Regulation and control method and platform based on resource partition balance aggregation, external characteristic representation and operation

Abstract

The invention relates to the technical field of electric energy allocation, and provides a regulation and control method and platform based on resource partition balance aggregation, external characteristic representation and operation. The aggregation method takes resource space distribution distance, aggregation cost, zoning plan electric quantity deviation and minimization of zoning loss caused by factors such as zoning balance between local consumption and cross-regional transaction electric quantity as targets, and determines aggregation nodes, zoning divided corresponding to the aggregation nodes and virtual units corresponding to the zoning. According to the external characteristic representation method, future output conditions of flexible resources are predicted based on a data-driven prediction mode so as to construct a resource feasible region and a polymer feasible region thereof considering perspectiveness. According to the operation regulation and control method, through two-stage combination of a maximum uncertain loss model and a minimum maximum expected loss model, resource output and climbing power conditions are considered to obtain an operation regulation and control decision of the minimum maximum uncertain loss. Partition aggregation optimization is achieved, external characteristics are described accurately, and regulation and control decision accuracy is improved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of electric energy allocation, and in particular to a method and platform based on resource partitioning, balancing aggregation, external characteristic representation, and operation regulation. Background Art

[0002] A virtual power plant is a medium that connects flexible resources that can provide electricity with the upper-level power grid. It participates in power grid operations by aggregating various types of flexible resources to form economies of scale. In order to improve operational management capabilities and market competitiveness, a virtual power plant needs to formulate an effective distributed flexible resource integration management strategy.

[0003] However, the current construction of virtual power plants still faces many problems that need to be solved. For example, on the one hand, the flexible resources called upon by virtual power plants have small capacity, large number, and complex relationships between the distribution network and flexible resource access points, which greatly increases the difficulty of aggregation. There is an urgent need for breakthroughs in large-scale aggregation technology. On the other hand, due to the rich variety of flexible resources and different technical characteristics, it is difficult to identify the external characteristic parameters of the aggregate after aggregation. The current aggregation technology lacks the accuracy of the characterization of the aggregate characteristics, which urgently needs breakthroughs. On the other hand, there are many uncertainties in the output of new energy. These factors lead to increased fluctuations in the operation of virtual power plants. The stable operation of virtual power plants faces great challenges. There is an urgent need for breakthroughs in virtual power plant operation and management technology to improve the stability of virtual power plants. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of the present disclosure is to provide a method and platform based on resource partitioning, balancing aggregation, external characteristic representation and operation regulation to solve the problems in the related art.

[0005] In a first aspect, the present disclosure provides a virtual power plant aggregation method based on resource partition balance, which is used to partition flexible resources connected to nodes in a distribution network. The method includes: determining an aggregation node in each node and a partition corresponding to each aggregation node based on minimization of a partition objective function. The partition objective function includes: a first term based on the total distance between global aggregation nodes, a second term based on the global aggregation cost, a third term based on the global partition planned power deviation, and a fourth term that does not meet the partition balance principle. The global aggregation node total distance is defined as the sum of the distances between each current aggregation node and each non-aggregate node between the adjacent aggregation node; the global aggregation cost is defined as the sum of the aggregation costs of the coordination controllers of each aggregation node; the global partition planned power deviation is defined as the sum of the partition deviations between the day-ahead planned power and the real-time power of each partition; the partition balance principle refers to the tendency for the sum of the local power consumption of each partition to increase and the sum of the cross-region transaction input power between the partitions to decrease; by aggregating the flexible resources in the partition, a virtual machine group corresponding to each partition is obtained.

[0006] In an embodiment of the first aspect, the second item includes a binary decision variable, where the binary decision variable is used to indicate whether the node is an aggregate node and is used as a coefficient for the cost associated with the aggregate node to exclude non-aggregate nodes from the calculation of the aggregate cost; and the first item includes an auxiliary variable, where the auxiliary variable is constructed based on the product of a binary decision variable of a current node and mutually exclusive variables of binary decision variables of other nodes and is used as a coefficient for the distance to exclude the distance between two non-aggregate nodes and the distance from an aggregate node to a node via another aggregate node from the sum of the distances.

[0007] In an embodiment of the first aspect, the partition objective function has a partition constraint set, and the partition constraint set includes: an upper limit on the total number of virtual machine groups; a maximum distance limit from an aggregation node to a boundary node of the partition to which it belongs; a first preset ratio of the amount of electricity consumed on-site within the partition being higher than the amount of power generated within the partition; the total amount of cross-regional transactions between each partition and other areas does not exceed a transaction capacity upper limit, and the transaction capacity upper limit is determined based on a second preset ratio of the total power generation of each of the partitions.

[0008] A second aspect of the present disclosure provides a method for representing external characteristics of resources, including: obtaining historical operating data of each flexible resource; preprocessing the historical operating data to obtain input data; using a data-driven output time series prediction model corresponding to each flexible resource arrangement to predict the predicted output sequence of each flexible resource at each future moment based on the input data; constructing a feasible domain for each flexible resource based on at least one operating characteristic constraint determined by the predicted output sequence; obtaining a virtual machine group feasible domain for each virtual machine group by aggregating the feasible domains of flexible resources belonging to the same virtual machine group; wherein the virtual machine group is determined based on the aggregation method based on resource partition balancing as described in any one of the first aspects; the aggregation method of the feasible domain includes calculating the Minkowski sum of the feasible domain.

[0009] In an embodiment of the second aspect, the feasible domain of each flexible resource is constructed based on at least one set of operating characteristic constraints determined according to the predicted output sequence, including: determining at least one operating characteristic constraint corresponding to each flexible resource based on the operating characteristics of the flexible resource; the operating characteristic constraints include: output constraints, climbing constraints and energy state constraints; the at least one operating characteristic constraint is applicable to all moments of the flexible resource or each periodic moment; and a unified form of feasible domain is obtained by combining the at least one operating characteristic constraint and the elements of the predicted output sequence.

[0010] In an embodiment of the second aspect, the feasible domains of flexible resources belonging to the same virtual machine group are aggregating to obtain the feasible domain of each virtual machine group, including: representing the feasible domains of flexible resources belonging to the same virtual machine group in the form of primitive convex polyhedrons by finding the intersection of half planes; selecting a homogeneous polyhedron based on each of the primitive convex polyhedrons; and obtaining the maximum inner approximation feasible domain of each flexible resource based on the same homogeneous polyhedron by different scaling and translation of the homogeneous polyhedrons for the original convex polyhedron of each flexible resource, for aggregation of the virtual machine group.

[0011] In an embodiment of the second aspect, the maximum inner approximation of the original convex polyhedron of each flexible resource by the isomorphic polyhedron is performed by different scaling and translation respectively, and the maximum inner approximation feasible domain of each flexible resource based on the same isomorphic polyhedron is obtained for the aggregation of the virtual machine group, including: defining the feasible domain in the form of an approximate polyhedron as the feasible domain in the form of an isomorphic polyhedron multiplied by the scaling factor and then added with the translation factor; solving each group of maximum inner approximation scaling coefficients and maximum inner approximation translation coefficients based on the maximum inner approximation scaling coefficients and maximum inner approximation translation coefficients of each flexible resource by the approximate polyhedron. and the feasible domain in the form of an isomorphic polyhedron to obtain the maximum inner approximation feasible domain of each flexible resource based on the same isomorphic polyhedron; the said obtaining the virtual machine group feasible domain of each virtual machine group by aggregating the feasible domains of the flexible resources belonging to the same virtual machine group also includes: aggregating the maximum inner approximation feasible domain of the flexible resources belonging to the same virtual machine group to obtain the virtual machine group feasible domain obtained by applying a first aggregation scaling coefficient and a first aggregation translation coefficient based on the isomorphic polyhedron; wherein the scaling coefficient and the translation coefficient of the virtual machine group feasible domain are respectively the sum of the scaling coefficients and the sum of the translation coefficients of the maximum inner approximation feasible domains of each flexible resource belonging to it.

[0012] In an embodiment of the second aspect, the method for representing external resource characteristics further includes: aggregating feasible domains of each of the virtual machine groups to obtain a feasible domain of a virtual power plant.

[0013] The third aspect of the present disclosure provides an operation control method, including: obtaining a maximum uncertain loss model, which is used to obtain the maximum uncertain loss with the goal of maximizing the uncertain loss between the maximum benefit that can be achieved under the operation control decision of the flexible resource output uncertainty scenario of the uncontrollable flexible resource and the current benefit with the minimum output deviation penalty cost that can be achieved by the current operation control decision, under the constraints of the first operation constraint set of the power supply unit in the virtual power plant and when the ramping power scenario is determined; the power supply unit is implemented as a virtual power plant or a virtual machine group; the operation control decision includes an action set for the controllable flexible resources in the power supply unit; the first operation constraint set includes external characteristic constraints for the power supply unit, and the external characteristic constraints are formed based on the feasible domain of the virtual power plant or virtual machine group obtained by the resource external characteristic representation method as described in any one of the second aspects; the maximum uncertain loss of each group of alternative operation control decisions in a set of the ramping power scenarios is obtained by the maximum uncertain loss model; each expected loss is obtained based on the maximum uncertain loss of each group by minimizing the maximum expected loss model, and the corresponding alternative operation control decision is determined as the target operation control decision according to the minimum value of each expected loss.

[0014] In an embodiment of the third aspect, the controllable flexible resources include one or more of energy storage equipment, gas / oil power generation equipment, transferable loads and curtailable loads; the action set includes one or more of gas / oil power generation equipment action, charging, discharging, power transfer and power curtailment; and / or, the uncontrollable flexible resources include new energy power generation equipment.

