and the preset number of decision trees, the number of feature selections, and the number of feature dimensions, and the target declaration quantity can be obtained in the case of only historical response data. Then, according to the current operation data and the posterior random forest model, the probability density function is determined to dynamically characterize the response success rate of the distributed load resources, so as to obtain the target declaration quantity. This not only does not require precise prior assumptions and reduces the dependence on data samples, but also can quantify the uncertainty of the distributed load resources aggregated by the virtual power plant, significantly improve the prediction accuracy of the response ability of the distributed load resources, reduce the calculation cost of the declaration quantity, improve the accuracy of the declaration quantity, and thus improve the response success rate and economic benefits of the virtual power plant participating in the demand response invitation.

[0073] Step S430: When receiving the demand response invitation information from the power grid, obtain the multiple current operation data of the multiple distributed load resources.

[0074] Among them, a single current operation data is used to characterize the current operation condition of the corresponding distributed load resource.

[0075] Among them, the demand response invitation information is used to indicate the current power consumption regulation demand of the power grid.

[0076] Step S440: Determine multiple probability density functions according to the multiple current operation data and the posterior random forest model.

[0077] Among them, a single probability density function is used to indicate the response success rate of the corresponding distributed load resource.

[0078] In a possible embodiment, the posterior random forest model includes a plurality of decision trees. Determining a plurality of probability density functions according to the plurality of current operation data and the posterior random forest model includes: determining a plurality of response data according to the plurality of current operation data, where a single response data is time series data, and the plurality of response data corresponds one-to-one with the plurality of distributed load resources; inputting each response data in the plurality of response data into the posterior random forest model to obtain a plurality of predicted observation subsets corresponding to each response data, where the plurality of predicted observation subsets corresponds one-to-one with the plurality of decision trees, and a single predicted observation subset is used to indicate the predicted response situation of the corresponding distributed load resource predicted by the corresponding decision tree; determining the probability density function corresponding to each response data according to the plurality of predicted observation subsets to obtain a plurality of probability density functions.

[0079] Among them, the plurality of current operation data can be directly collected by the server for all distributed load resources, or can be collected by a load aggregator managed by or cooperating with the virtual power plant for all distributed load resources and then sent to the server of the virtual power plant.

[0080] Among them, the single current operation data includes the current ambient temperature, current ambient humidity, current electricity price, and load nature of the corresponding distributed load resource.

[0081] Among them, specifically, determining a plurality of response data according to the plurality of current operation data can be: performing data cleaning processing and data standardization processing on the plurality of current operation data, where the data cleaning processing is used to correct data and fill in missing data, and the data standardization processing is used to improve data quality and adapt to algorithm requirements.

[0082] Among them, the data cleaning processing includes at least one of the following steps: predicting and filling the plurality of current operation data according to a preset data filling model to fill in missing data; deleting duplicate data in the plurality of current operation data to remove redundant data; deleting or correcting noise data and error data in the plurality of current operation data to remove redundant data or correct error data.

[0083] Among them, the data standardization processing includes at least one of the following steps: encoding categorical variables for the plurality of current operation data to adapt to algorithm requirements; for the plurality of current operation data, to balance the quantity difference of data in different categories; performing feature selection on the plurality of current operation data to reduce the feature dimension.

[0084] Among them, inputting each of the multiple response data into the posterior random forest model to obtain multiple predicted observation subsets corresponding to each response data. For example, it can be: when the number of the multiple decision trees is N, each response data is input into these N decision trees, and each decision tree outputs a predicted observation subset corresponding to the input response data, obtaining N predicted observation subsets corresponding to each response data.

[0085] Among them, the data in a single predicted observation subset includes: the predicted response amount of the corresponding distributed load resource, the predicted power consumption, the response period, the ambient temperature during the response period, the ambient humidity during the response period, the predicted electricity price, and the load nature.

