Random unit combination accelerated solution method and related device

Abstract

The invention discloses a random unit commitment acceleration solving method and a related device. The method comprises the following steps: acquiring power system input characteristics of a current scheduling period; the power system input characteristics of the current dispatching period are input into a trained unit combination start-stop variable prediction model, a unit start-stop state initial solution is obtained, and the unit combination start-stop variable prediction model is trained according to a variable fixing mechanism of a confidence coefficient threshold value; and recovering the initial solution of the start-stop state of the unit to obtain an optimal start-stop scheme of the unit, and the method and the related device can greatly shorten the solving time while ensuring the feasibility of the solution.

Description

Technical Field

[0001] This invention belongs to the field of power system planning and operation technology, and relates to a method and related apparatus for accelerating the solution of stochastic unit combinations. Background Technology

[0002] With the increasing penetration of renewable energy sources such as wind and solar power (RES) in modern power systems, their inherent uncertainties pose a severe challenge to the safe and stable operation of the power grid, and significantly increase the complexity and difficulty of solving the Unit Commitment (UC) problem. As the core task of power system optimal dispatch, the UC problem aims to determine the start-up and shutdown status and output plan of each generating unit within the dispatch cycle, while ensuring supply-demand balance and system reliability, thereby maximizing economic efficiency. Traditional methods typically establish deterministic mixed-integer linear programming (MILP) models based on the predicted values ​​of renewable energy. However, under the conditions of large-scale, highly volatile renewable energy, prediction errors can seriously affect system security. To enhance the robustness of dispatch schemes, the academic community has proposed stochastic UC models to explicitly introduce the uncertainties of renewable energy. Currently, related research mainly includes three types of modeling methods: scenario-based, uncertainty set-based, and chance-constraint-based. Although stochastic UC models can more accurately characterize uncertainty, the high-dimensional coupling between binary and continuous variables involved, and the rapidly expanding constraint scale due to the increase in the number of generating units, dispatch cycles, and scenarios, make it difficult for existing solvers to obtain the optimal solution within a reasonable time, even when considering only a finite number of scenarios. To address this problem, researchers have proposed various acceleration strategies, such as acceleration algorithms based on sample statistics and optimization kernels, data modeling methods based on load-solution space databases, and variable dimensionality reduction techniques based on global sensitivity analysis. With the rapid development of artificial intelligence, more and more data-driven algorithms (especially machine learning methods) are being introduced into UC problems to improve modeling and solution efficiency. Some methods attempt to completely replace the traditional optimization process with learning models, while others simplify the model structure and reduce computational complexity through constraint filtering and integer variable preprocessing. Under wind power uncertainty, CNN-based schemes are also used to accelerate binary decision-making, or deep reinforcement learning is introduced to generate efficient scheduling strategies that satisfy constraints. However, current methods still have shortcomings, such as a lack of neural network structures adapted to complex and uncertain inputs, neglect of redundancy and inherent patterns among scene data, and a lack of systematic analysis and assurance regarding the feasibility and reliability of data-driven results, which urgently require further research and solutions. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and apparatus for accelerating the solution of random unit combinations. This method and apparatus can significantly shorten the solution time while ensuring the feasibility of the solution.

[0004] To achieve the above objectives, this invention discloses a method for accelerating the solution of randomized unit combinations, comprising:

[0005] Obtain the power system input characteristics for the current scheduling period;

[0006] The power system input features of the current scheduling cycle are input into the trained unit combination start-up and shutdown variable prediction model to obtain the initial solution of the unit start-up and shutdown state. The unit combination start-up and shutdown variable prediction model is trained according to the variable fixing mechanism of confidence threshold.

[0007] The initial solution of the unit start-up and shutdown state is recovered to obtain the optimal unit start-up and shutdown scheme.

[0008] A further improvement of the accelerated solution method for randomized unit combination described in this invention is as follows:

[0009] Furthermore, before inputting the power system input features of the current scheduling period into the trained unit combination start-up and shutdown variable prediction model, the following steps are also included:

[0010] Build an optimized solution database;

[0011] Construct a predictive model for unit combination start-up and shutdown variables;

[0012] Based on the confidence threshold variable fixing mechanism, the optimized solution database is used to train the unit combination start-up and shutdown variable prediction model to obtain the trained unit combination start-up and shutdown variable prediction model.