[0015] In an embodiment of the third aspect, the first operating constraint set includes a power balance constraint, which means that the sum of the actual output power of the power supply unit and the output deviation is equal to the sum of the day-ahead power and the climbing power; the output deviation refers to the deviation of the actual output power of the power supply unit compared to the predetermined output power, and the output deviation penalty cost is related to the output deviation; the operating control decision includes an action set regarding controllable flexible resources; the maximum benefit is obtained based on searching for the optimal operating control decision and the output deviation of the virtual machine group under the operating control decision in the output uncertainty scenario with the goal of maximizing the first benefit; the current benefit is obtained based on searching for the output deviation under the current operating control decision to minimize the power supply deviation cost with the goal of maximizing the second benefit; the first benefit and the second benefit are represented by the difference between the sum of the day-ahead power benefit and the climbing power benefit and the sum of the total flexible resource scheduling cost and the output deviation penalty cost.

[0016] In an embodiment of the third aspect, the output deviation penalty cost is related to the output deviation; the output deviation refers to the deviation of the actual output power of the virtual machine group compared to the predetermined output power; the calculation method of the output deviation includes: when the output deviation is greater than zero, the output deviation penalty cost is a positive value positively correlated with the output deviation; when the output deviation is not greater than zero, the output deviation penalty cost is a negative value negatively correlated with the output deviation.

[0017] In an embodiment of the third aspect, the maximum uncertainty loss model is expressed as:

[0018] ;

[0019] in, Indicates the current operation and control decision and the maximum uncertain loss under a determined ramp capacity scenario; Indicates various output uncertainty scenarios; current operation and control decisions , p ope represents the set of actions on controllable flexible resources; P m and P m,u They represent the day-ahead power under the current operation and control decision and the operation and control decision under the output uncertainty scenario respectively; Indicates that the power supply unit is The total cost of flexible resource scheduling to execute operation and control decisions in the scenario, Represents the total cost of flexible resource scheduling for the power supply unit to execute the current operation control decision; and C B They represent the output deviation penalty costs to be borne under the operation and control decision of the output uncertainty scenario and the current operation and control decision respectively; R m and They represent the day-ahead power benefits under the current operation and control decision and the operation and control decision under the output uncertainty scenario respectively; R R Indicates the climbing power gain; R m and The calculation method is as follows: , is the day-ahead power value coefficient; R R The calculation method is as follows: , is the climbing power value coefficient; and C B The calculation method is as follows:

[0020] ;

[0021] in, and is the deviation power penalty coefficient.

[0022] In an embodiment of the third aspect, the first operating constraint set includes: external characteristic constraints of the power supply unit, output power characteristic constraints of various controllable flexible resources inside the power supply unit, and at least one of adjustable power constraints; wherein the external characteristic constraints are used to describe the characteristics of the power supply unit; the output power characteristic constraints include an action constraint set for the action set.

[0023] In an embodiment of the third aspect, the minimization of maximum expected loss model is expressed as:

[0024] ;

[0025] Among them, p VPP Indicates the current operation and control decision. represents the climbing power under the Kth climbing power scenario, represents the probability of the Kth climbing power scenario.

[0026] In an embodiment of the third aspect, each of the expected losses is obtained by calculating the weighted sum of a corresponding set of losses with the probability of occurrence of each of the climbing power scenarios as a weight value; the minimization of the maximum expected loss model includes a second operating constraint set, and the second operating constraint set includes: the first operating constraint set, the adjustable power constraint under each climbing power scenario, and the sum of each weight value is 1.

[0027] The fourth aspect of the present disclosure provides an operation and control platform, including: a processor and a memory; the memory stores program instructions; the processor is used to run the program instructions to execute at least one of the following: an aggregation method based on resource partition balancing as described in any one of the first aspects; a resource external characteristic representation method as described in any one of the second aspects; an operation and control method as described in any one of the third aspects.

[0028] As described above, the present disclosure relates to the technical field of electric energy allocation, and provides a method and platform based on resource partition balance aggregation, external characteristic representation and operation control. The aggregation method takes into account the resource spatial distribution distance, aggregation cost, partition plan power deviation, and the partition balance between local consumption and cross-regional transaction power, etc., and aims to minimize the partition loss caused by factors such as the partition, and determines the aggregation node and the partitions corresponding to the aggregation node and their corresponding virtual machine groups. The external characteristic representation method predicts the future output of flexible resources based on a data-driven prediction method to construct a resource feasible domain and its aggregate feasible domain based on foresight. The operation control method considers the resource output and ramp power conditions through a two-stage combination of the maximum uncertain loss model and the minimization of the maximum expected loss model to obtain an operation control decision that minimizes the maximum uncertain loss. Partition aggregation optimization is achieved respectively, and the external characteristic characterization accurately supports more decentralized user-side heterogeneous flexible resource control and improves the accuracy of control decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A schematic diagram showing an application scenario of a virtual power plant in one embodiment of the present disclosure.

[0030] Figure 2 A schematic diagram showing a flow chart of an aggregation method based on resource partition balancing in an embodiment of the present disclosure.

[0031] Figure 3 A flowchart illustrating a method for representing external resource characteristics in an embodiment of the present disclosure is shown.

[0032] Figure 4 A schematic diagram showing the structure of the LSTM network.

[0033] Figure 5 A schematic diagram showing a specific flow chart of step S304 in one embodiment of the present disclosure is shown.

[0034] Figure 6 Schematic diagram showing the geometric principles of Minkowski and computation.

[0035] Figure 7 A schematic diagram showing a specific flow chart of step S305 in one embodiment of the present disclosure is shown.

[0036] Figure 8 A schematic diagram showing specific implementation steps of step S703 in one embodiment of the present disclosure is shown.

[0037] Figure 9 A flow chart showing an operation control method in an embodiment of the present disclosure is shown.

[0038] Figure 10 A schematic diagram showing the principle of the combined application of Examples 1 to 3 of the present disclosure.

[0039] Figure 11 A schematic diagram showing the effect of using the aggregation method based on resource partition balance in an embodiment of the present disclosure to partition flexible resources in a network.

[0040] Figure 12 A schematic diagram showing the structure of a computer device in one embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0041] The following describes the embodiments of the present disclosure through specific examples. Those skilled in the art can easily understand the other advantages and effects of the present disclosure from the information disclosed in this disclosure. The present disclosure can also be implemented or applied through different specific embodiments. The details of the present disclosure can also be modified or changed according to different viewpoints and application modules without departing from the spirit of the present disclosure. It should be noted that the embodiments and features in the embodiments of the present disclosure can be combined with each other unless there is a conflict.

[0042] The following is a detailed description of the embodiments of the present disclosure with reference to the accompanying drawings so that those skilled in the art can easily implement the present disclosure. The present disclosure can be embodied in many different forms and is not limited to the embodiments described herein.

[0043] In the present disclosure, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or at least one embodiment or example in a suitable manner. In addition, those skilled in the art may combine and integrate different embodiments or examples described in the present disclosure, as well as features of different embodiments or examples, unless otherwise contradictory.

[0044] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of such features. In the context of this disclosure, "at least one" means two or more, unless otherwise specifically defined.

[0045] In order to clearly describe the present disclosure, components not related to the description are omitted, and the same or similar components throughout the specification are denoted by the same reference numerals.

[0046] Throughout this specification, when a device is said to be "connected" to another device, this includes not only "direct connection" but also "indirect connection" with other elements interposed therebetween. Furthermore, when a device is said to "include" a certain component, unless otherwise stated, this does not exclude the inclusion of other components but rather implies that the device may include other components.

[0047] Although the terms first, second, etc. are used in this document to represent various elements in some examples, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, the first interface and the second interface, etc. are represented. Furthermore, as used in this document, the singular forms "one," "an," and "the" are intended to also include the plural forms, unless there is a contrary indication in the context. It should be further understood that the terms "comprise" and "include" indicate the presence of the described features, steps, operations, elements, modules, projects, types, and / or groups, but do not exclude the presence, occurrence, or addition of one or at least one other feature, step, operation, element, module, project, type, and / or group. The terms "or" and "and / or" used herein are interpreted as inclusive, or mean any one or any combination. Therefore, "A, B, or C" or "A, B, and / or C" means "any of the following: A; B; C; A and B; A and C; B and C; A, B, and C." Exceptions to this definition only occur when the combination of elements, functions, steps, or operations is inherently mutually exclusive in some way.

[0048] The technical terms used herein are intended only to refer to specific embodiments and are not intended to limit the present disclosure. The singular form used herein also includes the plural form unless the statement explicitly indicates otherwise. The term "comprising" as used in this specification is intended to specify specific features, regions, integers, steps, operations, elements, and / or components and does not exclude the presence or addition of other features, regions, integers, steps, operations, elements, and / or components.

[0049] Although not defined differently, all terms used herein, including technical and scientific terms, have the same meanings as those generally understood by those skilled in the art to which this disclosure belongs. Terms defined in commonly used dictionaries are additionally interpreted as having meanings consistent with relevant technical literature and the current message. Unless otherwise defined, they should not be overly interpreted as ideal or highly formalized meanings.

[0050] Virtual power plants can integrate and regulate distributed flexible energy resources such as sources, networks, loads, and storage, and become an equivalent controllable power source to the outside world. It has strong flexibility and adjustment capabilities. It can be used as a power generation source to supply power to the power system, or as power for the load absorption system, and has the ability to adjust the load upward and downward.