[0086] It can be seen that in this example, multiple predicted observation subsets corresponding to each response data are obtained according to the current operation data and the multiple decision trees in the posterior random forest model. Then, the probability density function corresponding to each response data is determined according to the multiple predicted observation subsets to dynamically characterize the response success rate of each distributed load resource. The multiple predicted observation subsets can reflect the uncertainty of the distributed load resources, and the data information has a rich dimension, with higher prediction accuracy. It can obtain the target declaration quantity only with historical response data, not only without precise prior assumptions, reducing the dependence on data samples, but also being able to quantify the uncertainty of the distributed load resources aggregated by the virtual power plant, significantly improving the prediction accuracy of the response ability of the distributed load resources, reducing the calculation cost of the declaration quantity, improving the accuracy of the declaration quantity, and thus improving the response success rate and economic benefits of the virtual power plant participating in the demand response invitation.

[0087] In a possible embodiment, determining the probability density function corresponding to each response data according to the multiple predicted observation subsets to obtain multiple probability density functions includes: sampling the multiple predicted observation subsets to obtain a target sampling result; determining the probability density function corresponding to each response data according to the target sampling result to obtain the multiple probability density functions.

[0088] Among them, sampling the multiple predicted observation subsets to obtain a target sampling result includes: combining the multiple predicted observation subsets into a target observation set; performing Monte Carlo sampling on the target observation set to obtain the target sampling result. For example, it can be: N predicted observation subsets {y ∈ Y|T j (X i )} are combined into a target observation set Among them, X i represents the predicted observation subset corresponding to the distributed load resource i, and T j represents the jth decision tree, T j (X idenotes the predicted observation subset of distributed load resource i predicted by the j-th decision tree, y denotes the predicted response amount of distributed load resource i predicted by the j-th decision tree, and Y denotes the set of predicted response amounts of distributed load resource i predicted by all decision trees in the random forest.

[0089] Among them, the target sampling result includes a plurality of sampling samples. Specifically, Monte Carlo sampling can be: obtaining the number of samplings and the sampling rule; extracting a plurality of sampling samples from the target observation set according to the number of samplings and the sampling rule to obtain the target sampling result. A single sampling sample includes one or more data in the target observation set, and the number of the plurality of sampling samples is equal to the number of samplings.

[0090] Among them, the sampling rule includes any one or more of simple random sampling, stratified sampling, and Markov chain sampling. The simple random sampling is to randomly sample completely, and the probability of each data being sampled is equal. The stratified sampling is to divide the data in the target observation set into different dimensions according to the data type, and then perform simple random sampling independently from each dimension. The Markov chain sampling is to construct a Markov chain according to the target observation set so that the data in the target observation set is stably distributed into a preset target distribution, and samples are drawn by performing a random walk on the Markov chain.

[0091] Among them, please refer to Figure 6 , Figure 6 is a schematic flowchart of another demand response method for distributed load resources provided by an embodiment of the present application. As Figure 6 shown, a demand response method for distributed load resources provided by an embodiment of the present application includes the following steps:

[0092] Step S441, determining a plurality of response data according to the plurality of current operation data.

[0093] Among them, a single response data is time series data, and the plurality of response data corresponds to the plurality of distributed load resources one by one.

[0094] Step S442, inputting each response data in the plurality of response data into the posterior random forest model to obtain a plurality of predicted observation subsets corresponding to each response data.

[0095] Among them, the plurality of predicted observation subsets correspond to the plurality of decision trees one by one, and a single predicted observation subset is used to indicate the predicted response situation of the corresponding distributed load resource predicted by the corresponding decision tree.

[0096] Step S443, sampling the plurality of predicted observation subsets to obtain a target sampling result.

[0097] Step S444: Determine the probability density function corresponding to each response data according to the target sampling result, and obtain the multiple probability density functions.