[0013] Furthermore, the process of constructing the optimized solution database is as follows:

[0014] Constructing a randomized set combination problem;

[0015] Several uncertainties regarding renewable energy are generated;

[0016] Under various renewable energy uncertainty scenarios, the stochastic unit combination problem is solved to obtain the unit start-up and shutdown schemes under each renewable energy uncertainty scenario;

[0017] An optimized solution database is constructed based on the unit start-up and shutdown schemes under the aforementioned uncertainties of renewable energy.

[0018] Furthermore, the objective function of the randomized grouping problem is:

[0019]

[0020] Where t, s, g and e represent the indexes of scheduling time step, scheduling scenario, thermal power unit and energy storage system, respectively; and ε respectively represent the scheduling time step set, the scenario set, the thermal power unit set and the energy storage system set; π s represents the occurrence probability of scenario s; π s is u g,t and d g,t respectively represent whether the thermal power unit g starts or stops at time step t; p g,t,s represents the active power output of the thermal power unit g at time step t under scenario s; p e,t,s represents the charge and discharge power of the energy storage system e at time step t under scenario s; C g (·) represents the power generation cost function of the thermal power unit g; and respectively represent the start-up cost coefficient and the shutdown cost coefficient of the thermal power unit g; C e represents the charge and discharge cost coefficient of the energy storage system e.

[0021] Further, the constraint conditions of the random unit commitment problem include unit start-stop logic constraints, output upper and lower limits and ramping constraints, minimum start-stop time constraints, energy storage operation constraints, system supply-demand balance constraints, reserve constraints and renewable energy accommodation constraints.

[0022] Further, the process of recovering the initial solution of the unit start-stop state to obtain the optimal unit start-stop scheme is:

[0023] The initial solution of the unit start-stop state is recovered by a feasibility recovery mechanism to obtain the optimal unit start-stop scheme.

[0024] The application discloses a random unit commitment accelerated solving system, comprising:

[0025] An acquisition module is configured to acquire power system input features of a current scheduling period.

[0026] A solving module is configured to input the power system input features of the current scheduling period into a trained unit commitment start-stop variable prediction model to obtain an initial solution of a unit start-stop state, wherein the unit commitment start-stop variable prediction model is trained according to a variable fixing mechanism of a confidence threshold.

[0027] A recovery module is configured to recover the initial solution of the unit start-stop state to obtain an optimal unit start-stop scheme.

[0028] The random unit commitment accelerated solving system further improves in that:

[0029] Further, the system further comprises:

[0030] A first construction module is configured to construct an optimization solution database.

[0031] a second construction module, configured to construct a unit commitment start-stop variable prediction model;

[0032] a training module, configured to train the unit commitment start-stop variable prediction model according to a variable fixing mechanism of a confidence threshold, by using the optimization solution database, to obtain a trained unit commitment start-stop variable prediction model.

[0033] Further, the first construction module comprises:

[0034] a first construction unit, configured to construct a stochastic unit commitment problem;

[0035] a generation unit, configured to generate a plurality of renewable energy uncertainty scenarios;

[0036] a solution unit, configured to solve the stochastic unit commitment problem under each renewable energy uncertainty scenario, to obtain a unit start-stop scheme under each renewable energy uncertainty scenario;

[0037] a second construction unit, configured to construct an optimization solution database based on the unit start-stop scheme under each renewable energy uncertainty scenario.

[0038] Further, an objective function of the stochastic unit commitment problem is:

[0039]

[0040] where t, s, g and e respectively represent indexes of a scheduling time step, a scheduling scenario, a thermal power unit and an energy storage system; and ε respectively represent a scheduling time step set, a scenario set, a thermal power unit set and an energy storage system set; π s represents an occurrence probability of the scenario s; π s u g,t and d g,t respectively represent whether the thermal power unit g starts or stops at the time step t; p g,t,s represents active power output of the thermal power unit g at the time step t under the scenario s; p e,t,s represents charge-discharge power of the energy storage system e at the time step t under the scenario s; C g (·) represents a power generation cost function of the thermal power unit g; and respectively represent a start cost coefficient and a stop cost coefficient of the thermal power unit g; C e represents a charge-discharge cost coefficient of the energy storage system e.

[0041] ​Furthermore, the constraints of the stochastic unit combination problem include unit start-up and shutdown logic constraints, output upper and lower limits and ramping constraints, minimum start-up and shutdown time constraints, energy storage operation constraints, system supply and demand balance constraints, reserve constraints, and renewable energy acceptance constraints.