[0051] However, precisely because virtual power plants (VPPs) are essentially a technology for intelligently deploying and aggregating existing flexible resources, they face challenges arising from the specific characteristics of these resources. For example, the energy provided by renewable energy flexible resources, such as photovoltaics and wind farms, is characterized by significant randomness and uncertainty. This, combined with other uncertainties (such as energy market uncertainty), can lead to increased operational volatility in VPPs, posing significant challenges to their stable operation. Improving VPP stability is crucial. Furthermore, given the diverse types of flexible resources—sources, grids, loads, and storage—the identification of the external characteristic parameters of aggregated resources is complex, and the accuracy of characterizing these aggregates remains limited. Furthermore, the flexible resources deployed by VPPs often have small capacities, large numbers, and complex relationships between distribution networks and flexible resource access points. This significantly increases the difficulty of aggregation, necessitating breakthroughs in scalable aggregation technologies.

[0052] In view of this, the embodiments of the present disclosure provide a method and platform based on resource partition balance aggregation, external characteristic representation and operation regulation to solve the problems in related technologies.

[0053] like Figure 1 As shown, a schematic diagram of a virtual power plant application scenario in one embodiment of the present disclosure is shown.

[0054] exist Figure 1 The diagram shows the connections between various flexible resources in physical space, forming energy flows. In some embodiments, these flexible resources may include gas / oil power generation equipment, renewable energy power generation equipment, energy storage, transferable loads, and curtailable loads. Specifically, gas / oil power generation equipment may include gas turbines. Renewable energy power generation equipment may include photovoltaic power plants and wind farms. Energy storage may include batteries. Transferable loads and renewable energy resources correspond to different types of flexible resources, and the energy flow connection carriers may be implemented differently.

[0055] Specifically, flexible resources enter the distribution network 101 by connecting to nodes in the distribution network 101. The operating information of the flexible resources can be collected (such as through sensors set up corresponding to the flexible resources) and the information flow is transmitted to the operation and control platform 102 through a communication connection. The operation and control platform 102 can analyze the information flow and allocate and control the flexible resources through the distribution network 101 to achieve the aggregation of scattered flexible resources into, for example, "virtual machine groups" to be called like real physical units, thereby achieving the function of a "virtual power plant".

[0056] In some embodiments, the operation and control platform 102 can analyze the resource information of the flexible resources and implement means for regulating the production capacity / storage of the flexible resources. In some embodiments, the operation and control platform 102 can include terminals with processing capabilities, such as server groups or servers, which can be one or more. In some embodiments, the operation and control platform 102 can be implemented as a power grid control center or as a virtual power plant management platform integrated with or communicating with the power grid control center. The virtual power plant management platform can be used to analyze the information flow and issue dispatch control instructions to form flexible resource aggregation control, realizing the functionality of a "virtual power plant." In some embodiments, the implementation of the virtual power plant management platform is related to the operating mode of the virtual power plant. For example, in an operating scenario with separate transmission and distribution, the virtual power plant management platform can include, for example, a distribution system operator (DSO) and a transmission system operator (TSO). In some embodiments, the virtual power plant management platform can also include only the distribution system operator (DSO), depending on the number of operators corresponding to the operating mode.

[0057] The "power supply unit" in the virtual power plant referred to herein is defined herein as an object to be regulated. As an example, the "power supply unit" can be implemented as a flexible resource aggregate for the entire or local virtual power plant. That is, the operation and control platform 102 can generate an operation and control decision for regulating the aggregate of all flexible resources in the virtual power plant, i.e., a global operation and control decision for the virtual power plant; it can also generate an operation and control decision for optimizing the day-ahead operation of an aggregate of local flexible resources, i.e., a local operation and control decision for the virtual power plant. This allows for more flexible operation and control optimization of different scopes of the virtual power plant. In some embodiments, the aggregate includes at least one of the following: a virtual power plant, a virtual machine group, a flexible resource cluster corresponding to a flexible resource type, a flexible resource cluster within a geographical area, etc. As an example, a flexible resource cluster corresponding to a flexible resource type, such as an energy storage device cluster, etc. Of course, the above local flexible resource cluster is only an example and can be changed according to actual needs and is not a limitation.

[0058] Specifically, the operation control platform 102 obtains relevant information about each flexible resource, including but not limited to historical and current operational data of the flexible resource, such as the day-ahead planned output, actual output, day-ahead planned electricity, actual electricity, historical operating power, state of charge, etc. Resource-related information may also be included, including but not limited to meteorological parameters, electricity price data, resource aggregation costs, and the electrical distance of the resource in the distribution network. The resource aggregation costs include but are not limited to the installation and operation costs of aggregation nodes and the setup and operation costs of communication infrastructure.

[0059] Example 1:

[0060] Example 1 is about the specific implementation of the aggregation method based on resource partition balance. The aggregation method based on resource partition balance is used to determine the partitions of resources and further aggregate the resources in the determined partitions into virtual machine groups. This can effectively support the efficient regulation of decentralized heterogeneous and flexible resources.

[0061] like Figure 2 FIG. 1 is a flow chart showing a method for aggregation based on resource partition balance in an embodiment of the present disclosure. The method for aggregation based on resource partition balance can be applied to the operation control platform 102 described in the previous embodiment.

[0062] In a specific scenario, the distribution network includes multiple nodes, and the nodes can be the locations where various electrical equipment are connected to the distribution network. These devices include transformers, switchgear, loads, etc., which are connected to other parts of the distribution network through nodes. The node can also be the place where multiple feeders, branch lines or electrical equipment in the distribution network are interconnected. At these connection points, the current can be diverted or merged. A coordination controller can be configured at the node to coordinate and control the flexible resources within the partition where it is located, so as to realize the control of the resource aggregate corresponding to the partition, that is, the virtual machine group. Each coordination controller can be communicated with the central aggregation controller to execute the regulation of the flexible resources within the partition in accordance with the instructions issued by the central aggregation controller. In some embodiments, the coordination controller can be arranged only at the node, and the node where the coordination controller is arranged in each node serves as the "aggregation node" in the "partition" where it is located.

[0063] The aggregation method based on resource partition balancing of the virtual power plant includes:

[0064] Step S201: Based on minimization of the partition objective function as the goal, determine the aggregation nodes in each node and the partitions divided corresponding to each aggregation node.

[0065] Resource partitioning requires considering multiple optimization objectives (such as maximizing benefits or minimizing losses). For example, how should the aggregation nodes in each partition be set to minimize the total distance between the aggregation node and other non-aggregated nodes, thereby maximizing current transmission speed and reducing losses? Another example is how to set the aggregation nodes in each partition to minimize aggregation costs. Aggregation costs may include, but are not limited to, the installation and operation costs of the coordination controller and the cost of communication facilities. Based on these multiple optimization objectives, a partitioning objective function can be constructed. By minimizing its value, the partitioning result, i.e., the aggregation nodes and non-aggregated nodes belonging to each partition, can be determined.

[0066] Thus, in some embodiments, the partition objective function may include: a first item based on the total distance between global aggregation nodes, a second item based on the global aggregation cost, a third item based on the global partition planned electricity deviation, and a fourth item that does not meet the partition balance principle. Among them, the global aggregation node total distance can be defined as the sum of the distances between each current aggregation node and each non-aggregate node between the adjacent aggregation node. The global aggregation cost can be defined as the sum of the aggregation costs of the coordination controllers of each aggregation node. The global partition planned electricity deviation can be defined as the sum of the partition deviations between the day-ahead planned electricity and the real-time electricity of each partition. The partition balance principle refers to the tendency of the sum of the on-site electricity consumption of each partition to increase, and the sum of the cross-regional transaction input electricity between partitions to decrease.

[0067] It should be noted that in the fourth item, by introducing the zone balance principle, the local consumption of new energy within the zone can be maximized, thereby reducing the load impact on the superior dispatching system, utilizing the differences in the flexibility resource endowments of each zone, achieving complementarity through cross-regional dispatching, and improving the overall system stability and economy.

[0068] The above principle is explained in detail. First, the implementation of the partition objective function is briefly explained through the following formula (1):

[0069] (1);

[0070] in, l ij is the distance from the jth node to the ith aggregation node; b i is a binary decision variable describing whether the i-th node has an aggregation node; N n is the total number of nodes in a given network; c op,i and c ins,i are the operation and installation costs of the coordination controller associated with the aggregation node on node i; N z is the total number of partitions; represents the planned electricity consumption of zone i in the day-ahead market (t=d represents day-ahead, t=r represents real-time); represents the planned electricity consumption of partition i in the real-time market; The local power consumption in zone k; is the inter-regional power input to partition k; β is the inter-regional transaction penalty factor. It can be understood that the terms added to the right side of the equal sign in the formula are the first, second, third, and fourth terms mentioned above, respectively.

[0071] In addition, B ij As auxiliary variables, use binary decision variables b i Expressed as:

[0072] ;

[0073] It can be understood that in the above formula, the binary decision variable b i It is used as the coefficient of the aggregation cost when the node is an aggregation node, so as to exclude non-aggregation nodes from the calculation of the aggregation cost. ij The coefficient of the distance is constructed based on the product of the binary decision variable of the current node and the mutually exclusive variables of the binary decision variables of other nodes. Specifically, when the binary decision variable of node i is 0 or the binary decision variable b of any other node is jWhen it is 0, the corresponding distance will be excluded. Among them, the binary decision variable of node i is 0, which corresponds to node i not being an aggregation node, or the binary decision variable of any other node is 1. Let 1 minus the binary decision variable be 0, which corresponds to the other node being an aggregation node. In other words, if any node on the path from the i-th aggregation node to the j-th node is an aggregation node, then B ij is zero. So the auxiliary variable B ij It is not used to consider the spacing of simple aggregation nodes, but combined with l ij Calculate the shortest distance from the i-th aggregation node to the j-th aggregation node.