[0098] It can be seen that in this example, multiple predicted observation subsets corresponding to each response data are obtained based on the current operation data and multiple decision trees in the posterior random forest model, and then sampling is performed on the multiple predicted observation subsets to further quantify the uncertainty of distributed load resources. Thus, according to the multiple predicted observation subsets, the probability density function corresponding to each response data is determined to dynamically characterize the response success rate of each distributed load resource. This not only eliminates the need for precise prior assumptions and reduces the dependence on data samples, but also can quantify the uncertainty of the distributed load resources aggregated by the virtual power plant, significantly improving the prediction accuracy of the response ability of distributed load resources, reducing the calculation cost of the declared quantity, improving the accuracy of the declared quantity, and thus enhancing the response success rate and economic benefits of the virtual power plant participating in the demand response invitation.

[0099] In a possible embodiment, the determining the probability density function corresponding to each response data according to the target sampling result and obtaining the multiple probability density functions includes: performing kernel density estimation on the target sampling result to obtain a target probability distribution, where the target probability distribution is used to characterize the posterior probability density distribution of the response success rate of the corresponding distributed load resource; obtaining multiple standard deviations of the multiple decision trees, where the multiple standard deviations correspond to the multiple distributed load resources one by one; and determining the probability density function corresponding to each response data according to the target probability distribution and the standard deviation corresponding to each response data, and obtaining the multiple probability density functions.

[0100] Among them, the performing kernel density estimation on the target sampling result to obtain a target probability distribution is specifically carried out through the following formula:

[0101]

[0102] Among them, Θ represents the kernel density estimation method, and f(X i ) represents the target probability distribution.

[0103] Among them, a single standard deviation is used to characterize the standard deviation of the predicted response quantity of the distributed load resource i predicted by the multiple decision trees.

[0104] Among them, the determining the probability density function corresponding to each response data according to the target probability distribution and the standard deviation corresponding to each response data and obtaining the multiple probability density functions is specifically carried out through the following formula:

[0105]

[0106] Among them, λ represents the standard deviation corresponding to the distributed load resource i, and E ra，i represents the predicted response amount of the distributed load resource i, and T j (X i ) represents the predicted response amount of the distributed load resource i predicted by the j-th decision tree, and f(E ra,i ) represents the probability density function corresponding to the load resource i.

[0107] Among them, please refer to Figure 7 , Figure 7 which is a schematic flowchart of another demand response method for distributed load resources provided by the embodiments of the present application. As Figure 7 shown, a demand response method for distributed load resources provided by the embodiments of the present application includes the following steps:

[0108] Step S441, determining a plurality of response data according to the plurality of current operation data.

[0109] Among them, a single response data is time series data, and the plurality of response data corresponds one-to-one to the plurality of distributed load resources.

[0110] Step S442, inputting each response data in the plurality of response data into the posterior random forest model to obtain a plurality of predicted observation subsets corresponding to each response data.

[0111] Among them, the plurality of predicted observation subsets corresponds one-to-one to the plurality of decision trees, and a single predicted observation subset is used to indicate the predicted response situation of the corresponding distributed load resource predicted by the corresponding decision tree.

[0112] Step S443, sampling the plurality of predicted observation subsets to obtain a target sampling result.

[0113] Step S4441, performing kernel density estimation on the target sampling result to obtain a target probability distribution.

[0114] Among them, the target probability distribution is used to characterize the posterior probability density distribution of the response success rate of the corresponding distributed load resource.

[0115] Step S4442, obtaining a plurality of standard deviations of the plurality of decision trees.

[0116] Among them, the plurality of standard deviations corresponds one-to-one to the plurality of distributed load resources;

[0117] Step S4443, determining the probability density function corresponding to each response data according to the target probability distribution and the standard deviation corresponding to each response data to obtain the plurality of probability density functions.

[0118] It can be seen that in this example, based on the target sampling results obtained by sampling, kernel density estimation is performed, and then a probability density function is constructed to quantify the uncertainty of distributed load resources, providing more sufficient decision-making information. Compared with the traditional method of predicting based on a random forest model, it realizes the leap from deterministic prediction to probabilistic decision-making, significantly improving the prediction accuracy of the response ability to distributed load resources, reducing the calculation cost of the declared quantity, improving the accuracy of the declared quantity, and thus improving the response success rate and economic benefits of the virtual power plant participating in the demand response invitation.