[0042] Furthermore, the process of recovering the initial solution of the unit start-up and shutdown state to obtain the optimal unit start-up and shutdown scheme is as follows:

[0043] The initial solution of the unit start-up and shutdown state is restored by a feasibility recovery mechanism to obtain the optimal unit start-up and shutdown scheme.

[0044] The present invention discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the random group combination accelerated solution method.

[0045] The present invention discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the accelerated solution method for random unit combination.

[0046] The present invention has the following beneficial effects:

[0047] In practical operation, the stochastic unit combination accelerated solution method and related apparatus described in this invention input the power system input characteristics of the current scheduling cycle into the trained unit combination start-up and shutdown variable prediction model to obtain the initial solution of the unit start-up and shutdown state, thus ensuring the feasibility of the solution. Furthermore, the unit combination start-up and shutdown variable prediction model in this invention is trained based on a confidence threshold variable fixing mechanism. This mechanism allows for the selection and fixing of reliable variables, thereby transforming the original large-scale mixed integer model into multiple sub-problems with fewer constraints and dominated by continuous variables. This improves solution efficiency, reduces computational resource usage, and shortens solution time. Attached Figure Description

[0048] The accompanying drawings, which form part of this specification, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0049] Figure 1 This is a flowchart of the method of the present invention;

[0050] Figure 2 This is a diagram illustrating the architecture of the generator start-up and shutdown variable prediction model in this invention.

[0051] Figure 3 This is a system structure diagram of the present invention. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] In the description of this invention, it should be understood that the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0054] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0055] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Additionally, the character " / " in this invention generally indicates that the preceding and following objects have an "or" relationship.

[0056] It should be understood that although terms such as first, second, third, etc., may be used in the embodiments of the present invention to describe the preset range, these preset ranges should not be limited to these terms. These terms are only used to distinguish the preset ranges from one another. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0057] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0058] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0059] The accompanying drawings illustrate various structural schematic diagrams according to embodiments disclosed in this invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0060] Example 1

[0061] refer to Figure 1 The accelerated solution method for randomized unit combinations described in this invention includes the following steps:

[0062] 1) Obtain the power system input characteristics for the current scheduling period, which include the predicted load demand of each bus and the multi-dimensional renewable energy output scenario;

[0063] 2) Input the power system input features of the current scheduling cycle into the trained unit combination start-up and shutdown variable prediction model to obtain the initial solution of the unit start-up and shutdown state;

[0064] The specific process of step 2) is as follows:

[0065] 21) Constructing a randomized grouping problem;

[0066] The Stochastic Unit Commitment (SUC) problem is a scheduling problem that optimizes the start-up and shutdown schedules and output arrangements of generating units while considering uncertainties such as renewable energy sources in the power system. Unlike traditional deterministic models, SUC introduces the uncertainty of renewable energy output, such as wind and solar power, into the model as scenarios, simultaneously optimizing unit combination schemes under multiple possible future conditions to improve the robustness of the scheduling strategy and the operational reliability of the system. The SUC is essentially a large-scale mixed-integer optimization problem, requiring a balance between operating costs and feasibility under multiple scenarios while ensuring system supply and demand balance, start-up and shutdown logic constraints, and ramp-up constraints. Therefore, it has high computational complexity and is a key issue in power system scheduling and optimization.

[0067] The objective function of the randomized grouping problem is:

[0068]

[0069] Where t, s, g, and e represent the indexes of the scheduling time step, scheduling scenario, thermal power unit, and energy storage system (ESS), respectively. And ε represents the set of scheduling time steps, the set of scenarios, the set of thermal power units, and the set of energy storage systems, respectively, and π s π represents the probability of scenario s occurring. s for u g,t and d g,t These are binary variables, representing whether thermal power unit g starts or stops at time step t; p g,t,s This represents the active power output of thermal power unit g in scenario s at time step t; C represents the charging and discharging power of energy storage system e in scenario s at time step t; g (·) represents the power generation cost function of thermal power unit g; and C represents the start-up cost coefficient and shutdown cost coefficient of thermal power unit g, respectively; e ESSs represents the charge and discharge cost coefficient of energy storage system e.

[0070] The constraints include unit start-up and shutdown logic constraints, output upper and lower limits and ramping constraints, minimum start-up and shutdown time constraints, energy storage operation constraints, system supply and demand balance constraints, reserve constraints and renewable energy acceptance constraints.