[0074] Through the auxiliary variable B ij The distance between two non-aggregated nodes and the distance from an aggregator node to other nodes on the path across an aggregator node can be excluded (that is, other aggregator nodes are excluded from the range of the partition where the aggregator node is located). In this way, the distance between two non-aggregated nodes and the distance from an aggregator node to a node via another aggregator node can be excluded from the sum of the distances.

[0075] Since the first item excludes the distances of some non-aggregate nodes that are not expected to be included in the accumulation of the distances of each aggregated node through binary decision variables and auxiliary variables, it mainly performs calculations from the perspective of the distances related to the aggregated nodes. This can effectively reduce the amount of calculation compared to the solution that considers the distances between all nodes.

[0076] Because l ij It depends on the spatial distribution of nodes and consumers in the distribution network, and since the aggregation node is the only location where the aggregation coordination controller is set, it is assumed that l ij is a constant.

[0077] For n nodes, l ij It can be expressed in matrix form as follows:

[0078] ;

[0079] The above formula (1) is a conceptual expression. In actual calculation, the values ​​of the four terms of the function need to be normalized to the same interval, thereby obtaining the normalized partition objective function, which is expressed as:

[0080] (2);

[0081] in, 、 、 and Corresponding to the four terms of the original objective function, w1, w2, w3 and w4 are the weights assigned to the four terms, which are related to the preference or requirements of the virtual power plant for overall aggregation. max is the maximum possible distance for a given system, i.e., the total distance from the central aggregator to each distributed resource without aggregation; or, D max Alternatively, the maximum value of the total distance between each aggregation node and other nodes can be calculated. Of course, the amount of calculation will be greater than the former, as the former only needs to be calculated once based on the central aggregator. max is the maximum number of aggregation nodes that may exist in a given network. Since each node may appear as an aggregation node, N max Can be equal to the total number of nodes in a given network. E max is the maximum total power planned for the entire network on the day before; T max It is the sum of the upper limits of cross-region transaction capacity.

[0082] In addition, considering that the virtual power plant has restrictions on the total number of internal virtual machine groups and the maximum possible distance of each virtual machine group, the partition objective function has a partition constraint set.

[0083] The partition constraint set includes:

[0084] 1) Upper limit on the total number of virtual machine groups;

[0085] ;

[0086] 2) Maximum distance limit from an aggregation node to other non-aggregation nodes:

[0087] .

[0088] in, is the maximum possible distance between aggregation node i and node j in the partition, that is, the maximum possible distance between node i and the flexibility resource connected to node j. It describes the geographical and communication infrastructure limitations and limits the number of virtual machine groups.

[0089] Furthermore, considering the limitations of the zoning balance principle, zoning aggregates are allowed to share flexibility resources through market transactions or coordinated scheduling. However, the proportion of local power consumption within the zoning must not be lower than the threshold value, and the cross-zoning transaction volume between the zoning and other regions must not exceed the capacity limit. Therefore, the zoning constraint set includes the following constraints:

[0090] 3) The local power consumption within the zone is higher than the first preset ratio of the power generation within the zone:

[0091] ;

[0092] in, The total power generation of distributed energy in zone k; γ is the first preset ratio, i.e., the threshold of local consumption rate, which ensures the self-balancing ability of the region;

[0093] 4) The total amount of cross-regional transactions between each zone and other areas does not exceed the upper limit of transaction capacity. The upper limit of transaction capacity is determined based on a second preset ratio of the total power generation of each zone, expressed as:

[0094]

[0095] Among them, δ is the second preset ratio, which represents the upper limit rate of cross-regional transaction capacity, to avoid each zone from over-relying on external transaction purchases to regulate itself.

[0096] Back to Figure 2 , also includes:

[0097] Step S202: obtaining a virtual machine group corresponding to each partition by aggregating flexible resources in the partitions.

[0098] This aggregation method based on resource partition balancing innovatively considers the partition balancing issue in distributed resource aggregation, particularly in the partition objective function. By optimizing resource allocation, partition balancing is ultimately achieved. First, the spatial distribution of user resources (the first term) and the aggregation cost (the second term) are considered. Next, the deviation between the day-ahead plan and actual power consumption is considered, minimizing the deviation to reduce the penalty for failing to meet the plan. Finally, combining the energy flow between different regions, a multi-objective optimized resource partitioning result is obtained for aggregation, effectively improving the autonomous balancing capability and operational resilience of the partitioned aggregate. As a result, each partition functions as a "micro-balancing unit," independently managing local resources and reducing real-time reliance on the power grid. A cross-partition trading mechanism allows for inter-partition energy flow and mutual trading, enhancing overall system resilience.

[0099] Example 2:

[0100] like Figure 3 FIG. 1 is a flow chart showing a method for representing external resource characteristics in an embodiment of the present disclosure. Figure 3 The resource external characteristic representation method in the paper is used to accurately identify the feasible domain of flexible resources and even the characteristics of aggregates of flexible resources (such as virtual machine groups and virtual power plants) based on the characteristics of flexible resources (such as type, output, ramp rate and other restrictions). The obtained feasible domain can be applied as the external characteristic constraints of the virtual machine group, thereby obtaining more accurate operation and control decisions.

[0101] A feasible domain constructed directly using real historical data can only reflect past operational behavior and cannot account for potential changes under different future scenarios. While historical data can provide an "empirical response domain" for resources, it cannot predict future responsiveness under varying external conditions (such as load fluctuations, weather changes, and electricity price fluctuations). Furthermore, a feasible domain constructed directly from historical data can be distorted by data anomalies, leading to inaccurate modeling results. Therefore, directly using historical data can lead to overly conservative decisions and fail to effectively address future dynamic challenges.

[0102] To improve the accuracy and flexibility of identifying the external characteristics of a virtual power plant, this embodiment employs a data-driven forecasting approach. This method uses historical operational data to generate time series forecasts of resource output, thereby constructing a feasible domain for future forecasts. Unlike traditional static feasible domains constructed directly from historical data, the predictive feasible domain considers the potential changes in resource responses under different future external conditions, offering greater temporal adaptability and foresight.

[0103] exist Figure 3 In the method for expressing external resource characteristics, the method includes:

[0104] Step S301: Acquire historical operation data of each flexible resource.

[0105] In some embodiments, the historical operating data may include output-related data such as historical operating power, state of charge, meteorological parameters, electricity price curves, etc.

[0106] Step S302: pre-process the historical operation data to obtain input data.

[0107] In some embodiments, the preprocessing may include outlier removal, normalization, etc. As an example, the collected data may be normalized using the min-max method, as shown in the following formula:

[0108] ;

[0109] Among them, x norm is the value after data normalization, x is the data before normalization, and x max is the maximum value of the data before normalization, x min is the minimum value of the data before normalization.

[0110] Step S303: using the data-driven output time series prediction model corresponding to each flexible resource arrangement, predicting the predicted output sequence of each flexible resource at each future moment according to the input data.

[0111] In some embodiments, the output timing prediction model can be obtained based on, for example, a recurrent neural network model training, such as RNN, LSTM, or their variant models.

[0112] Taking LSTM as an example, Figure 4 As shown in the figure, a structural diagram of the LSTM network is shown.

[0113] The unit of the LSTM network consists of a cell, an input gate, an output gate, and a forget gate. The cell remembers the value of any time interval and the three gates regulate the flow of information in and out of the cell.

[0114] (1) Forget Gate

[0115] The forget gate decides which information to discard from the previous state by assigning it a value between 0 and 1. A value of 1 means to keep the information, and a value of 0 means to discard it, as shown in the following formula:

[0116] ;

[0117] Where W f is the weight matrix of the forget gate, It means connecting two vectors into a longer vector, b f is the bias term of the forget gate, is the Sigmoid activation function.

[0118] (2) Input gate

[0119] The input gate determines what new information to store in the current state, using the same system as the forget gate. The output gate controls the output in the current state by assigning a value between 0 and 1 to the information, taking into account the previous state output value and the current state input value. Selectively outputting relevant information from the current state allows the LSTM network to maintain useful long-term dependencies for prediction in the current and future time steps. This is shown in the following equation:

[0120] ;

[0121] ;

[0122] (3) Output gate

[0123] The output gate outputs the current cell state c t The calculation is based on the previous cell state c t-1 , element-wise multiplication by the forget gate f t , and then use the current input unit state Element-wise multiplication of the input gate i t , then add the two products together, so that the current memory c tand long-term memory c t-1 Combined together to form a new unit state c t . As shown in the following formula:

[0124] ;

[0125] ;

[0126] .

[0127] In some embodiments, each flexible resource can be locally configured with its own output timing prediction model, which can be trained based on the historical operating data of each flexible resource. Each flexible resource's local output timing prediction model can be sent to the local flexible resource for use after external training is completed. In other embodiments, the output timing prediction model corresponding to each flexible resource can also be configured externally to the flexible resource, such as centrally stored remotely, and can be trained remotely. The implementation method is not limited.

[0128] Step S304: constructing a feasible region for each flexible resource according to at least one operating characteristic constraint determined by the predicted output sequence.

[0129] Specifically, such as Figure 5 As shown, step S304 may include:

[0130] Step S501: determining at least one operating characteristic constraint corresponding to each flexible resource according to the operating characteristics of the flexible resources.

[0131] The operational characteristic constraints include output constraints, ramping constraints, and energy state constraints. These operational characteristic constraints can include time-invariant physical characteristic constraints inherent to the resource, such as the known upper and lower output limits, ramping limits, corresponding energy state limits, and the efficiency of converting output to energy state changes of the electrical equipment corresponding to the flexible resource. These can all be obtained from the electrical equipment's specifications and measured data.