[0119] Step S450: Determine the target declared quantity according to the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declared quantity.

[0120] Among them, the demand response invitation information includes a response mode, a response start time, a response end time, a total response capacity, a clearing mode, and a clearing price. The response mode includes peak shaving or valley filling, and the clearing mode includes fixed-price clearing or competitive bidding clearing.

[0121] In a possible embodiment, the determining the target declared quantity according to the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declared quantity includes: obtaining a response success rate boundary, where the response success rate boundary includes an upper limit of the response success rate and a lower limit of the response success rate; determining multiple response failure probability functions corresponding to the multiple distributed load resources according to the lower limit of the response success rate and the multiple probability density functions; determining the clearing price according to the demand response invitation information; and determining the target declared quantity according to the clearing price and the multiple response failure probability functions.

[0122] Among them, the response success rate boundary can be an empirical value set according to the operation preferences of the virtual power plant or the load aggregator.

[0123] Among them, the determining the multiple response failure probability functions corresponding to the multiple distributed load resources according to the lower limit of the response success rate and the multiple probability density functions can specifically be performed through the following formula:

[0124]

[0125] Among them, P lb is the lower limit of the response success rate, and P f,i is the response failure probability function corresponding to the distributed load resource i.

[0126] Among them, please refer to Figure 8 , Figure 8It is a schematic diagram of a scenario of a demand response method for distributed load resources provided by an embodiment of the present application. As Figure 8 shown, for distributed load resource i, the target probability distribution f(X1) is obtained through the posterior random forest model M, and then the probability density function f(E ra,1 ) is obtained. Based on this, the response failure probability function P f,1 is obtained.

[0127] Among them, the determining the target declaration volume according to the clearing price, the upper limit of the response success rate, and the multiple response failure probability functions includes: obtaining a preset demand response cost function; determining a target revenue function according to the demand response cost function, the clearing price, and the multiple response failure probability functions; determining a target revenue expectation function according to the target revenue function; and determining the target declaration volume according to the target revenue expectation function and the multiple response failure probability functions.

[0128] Among them, the demand response cost function can be, for example:

[0129]

[0130] Among them, is the unit response cost paid by the virtual power plant to the distributed load resources participating in the response, and C ra,i is the demand response cost function.

[0131] Among them, the target revenue function can be, for example:

[0132]

[0133] Among them, R s is the target revenue obtained by responding to the demand response invitation information, M is the clearing price, and P ub is the upper limit of the response success rate.

[0134] Among them, the target revenue expectation function can be, for example:

[0135]

[0136] Among them, the determining the target declaration volume according to the target revenue expectation function and the multiple response failure probability functions can specifically be: obtaining a preset optimization problem and a maximum allowable value, where the maximum allowable value is used to represent the maximum demand response failure probability; solving the optimization problem according to the target revenue expectation function and the multiple response failure probability functions to obtain the target declaration volume.

[0137] Among them, the maximum allowable value can be an empirical value set according to the operation preferences of the virtual power plant or the load aggregator.

[0138] Among them, the optimization problem can be, for example:

[0139]

[0140] Among them, F is the highest allowable value.

[0141] It can be seen that in this example, multiple response failure probability functions are determined according to the lower limit of the response success rate and multiple probability density functions, and then the target declaration volume is determined according to the response failure probability function, the clearing price, and the upper limit of the response success rate, quantifying the uncertainty of the distributed load resources aggregated by the virtual power plant, significantly improving the prediction accuracy of the response ability to the distributed load resources, reducing the calculation cost of the declaration volume, improving the accuracy of the declaration volume, and thus improving the response success rate and economic benefits of the virtual power plant participating in the demand response invitation.

[0142] In a possible embodiment, after determining the target declaration volume according to the demand response invitation information and the multiple probability density functions so that the virtual power plant responds to the demand response invitation information with the target declaration volume, the method further includes: accepting the actual response volume from the power grid; and updating the multiple historical response data according to the target declaration volume and the actual response volume to iteratively update the posterior random forest model.