[0071] The minimum start-stop constraint for thermal power units is expressed as:

[0072]

[0073] Where, x g,tA binary variable representing whether thermal power unit g is started or stopped at time t; and These represent the minimum continuous start-up time and minimum continuous shutdown time of unit g, respectively.

[0074] The logical constraints related to start / stop variables are represented as follows:

[0075]

[0076] The output constraint of thermal power units is expressed as:

[0077]

[0078] in, as well as These represent the minimum and maximum active power output limits for thermal power unit g, respectively.

[0079] The ramping constraint for thermal power units is expressed as:

[0080]

[0081] in, These represent the upper and lower limits of the ramp gradient for thermal power unit g, respectively. and These represent the starting output limit and the shutdown output limit of thermal power unit g, respectively.

[0082] The operating constraints of an energy storage system are expressed as follows:

[0083]

[0084] Among them, E e,t,s This represents the state of charge (SOC) of the energy storage system (ESS) e at time step t in scenario s; as well as These represent the minimum state-of-charge limit and the maximum state-of-charge limit for energy storage system e, respectively. as well as Δt represents the maximum discharge power limit and the maximum charging power limit of the energy storage system e, respectively; Δt represents the duration of the scheduling time interval.

[0085] The system standby constraint is expressed as:

[0086]

[0087] Where α represents the system reserve factor; d represents the load index; D represents the load set; p d,t,s This represents the active power demand of load d at time step t in scenario s.

[0088] The system power balance constraint is expressed as:

[0089]

[0090] Where r represents the index of renewable energy; p represents a set of renewable energy sources. r,t,s This represents the active power output of renewable energy r at time step t in scenario s.

[0091] The system line power flow constraints are expressed as follows:

[0092]

[0093] Where n and l represent the indices of the busbar and transmission line, respectively; and These represent the sets of busbars and transmission lines, respectively. K represents the maximum permissible transmission power of transmission line l. l,n denoted as the power flow distribution factor (PTDF) from bus n to line l.

[0094] 22) Construct an optimized solution database;

[0095] To construct a high-quality database of optimized solutions covering various load levels and renewable energy fluctuation scenarios, this invention employs a conditional generative adversarial network (CGAN) to generate large-scale renewable energy uncertainty scenarios, and combines an optimization solver to solve the stochastic unit combination problem offline.

[0096] This paper utilizes CGAN to generate realistic renewable energy output scenarios under given load levels and weather conditions. CGAN consists of a generator G and a discriminator D, and its objective is to learn the conditional probability distribution P(x). r c), where x r The output sequence of new energy sources is represented by the time step T; c represents the conditional input features, including load features and various types of weather features.

[0097] The optimization objective of CGAN is expressed as:

[0098]

[0099] Trained CGANs can be used to quickly generate large-scale new energy power output scenarios that conform to historical statistical characteristics.

[0100] For each generated scene Using these as input scenario parameters, the mixed integer optimization method is called to solve the random unit combination problem. Each set of uncertain scenarios, a fixed load water and its corresponding unit start-up and shutdown scheme are used as training samples, and an optimized solution database is constructed based on this.

[0101] 23) Construct a predictive model for unit start-up and shutdown variables;

[0102] The input features of the unit start-up and shutdown variable prediction model consist of multiple renewable energy (RES) output scenarios and deterministic load demand curves. The input scenarios are represented as follows:

[0103]

[0104] in, This represents the total load on bus n minus the total renewable energy generation at time t and under scenario s. By calculating the net load, the actual dispatch pressure of the system under different scenarios can be effectively reflected, thereby transforming complex renewable energy uncertainty information into structured input features.

[0105] refer to Figure 2 The unit combination start-up and shutdown variable prediction model includes a time-series coding module, a topology modeling module, and a time modeling module. The time-series coding module uses a convolutional neural network (CNN) to extract multi-scale features from renewable energy output and load data in multiple scenarios, capturing temporal dynamics and scenario correlations. The topology modeling module is based on a graph convolutional network (GCN) and uses the grid adjacency matrix to aggregate node and neighbor information to mine the spatial dependencies of the power system. The time modeling module uses a recurrent neural network based on LSTM and models temporal dependencies through a gating mechanism to capture the continuity and logical structure of unit start-up and shutdown states over time, thereby more accurately focusing on the state evolution at critical scheduling moments.