[0132] Step S502: combining the at least one operating characteristic constraint and the elements of the predicted output sequence to obtain a unified feasible region.

[0133] For example, the combined operating characteristic constraints can be expressed as:

[0134] ;

[0135] Among them, the output power of flexible resource i at time t is expressed as P i,t , the corresponding energy state is , the upper and lower limits of the climbing rate are R i,min and Ri,max The upper and lower limits of the output power are P i,min and P i,max The upper and lower limits of the energy state are and , the efficiency coefficient of output power affecting energy state conversion is .

[0136] It should be noted that the meaning of this formula is that although the output P i,t It is in P i,min and P i,max However, the actual feasible region does not cover all P i,min and P i,max The entire range of values ​​between , but based on the predicted P i,t The range that can be achieved within this entire range shall prevail.

[0137] In some embodiments, the at least one operating characteristic constraint may also include multiple constraints applicable to each periodic moment of the flexible resource, for example, there is an output upper and lower limit constraint from 2:00 to 3:00 every Monday, and another output upper and lower limit constraint from 3:00 to 4:00; or, there is an output upper and lower limit constraint from 2:00 to 3:00 every day, and another output upper and lower limit constraint from 3:00 to 4:00, and so on.

[0138] In linear programming problems, the set of decision variables that satisfies all constraints is called the feasible region. The feasible region is a polyhedron, defined in the same way as the polyhedron: {x∈Rn|Ax≤b}. Based on the power and energy constraints of the flexible resource, the feasible region of flexible resource i can be expressed as a convex polyhedron by finding the intersection of the half-planes corresponding to the various constraints:

[0139] (3)

[0140] Among them, M i is the feasible region of the i-th flexible resource, P i is the output power vector of the i-th flexible resource in the scheduling period, A i and b i are the inequality coefficient matrix and constraint vector of the original feasible region of the i-th flexible resource; E is the identity matrix, is the energy dissipation coefficient, and They represent power characteristic constraints and power constraints respectively.

[0141] After accurately representing the feasible domain of flexible resources, the feasible domain of the aggregate of flexible resources can be obtained by aggregating the feasible domains of flexible resources. For example, the feasible domain of a virtual machine group of flexible resource aggregates can be obtained. Furthermore, the feasible domain of the virtual power plant response can be obtained by aggregating the feasible domains of various flexible resources under the virtual power plant, thus accurately characterizing the characteristics of the virtual power plant.

[0142] Step S305: A virtual machine group feasible region of each virtual machine group is obtained by aggregating the feasible regions of the flexible resources belonging to the same virtual machine group.

[0143] The virtual machine group may be determined based on the aggregation method based on resource partition balance as in Example 1. In some embodiments, the aggregation calculation method may be a method of solving the Minkowski sum.

[0144] like Figure 6 As shown in Figure 1, the Minkowski sum is the sum of the point sets of two Euclidean spaces, also known as the expansion set of these two spaces. It can be generally understood as the union of the areas passed by rotating along the boundary of the convex hull A for a convex hull B, that is, C.

[0145] The calculation of the usual Minkowski sum requires calculating the sum of the two vertices of the two convex hulls. Preferably, in order to reduce the computational complexity, an approximate calculation method can be used when aggregating the flexible resources of the virtual machine group. Specifically, an approximate polyhedron is obtained by translating and scaling the isomorphic polyhedron of the original convex polyhedron, and the original feasible domain of a single flexible resource is converted into an approximate polyhedron shape obtained by different scaling and translation of the same isomorphic polyhedron to approximate its maximum inner approximation, and then the Minkowski sum of the feasible domain in the form of the original convex polyhedron of each flexible resource is calculated and converted into the Minkowski sum between the maximum inner approximation feasible domains of each approximation based on the same isomorphic polyhedron, thereby achieving accurate acquisition of the feasible domain of the aggregate and effectively reducing the amount of calculation.

[0146] like Figure 7 As shown, the specific implementation process of step S305 is shown:

[0147] Step S701: The feasible regions of the flexible resources belonging to the same virtual machine group are represented as primitive convex polyhedrons by calculating the intersection of half planes.

[0148] That is, as shown in formula (3).

[0149] Step S702: selecting an isomorphic polyhedron based on each of the original convex polyhedrons.

[0150] As an example, similar to the feasible domain example of the original convex polyhedron form above, the isomorphic polyhedron can be expressed as:

[0151] ;

[0152] Optionally, the isomorphic polyhedron may select a portion of the operational characteristic constraints for construction.

[0153] Step S703: The isomorphic polyhedron is scaled and translated differently to obtain a maximum inner approximation to the original convex polyhedron of each flexible resource, thereby obtaining a maximum inner approximation feasible region of each flexible resource based on the same isomorphic polyhedron for aggregation of the virtual machine group.

[0154] Specifically, Figure 6 The figure shows two convex hulls of different shapes. If the two convex hulls have the same shape but differ only in size or position, then the union region obtained by moving one along the boundary of the other is still a convex hull of the same shape, with only the size and position changed. Therefore, the maximum inner approximation feasible domain of the feasible domain of the original convex polyhedron form of each flexible resource is constructed using the same isomorphic polyhedron, so that the Minkowski sum of the feasible domain of each original convex polyhedron form is approximately replaced by the Minkowski sum of the approximate maximum inner approximation feasible domain. When calculating the Minkowski sum of the maximum inner approximation feasible domain, since they are all replaced based on the same isomorphic polyhedron, the summation result will still be based on the isomorphic polyhedron, thus avoiding the calculation of adding the vertices of different convex polyhedrons one by one, and converting it into a superposition of the scaling and translation amounts of the isomorphic polyhedrons.

[0155] like Figure 8 As shown, the specific implementation of step S703 is shown.

[0156] exist Figure 8 In the embodiment, step S703 may specifically include:

[0157] Step S801: define the feasible domain of the approximate polyhedron form as the feasible domain of the isomorphic polyhedron form multiplied by the scaling factor and then added with the translation factor.

[0158] As an example, the feasible domain of the approximate polyhedral form of the isomorphic polyhedron is defined as:

[0159] ;

[0160] in, is the scaling factor, is the translation coefficient;

[0161] Step S802: Taking the original polyhedron of each flexible resource as the target of the maximum inner approximation of the approximate polyhedron, solve to obtain each set of maximum inner approximation scaling coefficients and maximum inner approximation translation coefficients, and based on each set of maximum inner approximation scaling coefficients, maximum inner approximation translation coefficients and the feasible domain in the form of the isomorphic polyhedron, obtain the maximum inner approximation feasible domain of each flexible resource based on the same said isomorphic polyhedron.

[0162] Continuing from the previous example, we take the maximum inner approximation of the feasible domain of the flexible resource i in the form of an approximate polyhedron as the goal and determine and , to obtain the maximum inner approximation feasible domain of flexible resource i. For example, the maximum value max that does not exceed the feasible domain of the original convex polyhedron form of flexible resource i can be firstly reached by scaling the coefficient In this case, we can solve the translation coefficient .

[0163] exist Figure 8 In the embodiment, step S305 may further include:

[0164] Step S803: Aggregate the maximum inner approximation feasible domain of the flexible resources belonging to the same virtual machine group to obtain the feasible domain of the virtual machine group obtained by applying the first aggregation scaling coefficient and the first aggregation translation coefficient based on the isomorphic polyhedron; wherein the scaling coefficient and the translation coefficient of the feasible domain of the virtual machine group are respectively the sum of the scaling coefficients and the sum of the translation coefficients of the maximum inner approximation feasible domains of each flexible resource belonging to it.

[0165] As an example, the feasible domain of the virtual machine group after flexible resource aggregation is defined as:

[0166] ;

[0167] In the formula is the flexible resource set in virtual machine group j, is the output power of the virtual machine group, ⊕ is the Minkowski sum symbol;

[0168] By accumulating the maximum inner approximation feasible regions of the same virtual machine group, the feasible region of the virtual machine group response is expressed as:

[0169] ;

[0170] in,

[0171] ; is the output power of virtual machine group j; is the flexible resource set in virtual machine group j; A VU is the inequality coefficient matrix of the virtual machine group.

[0172] Thus, after accurately representing the feasible domain of flexible resources, the feasible domain of the aggregate of flexible resources can be obtained by aggregating the feasible domain of flexible resources. For example, the above embodiment provides a feasible domain for a virtual machine group of virtual machines aggregated by flexible resources; furthermore, the feasible domain of the virtual power plant response can be obtained by aggregating the feasible domains of various flexible resources under the virtual power plant, thus accurately characterizing the characteristics of the virtual power plant.

[0173] Since the virtual machine group feasible domain of each virtual machine group has been obtained, the virtual machine group feasible domain can be aggregated to obtain the virtual power plant feasible domain.

[0174] Back to Figure 3 Optionally, the method further includes step S306: aggregating the feasible domains of each virtual machine group to obtain a feasible domain of a virtual power plant.

[0175] In some embodiments, step S306 may also adopt the above-mentioned method of calculating the simplified Minkowski sum of the isomorphic polyhedron approximating the maximum approximation to the original polyhedron to obtain the feasible region of the virtual power plant.

[0176] Through the described resource external characteristic representation method, resource endowment differences can be identified at the partition level and the feasible domain of regional virtual machine groups / virtual power plants can be accurately constructed, thereby realizing the external characteristic identification and modeling of virtual power plants with regional coordination capabilities, and providing accurate model support for subsequent virtual power plant operation scheduling optimization, trading decisions, etc.

[0177] Example 3:

[0178] like Figure 9 As shown, a flow chart of the operation control method in one embodiment of the present disclosure is shown.