[0143] It can be seen that in this example, the response situation to the demand response invitation can be collected, and the historical response data of the distributed load resources can be updated in real time, so that the prediction accuracy can be dynamically improved with the number of responses.

[0144] It can be seen that in this application, multiple historical response data of multiple distributed load resources are obtained, and a single historical response data is used to characterize the historical response situation of the corresponding distributed load resource; a posteriori random forest model is determined according to the multiple historical response data, and the a posteriori random forest model is used to indicate the predicted response situation corresponding to the multiple distributed load resources; when receiving demand response invitation information from the power grid, multiple current operation data of the multiple distributed load resources are obtained, and a single current operation data is used to characterize the current operation situation of the corresponding distributed load resource; multiple probability density functions are determined according to the multiple current operation data and the a posteriori random forest model, and a single probability density function is used to indicate the response success rate of the corresponding distributed load resource; a target declaration quantity is determined according to the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declaration quantity. In this way, the probability density function is determined according to the current operation data and the a posteriori random forest model to dynamically characterize the response success rate of the distributed load resource, so as to obtain the target declaration quantity. This not only does not require precise prior assumptions and reduces the dependence on data samples, but also can quantify the uncertainty of the distributed load resources aggregated by the virtual power plant, significantly improve the prediction accuracy of the response ability of the distributed load resources, reduce the calculation cost of the declaration quantity, improve the accuracy of the declaration quantity, and thus improve the response success rate and economic benefits of the virtual power plant participating in the demand response invitation.

[0145] The above mainly introduces the solution of the embodiment of this application from the perspective of the execution process on the method side. It can be understood that in order to implement the above functions, the controller includes the corresponding hardware structure and / or software module for executing each function. Those skilled in the art should easily realize that, combining the units and algorithm steps of each example described in the embodiments provided in this article, this application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed in the way of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0146] Consistent with the above-described embodiments, please refer to Figure 9 , Figure 9 is the functional unit composition block diagram of a demand response device for distributed load resources provided by an embodiment of this application. As Figure 9As shown in the figure, the demand response device 900 for distributed load resources includes: a first receiving unit 901, configured to obtain multiple historical response data of multiple distributed load resources, where a single historical response data is used to characterize the historical response situation of the corresponding distributed load resource; a first processing unit 902, configured to determine a posteriori random forest model according to the multiple historical response data, where the posteriori random forest model is used to indicate the predicted response situation of the multiple distributed load resources; when receiving demand response invitation information from the power grid, obtain multiple current operation data of the multiple distributed load resources, where a single current operation data is used to characterize the current operation situation of the corresponding distributed load resource; determine multiple probability density functions according to the multiple current operation data and the posteriori random forest model, where a single probability density function is used to indicate the response success rate of the corresponding distributed load resource; and determine a target declaration quantity according to the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declaration quantity.

[0147] In a possible embodiment, in terms of determining the posteriori random forest model according to the multiple historical response data, the first processing unit 902 is specifically configured to: determine training data according to the multiple historical response data, where the training data is time series data; obtain a preset number of decision trees, a feature selection number, and a feature dimension number, where the number of decision trees is used to indicate the number of decision trees in the multiple decision trees, the feature selection number is used to indicate the number of features to be selected for constructing a single decision tree, and the feature dimension number is used to indicate the number of features to be extracted from the training data; determine multiple features according to the feature dimension number and the training data; and construct the multiple decision trees according to the feature selection number and the multiple features.

[0148] In a possible embodiment, the posteriori random forest model includes multiple decision trees. In terms of determining multiple probability density functions according to the multiple current operation data and the posteriori random forest model, the first processing unit 902 is specifically configured to: determine multiple response data according to the multiple current operation data, where a single response data is time series data, and the multiple response data corresponds to the multiple distributed load resources one by one; input each response data in the multiple response data into the posteriori random forest model to obtain multiple predicted observation subsets corresponding to each response data, where the multiple predicted observation subsets correspond to the multiple decision trees one by one, and a single predicted observation subset is used to indicate the predicted response situation of the corresponding distributed load resource predicted by the corresponding decision tree; and determine the probability density function corresponding to each response data according to the multiple predicted observation subsets to obtain multiple probability density functions.