[0106] The fusion features of the input renewable energy output and load data from multiple scenarios are represented as a three-dimensional tensor, denoted as . Where N is the number of nodes; T is the number of time steps; and S is the number of scenes. A convolution operation is performed on the time series of each node n.

[0107] The convolution process is represented as:

[0108] H (1) =conv1D(X; W) (1) ,b (1) )

[0109] H (l) =Relu(conv1D(H (l-1) W (l) ,b (l) ))

[0110] Therefore, the input of this module is used as the input of the next topology modeling module. Let the adjacency matrix of the power grid be represented as:

[0111]

[0112] Where A is a binary matrix used to describe the topology of the power network, and each element a in the binary matrix is ​​a ij This represents the connection relationship between the i-th and j-th buses in the system. At the same time, considering the information of each node itself, the adjacency matrix A is usually added to the identity matrix I.

[0113] Let the degree matrix describe the number of edges connected to each node as follows:

[0114]

[0115] The output of the graph convolution operation layer l is:

[0116]

[0117] The final output is used as the spatial features of the nodes, and then as the input to the next temporal modeling module.

[0118] For each time step t, the LSTM controls the update process of cell states and hidden states through forget gates, input gates, and output gates, ensuring that the model can flexibly remember and forget historical information. The specific update process is represented as follows:

[0119] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0120] i t =σ(W i ·[h t-1 ,x t ]+b i )

[0121]

[0122]

[0123] o t =σ(W o ·[h t-1 ,x t ]+b o )

[0124] h t =o t ⊙tanh(C t )

[0125] Where, x t This represents the input at the current time step; h t-1 C t-1These represent the hidden state and memory state of the previous time step, respectively; σ represents the Sigmoid function; tanh represents the hyperbolic tangent function; ⊙ represents the element-wise product; W * and b * Here are the weight matrix and bias vector to be trained; [·] indicates that the two vectors are concatenated to form a long vector.

[0126] Finally, the fully connected neural network layer outputs the predicted probability of the integer control variable for each time step. The final output is represented as follows:

[0127]

[0128] The Binary Cross-Entropy (BCE) loss function is used to train the aircraft start-up and shutdown state prediction model. This loss function treats the prediction of start-up and shutdown states as a binary classification task, and its goal is to minimize the difference between the model output and the true label. The BCE loss is expressed as:

[0129]

[0130] Where, x g,t This indicates the actual start-up and shutdown status of unit g at time t; This represents the predicted probability of start-stop status, with a value range of [0,1].

[0131] 24) The unit combination start-up and shutdown variable prediction model is trained by optimizing the solution database to obtain the trained unit combination start-up and shutdown variable prediction model;

[0132] Specifically, based on the optimized solution database, data preprocessing is first performed to extract key variables such as load and renewable energy output of each bus at each time step. After standardization, a three-dimensional tensor is formed, which is used as the input to the unit combination start-up and shutdown variable prediction model, labeled with the optimal unit start-up and shutdown scheme under the corresponding load scenario. Then, the model training phase begins. The unit combination start-up and shutdown variable prediction model optimizes its parameters by minimizing the binary cross-entropy loss between the predicted output and the actual start-up and shutdown state. After training, the model parameters and structure are saved together, resulting in the trained unit combination start-up and shutdown variable prediction model.

[0133] 25) Input the power system input features of the current scheduling cycle into the trained unit combination start-up and shutdown variable prediction model to obtain the initial solution of the unit start-up and shutdown state.

[0134] 3) The initial solution of the unit start-up and shutdown state is restored through the feasibility recovery mechanism to obtain the optimal unit start-up and shutdown scheme.

[0135] Furthermore, to ensure that the initial solution of the unit start-up and shutdown states satisfies the feasibility constraints of the scheduling model, this invention introduces a feasibility recovery mechanism to recover the initial solution of the unit start-up and shutdown states, thereby obtaining the optimal unit start-up and shutdown scheme. Specifically, for the parts of the unit start-up and shutdown states that violate logical constraints such as minimum start-up and shutdown times, a local optimization method is used for correction. A small-scale integer optimization subproblem is constructed using the minimum distance metric method. Under the premise of keeping the states of other time steps unchanged, the start-up and shutdown variables in this interval are adjusted to meet the minimum continuous start-up or shutdown time requirements, and to maintain consistency with the original prediction results as much as possible. Since the subproblem is small in scale, it can be solved quickly. The integer optimization subproblem is expressed as:

[0136]

[0137] in, This represents the unit combination scheme of thermal power unit g predicted by the unit combination start-up and shutdown variable prediction model; x g denoted by , represents the optimal start-up and shutdown scheme for thermal power unit g after recovery; D(·) represents the distance metric function.