[0179] exist Figure 9 In the process, the process includes:

[0180] Step S901: Obtain a maximum uncertainty loss model.

[0181] Among them, the maximum uncertain loss model is used to obtain the maximum uncertain loss under the constraints of the first operating constraint set of the power supply unit in the virtual power plant and when the ramp power scenario is determined, with the goal of maximizing the uncertain loss between the maximum benefit that can be achieved under the operating control decision of the uncertain flexible resource output scenario of the uncontrollable flexible resources and the current benefit with the minimum output deviation penalty cost that can be achieved by the current operating control decision.

[0182] First, define the output power P of the power supply unit (such as a virtual power plant or a virtual machine group) VPP Output P for the controllable flexible resources included opeThe output uncertainty scenario corresponds to the situation where the output power of uncontrollable flexible resources is uncertain (i.e., flexible resource output information is incomplete). For example, new energy power generation equipment, such as photovoltaic power stations and wind farms, are uncontrollable flexible resources, and photovoltaic power stations are set up. Other types of resources, such as gas / oil power generation equipment, energy storage, transferable loads, and curtailable loads, are controllable flexible resources. In a deterministic scenario, for example, if the output power of photovoltaic power stations and wind farms can be predicted relatively accurately. Of course, even if the output power of uncontrollable flexible resources can be predicted, there is no guarantee that the output of uncontrollable flexible resources will not change in accordance with the prediction, so uncertainty also needs to be considered.

[0183] The maximum uncertainty loss can be defined as the difference between the optimal solution that can be achieved by different decisions under information uncertainty (i.e., output uncertainty) and the optimal solution that can be achieved by the current decision.

[0184] The operation control decision includes a set of actions for the controllable flexible resources in the power supply unit, and may also include the day-ahead power at the same time. As an example, the operation control decision can be expressed as , represents the set of actions on the controllable flexible resources at time t; represents the day-ahead power at time t.

[0185] As an example, for controllable flexible resources such as gas / oil power generation equipment, energy storage, shiftable loads, curtailable loads, It can be expressed as , where the actions superscripted g, dis, ch, tran, and cut represent the action of gas / oil power generation equipment, charging, discharging, power transfer, and power reduction, respectively. It is understandable that the corresponding action will change as the type of controllable flexible resource changes. In some embodiments, the "action" can be the output power of the corresponding flexible resource being controlled. In some embodiments, since the power of each controllable flexible resource should be limited, the action set should also have a set of constraints on the corresponding actions, represented by M ope .

[0186] The supplyable power of the power supply unit includes the day-ahead power and the ramping power. The day-ahead power refers to the forecasted power for the output power of the next day. The ramping power refers to the power that flexible resources can quickly increase the power generation capacity to meet the sudden increase in the power demand of the system or compensate for it. In some embodiments, the first set of operating constraints includes a power balance constraint, and the power balance constraint refers to the sum of the actual output power and the output deviation related to the action set being equal to the sum of the day-ahead power and the ramping power. The output deviation refers to the deviation of the actual output power of the power supply unit from the predetermined output power, and the output deviation penalty cost is related to the output deviation.

[0187] As an example, after removing t, the output deviations under the operation and control decision of the flexible resource output uncertainty scenario and the current operation and control decision are P and B,u and P B , the day-ahead power is P m,u and P m , the climbing power is determined as P R , the actual output power is P VPP,u and P VPP , then the power balance constraint can be expressed as:

[0188] ;

[0189] .

[0190] Through this power balance constraint, the maximum uncertain loss model can associate the actual output power and the ramp power to obtain the maximum uncertain loss obtained by implementing the current operation control decision under the conditions of each determined ramp capacity scenario, which is expressed as .

[0191] As an example, for simplicity, the time symbol t is uniformly deleted, and each variable is the value at time t. The maximum uncertainty loss model is expressed as:

[0192] (4);

[0193] in, Indicates the current operation and control decision and the determined ramp capacity scenario P R The maximum uncertain loss under Indicates various output uncertainty scenarios; current operation and control decisions ; Indicates that the power supply unit is The total cost of flexible resource scheduling to execute operation and control decisions in the scenario, represents the total cost of flexible resource scheduling for the power supply unit to execute the current operation control decision, and the action set p ope Related; and C B They represent the output deviation penalty costs to be borne under the operation and control decision of the output uncertainty scenario and the current operation and control decision respectively; R m and They represent the day-ahead power benefits under the current operation and control decision and the operation and control decision under the output uncertainty scenario respectively; R R Indicates the climbing power gain.

[0194] R m and The calculation method is as follows: , is the day-ahead power value coefficient; R R The calculation method is as follows: , is the climbing power value coefficient. In some embodiments, can be the clearing price of the day-ahead market, The electricity price in the ramping service market can be used to measure the value of day-ahead power and ramping power, which corresponds to the revenue.

[0195] In some embodiments, the output deviation is calculated as follows: when the output deviation is greater than zero, the output deviation penalty cost is a positive value positively correlated with the output deviation; when the output deviation is not greater than zero, the output deviation penalty cost is a negative value negatively correlated with the output deviation.

[0196] As an example, and C B The calculation method is as follows: ;in, and is the deviation power cost coefficient. In some embodiments, the predetermined output power can be implemented as a pre-agreed output power. If the actual output power deviates from the agreed output power, a penalty cost may be incurred. If the actual output power is lower than the predetermined output power, the penalty cost is positive; if the actual output power is not lower than the predetermined output power, the penalty cost is 0 (i.e., no penalty) or negative (the excess output power also constitutes a benefit). and May be obtained from conventions or other references.

[0197] In some embodiments, the first set of operating constraints may include, in addition to the power balance constraint, external characteristic constraints of the power supply unit. For example, the external characteristic constraints may be related to the output of the power supply unit, such as one or more characteristics of the output upper and lower limits, ramp rate, response, etc. VPP To express it, we can get Among them, the external characteristic constraint M VPP The feasible region of the virtual power plant or virtual machine group obtained based on the resource external characteristic representation method in Example 2 can be formed.

[0198] In some embodiments, the first execution constraint set also includes an action set The action constraint set M mentioned in the previous embodiment should be satisfied ope ,Right now: .

[0199] In some embodiments, the first set of operating constraints further includes adjustable power constraints, for example, the adjustable power constraints are implemented as power restrictions for market access, such as adjustable power not less than a megawatts, adjustable time not less than b hours, etc.

[0200] As an example, the corresponding constraint of the power limit is expressed as:

[0201] ;

[0202] Where P m- 、P m+ The upper and lower limits of the bidding quantity required for market access.

[0203] Based on the maximum uncertainty loss model, given the current operation control decision and P R The maximum uncertainty loss can be solved as follows .

[0204] A simple analysis of the model principle:

[0205] According to Equation (4), the maximum benefit is obtained by searching for the optimal operation and control decision and the output deviation of the power supply unit under the operation and control decision in the output uncertainty scenario, with the goal of maximizing the first benefit. The current benefit is obtained by searching for the output deviation under the current operation and control decision, minimizing the power supply deviation cost, and maximizing the second benefit. The first and second benefits are represented by the difference between the sum of the day-ahead power benefit and the ramp power benefit and the sum of the total flexible resource scheduling cost and the output deviation penalty cost.

[0206] It can be further seen that the maximum benefit is obtained by searching for the optimal operation and control decision and the output deviation of the power supply unit under the operation and control decision of the output uncertainty scenario to maximize the first benefit, that is, the first minimization problem in formula (4):

[0207] ;

[0208] In this minimization problem, based on trying various operation control decisions under the output uncertainty scenario, 、p ope And the output deviation of the power supply unit The goal is to find the optimal operation and control decision to maximize the first benefit under the operation and control decision in the output uncertainty scenario.

[0209] The current benefit can be obtained by searching for the output deviation under the current operation and control decision to minimize the power supply deviation cost so as to maximize the second benefit, that is, the second minimization problem in formula (4):

[0210] ;

[0211] Among them, in the case of uncertain output scenario and given the current operation control decision, the output deviation P is searched B , which aims to find the minimum output deviation penalty cost under a given operation and control decision.

[0212] The outermost maximization problem in formula (4) aims to find the maximum profit difference, that is, the maximum opportunity loss, between the optimal solution (maximum benefit, i.e., maximum first benefit) under the output uncertainty scenario and the optimal solution that can be achieved by the current operation and control decision (current benefit with minimum deviation penalty cost, i.e., maximum second benefit).

[0213] Step S902: Obtain each set of maximum uncertain losses of each alternative operation control decision under a set of ramp power scenarios through the maximum uncertain loss model.

[0214] Step S903: Obtain each expected loss based on each group of maximum uncertain losses by minimizing the maximum expected loss model, and determine the corresponding alternative operation control decision as the target operation control decision according to the minimum value of each expected loss.

[0215] It can be understood that the maximum uncertain loss model takes into account the uncertainty of flexible resource output, and in steps S902 to S903, by changing the ramp power scenario of the maximum uncertain loss model (i.e., different ramp powers), the uncertainty of the ramp capacity is further considered, and the multiple uncertainties of the virtual power plant are considered, thereby obtaining more accurate and efficient target operation and control decisions, reducing the output deviation generated during operation, improving operation stability and increasing profits.

[0216] In some embodiments, each of the expected losses is obtained by calculating the weighted sum of a corresponding set of losses using the probability of occurrence of each of the ramp power scenarios as a weight value.

[0217] As an example, the minimization of the maximum expected loss model is expressed as:

[0218] (5);

[0219] Among them, p VPP represents an alternative operation and control decision, represents the climbing power under the Kth climbing power scenario, represents the probability of the Kth climbing power scenario.