[0149] In a possible embodiment, in terms of determining the probability density function corresponding to each response data according to the multiple predicted observation subsets to obtain multiple probability density functions, the first processing unit 902 is specifically configured to: sample the multiple predicted observation subsets to obtain a target sampling result; determine the probability density function corresponding to each response data according to the target sampling result to obtain the multiple probability density functions.

[0150] In a possible embodiment, in terms of determining the probability density function corresponding to each response data according to the target sampling result to obtain the multiple probability density functions, the first processing unit 902 is specifically configured to: perform kernel density estimation on the target sampling result to obtain a target probability distribution, where the target probability distribution is used to characterize the posterior probability density distribution of the response success rate of the corresponding distributed load resource; obtain multiple standard deviations of the multiple decision trees, where the multiple standard deviations correspond to the multiple distributed load resources one by one; determine the probability density function corresponding to each response data according to the target probability distribution and the standard deviation corresponding to each response data to obtain the multiple probability density functions.

[0151] In a possible embodiment, in terms of determining a target declaration quantity according to the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declaration quantity, the first processing unit 902 is specifically configured to: obtain a response success rate boundary, where the response success rate boundary includes a response success rate upper limit and a response success rate lower limit; determine multiple response failure probability functions corresponding to the multiple distributed load resources according to the response success rate lower limit and the multiple probability density functions; determine a clearing price according to the demand response invitation information; determine the target declaration quantity according to the clearing price and the multiple response failure probability functions.

[0152] In a possible embodiment, after determining a target declaration quantity according to the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declaration quantity, the demand response device 900 for distributed load resources is further configured to: accept an actual response quantity from the power grid; update the multiple historical response data according to the target declaration quantity and the actual response quantity to iteratively update the posterior random forest model.

[0153] It can be understood that since the method embodiment and the device embodiment are different presentation forms of the same technical concept, the content of the method embodiment part in this application should be synchronously adapted to the device embodiment part, and will not be elaborated here.

[0154] In the case of adopting an integrated unit, such as Figure 10As shown Figure 10 is a block diagram of the functional units of another demand response device for distributed load resources provided by an embodiment of the present application. In Figure 10 , the demand response device 900 for distributed load resources includes: a processing module 1012 and a communication module 1011. The processing module 1012 is used to control and manage the actions of a demand response device 900 for distributed load resources. For example, it executes the steps of the first receiving unit 901 and the first processing unit 902, and / or is used to execute other processes of the technologies described herein. The communication module 1011 is used to support the interaction between a demand response device 900 for distributed load resources and other devices. As Figure 10 shown, the demand response device 900 for distributed load resources may further include a storage module 1013, and the storage module 1013 is used to store the program code and data of the demand response device 900 for distributed load resources.

[0155] Among them, the processing module 1012 may be a processor or a controller. For example, it may be a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an ASIC, an FPGA or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It can implement or execute various exemplary logic blocks, modules and circuits described in combination with the disclosure of the present application. The processor may also be a combination that realizes computing functions, such as a combination including one or more microprocessors, a combination of a DSP and a microprocessor, and so on. The communication module 1011 may be a transceiver, an RF circuit or a communication interface, etc. The storage module 1013 may be a memory.

[0156] Among them, all relevant contents of each scenario involved in the above method embodiment can be cited in the function description of the corresponding functional module, and will not be repeated here. The above demand response device 900 for distributed load resources can all execute the above Figure 4 shown demand response method for distributed load resources.

[0157] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center in a wired or wireless manner. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more collections of available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0158] Figure 11 It is a block diagram of the structure of a server provided by an embodiment of the present application. As Figure 11 shown, the server 120 may include one or more of the following components: a processor 310, and a memory 320 coupled to the processor 310. The memory 320 may store one or more computer programs 321, and the one or more computer programs 321 may be configured to implement the methods described in the above embodiments when executed by the one or more processors 310.