[0220] By using the minimization of the maximum expected loss model, we can obtain K maximum uncertain losses by solving the problems under K ramping power scenarios in sequence. Since each ramping capacity situation corresponds to a certain probability , so it is used as a weight, and the product of each maximum uncertain loss and the probability of its corresponding scenario is added, that is, the weighted sum is obtained, and the decision p for an alternative operation control can be obtained. VPP The maximum uncertain loss expected loss is obtained. Therefore, the one with the minimum maximum uncertain loss expected loss among the alternative operation and control decisions is found, which is taken as the target operation and control decision.

[0221] Specifically, the maximum uncertain loss expected loss represents the operation and control decision under the worst uncertainty situation (the one that deviates most from the maximum benefit), and minimizing it is a way to find a target operation and control decision that can minimize the opportunity loss under the worst situation.

[0222] In some embodiments, the various ramp capacity conditions and their probabilities can be obtained by performing distribution statistics based on historical ramp capacity data.

[0223] In some embodiments, each of the expected losses is obtained by calculating the weighted sum of a corresponding set of losses using the probability of occurrence of each of the ramp power scenarios as a weight value; the minimization of the maximum expected loss model includes a second set of operating constraints, which includes: the first set of operating constraints, the adjustable power constraints under each ramp power scenario, and the sum of the weight values ​​is 1 (i.e. As an example, the adjustable power constraint in each ramping power scenario is, for example, the upper and lower limits that can be achieved under the market access power restriction in the ramping power scenario.

[0224] As an example, the adjustable power constraint may be expressed as:

[0225] ;

[0226] Among them, the is the adjustment parameter, between [0, 1], P RE.f Indicates the output fluctuation range of uncontrollable flexible resources, Indicates the output upper limit of uncontrollable flexible resources to ensure their safety.

[0227] It is understandable that Examples 1 to 3 are for intuitively explaining the linkage process between Examples 1 to 3. Figure 10 As shown, a schematic diagram showing the principle of the combined application of Examples 1 to 3 of the present disclosure is shown.

[0228] exist Figure 10 In the example 1, the resource aggregation method can be used to divide the flexible resources in the power grid into aggregation areas, forming flexible resource divisions for each virtual machine group. Further, optionally, the resource characteristic representation method in Example 2 can be used to obtain the virtual machine group feasible region and the virtual power plant response feasible region of the virtual machine groups corresponding to each aggregation area, and external characteristic constraints can be formed to apply to the first operating constraint set of the maximum uncertain loss model 1001 and the second operating constraint set of the minimization maximum expected loss model 1002 in Example 3.

[0229] The maximum uncertain loss model 1001 calculates the maximum uncertain losses of the alternative operation and control decisions under each ramping power scenario under the constraints of the first operation constraint set based on the given current operation and control decision and a set of ramping power scenarios, thereby obtaining a set of maximum uncertain losses corresponding to each current operation and control decision. The minimum maximum expected loss model 1002 calculates the maximum uncertain expected loss of each alternative operation and control decision based on each set of losses under the constraints of the second operation constraint set, and selects the alternative operation and control decision corresponding to the minimum expected loss as the target operation and control decision. Therefore, the target operation and control decision determined based on the minimized maximum uncertain expected loss can be used by the virtual power plant to control the corresponding flexible resources, which can effectively reduce the risk of output deviation caused by the output uncertainty of new energy flexible resources, the uncertainty of ramping capacity, etc.

[0230] By combining and implementing the methods of Example 1 to Example 3, the advantages of these methods can be superimposed, thereby obtaining more accurate and reliable operation and control results of the virtual power plant.

[0231] like Figure 11 FIG. 1 is a schematic diagram showing the effect of using the aggregation method based on resource partition balance to partition flexible resources in a network in an embodiment of the present disclosure.

[0232] The optimized division of the original distribution network yields four subnetworks (or zones), represented in the figure by VU1, VU2, VU3, and VU4. Each subnetwork has its own regional control center, representing the nodes where the coordination controllers for each virtual machine group reside, labeled C1, C2, C3, and C4, respectively. The resulting divisions clearly demonstrate that the locations of each virtual machine group within the network are geographically convenient and highly feasible for practical engineering applications.

[0233] It should be noted that the various functional modules in the above embodiments (such as the operation and control platform, various models, methods, etc.) can be implemented in whole or in part through software, hardware, firmware, or any combination thereof. When implemented using software, they can be implemented in whole or in part in the form of a computer program or instruction product. A computer program or instruction product includes one or at least one computer program or instruction. When the computer program or instruction is loaded and executed on a computer, the process or function according to the present disclosure is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer program or instruction can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another.

[0234] Furthermore, the methods disclosed in the previous embodiments can be implemented through other module division methods. The embodiments shown above are merely illustrative. For example, the module division is merely a logical function division. In actual implementation, there may be other division methods, such as at least one module or modules can be combined or dynamically transferred to another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between each other shown or discussed can be through some interface, indirect coupling or communication connection between devices or modules, and can be electrical or other forms.

[0235] like Figure 12 FIG. 1 is a schematic diagram showing the structure of a computer device in one embodiment of the present disclosure.

[0236] The computer device 1200 may be used to implement, for example, Figure 1 The operation control platform 102 in.

[0237] The computer device 1200 includes a bus 1201, a processor 1202, and a memory 1203. The processor 1202 and the memory 1203 can communicate with each other via the bus 1201. The memory 1203 can store computer programs or instructions. The processor 1202 executes the computer programs or instructions in the memory 1203 to implement the method of any of the above embodiments 1 to 3.

[0238] Bus 1201 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. Buses can be classified into address buses, data buses, and control buses. For ease of illustration, although only one thick line is used in the figure, this does not mean that there is only one bus or only one type of bus.

[0239] In some embodiments, the processor 1202 may be implemented as a central processing unit (CPU), a microprocessor unit (MCU), a system on a chip (SoC), or a field programmable gate array (FPGA). The memory 1203 may include volatile memory, such as random access memory (RAM), for temporarily storing data while running programs.

[0240] The memory 1203 may also include a non-volatile memory for data storage, such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD), or a solid-state disk (SSD).

[0241] In some embodiments, the computer device 1200 may further include a communicator 1204. The communicator 1204 is used to communicate with the outside world. In a specific example, the communicator 1204 may include one or at least one wired and / or wireless communication circuit module. For example, the communicator 1204 may include one or more of a wired network card, a USB module, a serial interface module, etc. The wireless communication protocol followed by the wireless communication module includes, for example, one or more of Nearfield Communication (NFC) technology, Infrared (IR) technology, Global System for Mobile Communications (GSM), General Packet Radio Service (GPRS), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), Time-Division Code Division Multiple Access (TD-SCDMA), Long Term Evolution (LTE), Bluetooth (BT), Global Navigation Satellite System (GNSS), etc.

[0242] In summary, the present disclosure relates to the technical field of electric energy allocation, and provides a method and platform based on resource partition balance aggregation, external characteristic representation and operation control. The aggregation method takes into account the resource spatial distribution distance, aggregation cost, partition plan power deviation, and the partition balance between local consumption and cross-regional transaction power, etc., and aims to minimize the partition loss caused by factors such as the partition, and determines the aggregation node and the partitions corresponding to the aggregation node and their corresponding virtual machine groups. The external characteristic representation method predicts the future output of flexible resources based on a data-driven prediction method to construct a resource feasible domain and its aggregate feasible domain based on foresight. The operation control method considers the resource output and ramp power conditions through a two-stage combination of the maximum uncertain loss model and the minimum maximum expected loss model to obtain an operation control decision that minimizes the maximum uncertain loss. Partition aggregation optimization is achieved respectively, and the external characteristic characterization accurately supports more decentralized user-side heterogeneous flexible resource control and improves the accuracy of control decisions.

[0243] The above embodiments are merely illustrative of the principles and effects of this disclosure and are not intended to limit this disclosure. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of this disclosure. Therefore, any equivalent modifications or alterations made by a person skilled in the art without departing from the spirit and technical concepts disclosed herein shall be encompassed by the scope of protection of this disclosure.

Claims

1. An aggregation method based on resource partition balance, characterized in that: Applied to a virtual power plant, the method is used to partition the flexible resources connected to nodes of a distribution network; the method comprises: According to the minimization of the partition objective function as the goal, the aggregation nodes in each node and the partitions corresponding to each aggregation node are determined; wherein the partition objective function includes: a first item based on the total distance between the global aggregation nodes, a second item based on the global aggregation cost, a third item based on the global partition plan power deviation, and a fourth item that does not meet the partition balance principle; wherein the global aggregation node total distance is defined as the sum of the sum of the distances between each current aggregation node and the adjacent aggregation node of each non-aggregate node; the global aggregation cost is defined as the sum of the aggregation costs of the coordination controllers of each aggregation node; the global partition plan power deviation is defined as the sum of the partition deviations between the day-ahead planned power and the real-time power of each partition; the partition balance principle refers to the tendency of the sum of the local power consumption of each partition to increase, and the sum of the cross-region transaction input power between partitions to decrease; By aggregating the flexible resources in the partitions, a virtual machine group corresponding to each partition is obtained.

2. The aggregation method based on resource partition balance according to claim 1, characterized in that: The second item includes a binary decision variable, which is used to indicate whether the node is an aggregate node and is used as a coefficient for the cost related to the aggregate node to exclude non-aggregate nodes from the calculation of the aggregate cost; the first item includes an auxiliary variable, which is constructed based on the product of a binary decision variable of a current node and mutually exclusive variables of the binary decision variables of other nodes and is used as a coefficient for the distance to exclude the distance between two non-aggregate nodes and the distance from an aggregate node to a node via another aggregate node from the sum of the distances.