[0159] The processor 310 may include one or more processing cores. The processor 310 connects various parts within the entire server 120 using various interfaces and lines. By running or executing instructions, programs, code sets, or instruction sets stored in the memory 320, and by invoking the data stored in the memory 320, it performs various functions of the server 120 and processes data. Optionally, the processor 310 may be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), or programmable logic array (PLA). The processor 310 may integrate a combination of one or several of a central processing unit (CPU), a graphics processing unit (GPU), and a modem, etc. Among them, the CPU mainly processes the operating system, user interface, application programs, etc.; the GPU is responsible for rendering and drawing display content; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 310 and may be implemented separately through a communication chip.

[0160] The memory 320 may include random access memory (RAM) and may also include read-only memory (ROM). The memory 320 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 320 may include a program storage area and a data storage area. Among them, the program storage area may store instructions for implementing the operating system, instructions for implementing at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc. The data storage area may also store data created during the use of the server 120.

[0161] It can be understood that the server 120 may include more or fewer structural elements than those shown in the above structural block diagram, which is not limited herein. The embodiments of the present application provide a computer-readable storage medium, on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, it implements the steps of the method described in any possible embodiment.

[0162] It should be understood that in various embodiments of the present application, the magnitudes of the sequence numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0163] In several embodiments provided by the present application, it should be understood that the disclosed methods, apparatuses, and systems can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for example, the division of the units is only a logical function division, and there may be other division methods in actual implementation; for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections between each other can be through some interfaces, and the indirect couplings or communication connections of the apparatuses or units can be in electrical, mechanical, or other forms.

[0164] The unit described as a separate component may or may not be physically separated, and the component displayed as a unit may or may not be a physical unit, that is, it may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0165] In addition, each functional unit in various embodiments of the present invention can be integrated in a processing unit, or each unit can be physically included separately, or two or more units can be integrated in one unit. The above integrated unit can be implemented in the form of hardware or in the form of a hardware plus software functional unit.

[0166] The integrated unit implemented in the form of software functional units can be stored in a computer-readable storage medium. The above-mentioned software functional units are stored in a storage medium and include several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute some steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, mobile hard disks, magnetic disks, optical discs, volatile memories, or non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of random access memory (RAM) are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchlink dynamic random access memory (SLDRAM), and direct rambus random access memory (DR RAM), etc., and various media that can store program codes.

[0167] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions without departing from the spirit and scope of the present invention, and can make various modifications and alterations, including combinations of the above different functions and implementation steps, including software and hardware implementation manners, all within the protection scope of the present invention.

Claims

1. A demand response method for distributed load resources, characterized in that: Applied to a server in a virtual power plant, the method comprises: Acquire multiple historical response data of multiple distributed load resources, where a single historical response data is used to characterize the historical response of the corresponding distributed load resource; Determine a posterior random forest model according to the plurality of historical response data, wherein the posterior random forest model is used to indicate predicted response conditions corresponding to the plurality of distributed load resources; When receiving a demand response invitation message from the power grid, obtaining a plurality of current operation data of the plurality of distributed load resources, wherein a single current operation data is used to characterize a current operation status of a corresponding distributed load resource; Determining a plurality of probability density functions according to the plurality of current operating data and the posterior random forest model, wherein a single probability density function is used to indicate a response success rate of a corresponding distributed load resource; A target reporting amount is determined according to the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target reporting amount.

2. The method according to claim 1, characterized in that The posterior random forest model includes a plurality of decision trees, and determining a plurality of probability density functions according to the plurality of current operating data and the posterior random forest model includes: Determine a plurality of response data according to the plurality of current operation data, wherein a single response data is time series data, and the plurality of response data correspond one-to-one to the plurality of distributed load resources; Input each of the multiple response data into the posterior random forest model to obtain multiple prediction observation subsets corresponding to each response data, wherein the multiple prediction observation subsets correspond to the multiple decision trees one by one, and a single prediction observation subset is used to indicate the prediction response of the corresponding distributed load resource predicted by the corresponding decision tree; The probability density function corresponding to each response data is determined according to the multiple predicted observation subsets to obtain multiple probability density functions.

3. The method according to claim 2, characterized in that The determining the probability density function corresponding to each response data according to the multiple prediction observation subsets to obtain multiple probability density functions includes: Sampling the multiple prediction observation subsets to obtain target sampling results; The probability density function corresponding to each response data is determined according to the target sampling result to obtain the multiple probability density functions.

4. The method according to claim 3, characterized in that Determining the probability density function corresponding to each response data according to the target sampling result to obtain the multiple probability density functions includes: Performing kernel density estimation on the target sampling result to obtain a target probability distribution, where the target probability distribution is used to characterize a posterior probability density distribution of a response success rate of a corresponding distributed load resource; Acquire multiple standard deviations of the multiple decision trees, where the multiple standard deviations correspond one-to-one to the multiple distributed load resources; The probability density function corresponding to each response data is determined according to the target probability distribution and the standard deviation corresponding to each response data to obtain the multiple probability density functions.

5. The method according to any one of claims 2 to 4, characterized in that: Determining the posterior random forest model according to the plurality of historical response data comprises: Determine training data according to the plurality of historical response data, wherein the training data is time series data; Obtaining a preset number of decision trees, a number of feature selections, and a number of feature dimensions, wherein the number of decision trees is used to indicate the number of decision trees in the multiple decision trees, the number of feature selections is used to indicate the number of features that need to be selected to construct a single decision tree, and the number of feature dimensions is used to indicate the number of features that need to be extracted from the training data; Determine a plurality of features according to the number of feature dimensions and the training data; The plurality of decision trees are constructed according to the feature selection quantity and the plurality of features.

6. The method according to claim 5, characterized in that The determining the target declared amount according to the demand response invitation information and the multiple probability density functions so that the virtual power plant responds to the demand response invitation information with the target declared amount includes: Obtaining a response success rate boundary, wherein the response success rate boundary includes a response success rate upper limit and a response success rate lower limit; Determine a plurality of response failure probability functions corresponding to the plurality of distributed load resources according to the response success rate lower limit and the plurality of probability density functions; Determining a clearing price according to the demand response invitation information; The target declared quantity is determined according to the clearing price, the upper limit of the response success rate and the multiple response failure probability functions.

7. The method according to claim 6, characterized in that After determining the target declared amount according to the demand response invitation information and the multiple probability density functions so that the virtual power plant responds to the demand response invitation information with the target declared amount, the method further includes: receiving an actual response amount from said power grid; The multiple historical response data are updated according to the target declared amount and the actual response amount to iteratively update the posterior random forest model.

8. A demand response device for distributed load resources, characterized in that: Applied to a server in a virtual power plant, the device comprises: A first receiving unit is used to obtain multiple historical response data of multiple distributed load resources, and a single historical response data is used to characterize the historical response of the corresponding distributed load resource; A first processing unit is used to determine a posterior random forest model based on the multiple historical response data, and the posterior random forest model is used to indicate the predicted response status corresponding to the multiple distributed load resources; when receiving a demand response invitation information from the power grid, a plurality of current operating data of the multiple distributed load resources are obtained, and a single current operating data is used to characterize the current operating status of the corresponding distributed load resource; a plurality of probability density functions are determined based on the multiple current operating data and the posterior random forest model, and a single probability density function is used to indicate the response success rate of the corresponding distributed load resource; a target declaration amount is determined based on the demand response invitation information and the multiple probability density functions, so that the virtual power plant responds to the demand response invitation information with the target declaration amount.

9. A server, characterized in that: The method comprises a processor, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the programs include instructions for executing the steps in the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: A computer program / instruction is stored thereon, and when the computer program / instruction is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images