3. The aggregation method based on resource partition balance according to claim 1, characterized in that: The partition objective function has a partition constraint set, which includes: an upper limit on the total number of virtual machine groups; a maximum distance limit from an aggregation node to a boundary node of the partition to which it belongs; a first preset ratio of the amount of electricity consumed on-site within the partition being higher than the amount of power generated within the partition; the total amount of cross-regional transactions between each partition and other areas does not exceed an upper limit on the transaction capacity, and the upper limit on the transaction capacity is determined based on a second preset ratio of the total power generation of each of the partitions.

4. A method for expressing external resource characteristics, characterized in that: include: Obtain historical operating data for each flexible resource; Preprocessing the historical operation data to obtain input data; Using a data-driven output data time series prediction model corresponding to each flexible resource arrangement, predicting a predicted output sequence of each flexible resource at each future moment based on the input data; constructing a feasible region for each flexible resource according to at least one operating characteristic constraint determined by the predicted output sequence; A virtual machine group feasible domain for each virtual machine group is obtained by aggregating feasible domains of flexible resources belonging to the same virtual machine group; wherein the virtual machine group is determined based on the aggregation method based on resource partition balancing according to any one of claims 1 to 3; and the aggregation method of the feasible domain includes calculating a Minkowski sum of the feasible domain.

5. The method for expressing off-resource characteristics according to claim 4, wherein: The constructing of a feasible domain for each flexible resource according to at least one set of operating characteristic constraints determined by the predicted output sequence includes: Determining at least one operating characteristic constraint corresponding to each flexible resource based on operating characteristics of the flexible resource; the operating characteristic constraint includes: an output constraint, a ramp constraint, and an energy state constraint; the at least one operating characteristic constraint is applicable to the flexible resource at all times or at each periodic time; The at least one operating characteristic constraint and the elements of the predicted output sequence are combined to obtain a unified feasible domain.

6. The method for expressing off-resource characteristics according to claim 4, wherein: By aggregating the feasible domains of the flexible resources belonging to the same virtual machine group, the virtual machine group feasible domain of each virtual machine group is obtained, including: The feasible domain of flexible resources belonging to the same virtual machine group is represented as a primitive convex polyhedron by finding the intersection of half planes. Selecting a homomorphic polyhedron based on each of the original convex polyhedrons; The isomorphic polyhedrons are scaled and translated differently to obtain maximum inner approximation to the original convex polyhedron of each flexible resource, thereby obtaining a maximum inner approximation feasible region of each flexible resource based on the same isomorphic polyhedron for aggregation of the virtual machine group.

7. The method for expressing external resource characteristics according to claim 6, wherein: The method of obtaining a maximum inner approximation of an original convex polyhedron of each flexible resource by scaling and translating the isomorphic polyhedron in different ways, and obtaining a maximum inner approximation feasible region of each flexible resource based on the same isomorphic polyhedron for aggregating the virtual machine group, includes: The feasible domain of the approximate polyhedron form is defined as the feasible domain of the isomorphic polyhedron form multiplied by the scaling factor and then added with the translation factor; Taking the approximate polyhedron's maximum inner approximation of the original polyhedron of each flexible resource as the goal, solving to obtain each set of maximum inner approximation scaling coefficients and maximum inner approximation translation coefficients, and obtaining the maximum inner approximation feasible region of each flexible resource based on the same isomorphic polyhedron based on each set of maximum inner approximation scaling coefficients, maximum inner approximation translation coefficients and the feasible region of the isomorphic polyhedron form; The step of obtaining the virtual machine group feasible region of each virtual machine group by aggregating the feasible regions of the flexible resources belonging to the same virtual machine group further includes: Aggregate the maximum inner approximation feasible domains of the flexible resources belonging to the same virtual machine group to obtain the virtual machine group feasible domain obtained by applying a first aggregation scaling coefficient and a first aggregation translation coefficient based on the isomorphic polyhedron; wherein the scaling coefficient and the translation coefficient of the virtual machine group feasible domain are respectively the sum of the scaling coefficients and the sum of the translation coefficients of the maximum inner approximation feasible domains of each flexible resource belonging to it.

8. The method for expressing off-resource characteristics according to claim 4, wherein: Also includes: The feasible regions of the virtual machine groups are aggregated to obtain the feasible region of the virtual power plant.

9. An operation control method, characterized in that: include: Obtain a maximum uncertain loss model for obtaining the maximum uncertain loss under the constraints of a first operating constraint set of a power supply unit in a virtual power plant and when a ramping power scenario is determined, with the goal of maximizing the uncertain loss between the maximum benefit that can be achieved under the operating control decision of an uncertain flexible resource output scenario of an uncontrollable flexible resource and the current benefit with the minimum output deviation penalty cost that can be achieved by the current operating control decision; the power supply unit is implemented as a virtual power plant or a virtual machine group; the operating control decision includes an action set for controllable flexible resources in the power supply unit; the first operating constraint set includes external characteristic constraints for the power supply unit, and the external characteristic constraints are formed based on the feasible domain of the virtual power plant or virtual machine group obtained by the resource external characteristic representation method according to any one of claims 4 to 8; Obtaining each set of maximum uncertain losses of each alternative operation control decision under a set of ramp power scenarios through the maximum uncertain loss model; Each expected loss is obtained based on each group of maximum uncertain losses by minimizing the maximum expected loss model, and the corresponding alternative operation control decision is determined as the target operation control decision according to the minimum value of each expected loss.

10. The operation control method according to claim 9, characterized in that: The controllable flexible resources include one or more of energy storage equipment, gas / oil power generation equipment, transferable loads and curtailable loads; the action set includes one or more of gas / oil power generation equipment action, charging, discharging, power transfer and power curtailment; and / or, the uncontrollable flexible resources include new energy power generation equipment.

11. The operation control method according to claim 9, characterized in that: The first set of operating constraints includes a power balance constraint, which means that the sum of the actual output power of the power supply unit and the output deviation is equal to the sum of the day-ahead power and the ramp power; the output deviation refers to the deviation of the actual output power of the power supply unit from the predetermined output power, and the output deviation penalty cost is related to the output deviation; the operating control decision includes an action set regarding controllable flexible resources; the maximum benefit is obtained by searching for the optimal operating control decision and the output deviation of the virtual machine group under the operating control decision in the output uncertainty scenario, with the goal of maximizing the first benefit; The current benefit is obtained by searching for the output deviation under the current operation control decision to minimize the power supply deviation cost so as to maximize the second benefit obtained; The first benefit and the second benefit are represented by the difference between the sum of the day-ahead power benefit and the ramp power benefit and the sum of the total flexible resource scheduling cost and the output deviation penalty cost.

12. The operation control method according to claim 11, characterized in that: The output deviation penalty cost is related to the output deviation amount; the output deviation amount refers to the deviation of the actual output power of the virtual machine group compared to the predetermined output power; the calculation method of the output deviation amount includes: When the output deviation is greater than zero, the output deviation penalty cost is a positive value that is positively correlated with the output deviation; When the output deviation is not greater than zero, the output deviation penalty cost is a negative value that is negatively correlated with the output deviation.

13. The operation control method according to claim 9, characterized in that: The maximum uncertainty loss model is expressed as: ； in, Indicates the current operation and control decision and the determined ramp capacity scenario P R The maximum uncertain loss under Indicates various output uncertainty scenarios; current operation and control decisions , p ope represents the set of actions on controllable flexible resources; P m and P m,u They represent the day-ahead power under the current operation and control decision and the operation and control decision under the output uncertainty scenario respectively; Indicates the output deviation under various output uncertainty scenarios; Indicates that the power supply unit is The total cost of flexible resource scheduling to execute operation and control decisions in the scenario, Represents the total cost of flexible resource scheduling for the power supply unit to execute the current operation control decision; and C B They represent the output deviation penalty costs to be borne under the operation and control decision of the output uncertainty scenario and the current operation and control decision respectively; R m and They represent the day-ahead power benefits under the current operation and control decision and the operation and control decision under the output uncertainty scenario respectively; R R Indicates the climbing power gain; R m and The calculation method is as follows: , is the day-ahead power value coefficient; R R The calculation method is as follows: , is the climbing power value coefficient; and C B The calculation method is as follows: ； in, and is the deviation power penalty coefficient.

14. The operation control method according to claim 9, characterized in that: The first operating constraint set includes: external characteristic constraints of the power supply unit, output power characteristic constraints of various controllable flexible resources inside the power supply unit, and at least one of adjustable power constraints; wherein, the external characteristic constraints are used to describe the characteristics of the power supply unit; the output power characteristic constraints include an action constraint set for the action set.

15. The operation control method according to claim 9, characterized in that: The minimization of the maximum expected loss model is expressed as: ； Among them, p VPP Indicates the current operation and control decision. represents the climbing power under the Kth climbing power scenario, represents the probability of the Kth climbing power scenario.

16. The operation control method according to claim 9, characterized in that: Each of the expected losses is obtained by calculating the weighted sum of a corresponding set of losses using the probability of occurrence of each of the climbing power scenarios as a weight value; the model for minimizing the maximum expected loss includes a second operating constraint set, which includes: the first operating constraint set, the adjustable power constraints under each climbing power scenario, and the sum of each weight value is 1.

17. An operation control platform, characterized in that: include: processor and memory; The memory stores program instructions; The processor is configured to execute the program instructions to perform at least one of the following: The aggregation method based on resource partition balancing according to any one of claims 1 to 3; The method for expressing off-resource characteristics according to any one of claims 4 to 8; Execute the operation control method according to any one of claims 9 to 16.

Images