[0138] Additionally, it should be noted that this invention sets a confidence threshold. When the model's prediction result has a high confidence level and feasibility recovery is successfully completed, the corresponding variable is fixed; when the model's prediction result has a low confidence level, the variable remains unfixed and will be processed in subsequent optimization processes.

[0139]

[0140] in, This represents the predicted start-up and shutdown result of thermal power unit g at time t using the prediction model; x g This indicates the start-up and shutdown plan for thermal power units; δ represents the upper and lower limits of confidence, respectively.

[0141] Example 2

[0142] refer to Figure 3 The stochastic grouping accelerated solution system of the present invention includes:

[0143] The acquisition module is used to acquire the power system input characteristics for the current scheduling period;

[0144] The solution module is used to input the power system input features of the current scheduling cycle into the trained unit combination start-up and shutdown variable prediction model to obtain the initial solution of the unit start-up and shutdown state. The unit combination start-up and shutdown variable prediction model is trained according to the variable fixing mechanism of confidence threshold.

[0145] The recovery module is used to recover the initial solution of the unit start-up and shutdown state to obtain the optimal unit start-up and shutdown scheme.

[0146] This embodiment also includes:

[0147] The first building module is used to build the optimized solution database;

[0148] The second building module is used to build a predictive model for unit combination start-up and shutdown variables;

[0149] The training module is used to train the unit combination start-up and shutdown variable prediction model using the optimized solution database based on the variable fixing mechanism of confidence threshold, so as to obtain the trained unit combination start-up and shutdown variable prediction model.

[0150] In this embodiment, the first building module includes:

[0151] The first building unit is used to construct the randomized grouping problem;

[0152] A generation unit is used to generate several renewable energy uncertainty scenarios;

[0153] The solution unit is used to solve the stochastic unit combination problem under various renewable energy uncertainty scenarios, and obtain the unit start-up and shutdown schemes under various renewable energy uncertainty scenarios;

[0154] The second construction unit is used to construct an optimized solution database based on the unit start-up and shutdown schemes under the various renewable energy uncertainty scenarios.

[0155] In this embodiment, the objective function of the randomized grouping problem is:

[0156]

[0157] Where t, s, g and e represent the indexes of scheduling time step, scheduling scenario, thermal power unit and energy storage system, respectively; And ε represents the set of scheduling time steps, the set of scenarios, the set of thermal power units, and the set of energy storage systems, respectively; π s π represents the probability of scenario s occurring; s for u g,t and d g,t These represent whether thermal power unit g starts or stops at time step t; p g,t,s p represents the active power output of thermal power unit g in scenario s at time step t; e,t,s C represents the charging and discharging power of energy storage system e in scenario s at time step t; g (·) represents the power generation cost function of thermal power unit g; and C represents the start-up cost coefficient and shutdown cost coefficient of thermal power unit g, respectively; eESSs represents the charge and discharge cost coefficient of energy storage system e.

[0158] In this embodiment, the constraints of the stochastic unit combination problem include unit start-up and shutdown logic constraints, output upper and lower limits and ramping constraints, minimum start-up and shutdown time constraints, energy storage operation constraints, system supply and demand balance constraints, reserve constraints, and renewable energy acceptance constraints.

[0159] In this embodiment, the process of restoring the initial solution of the unit start-up and shutdown state to obtain the optimal unit start-up and shutdown scheme is as follows:

[0160] The initial solution of the unit start-up and shutdown state is restored by a feasibility recovery mechanism to obtain the optimal unit start-up and shutdown scheme.

[0161] The module division in this embodiment is illustrative and represents only one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional modules in each embodiment of this application can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0162] Example 3

[0163] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the accelerated solution method for stochastic unit combination, for example, including: acquiring power system input characteristics for the current scheduling period; inputting the power system input characteristics for the current scheduling period into a trained unit combination start-up and shutdown variable prediction model to obtain an initial solution for the unit start-up and shutdown state, wherein the unit combination start-up and shutdown variable prediction model is trained according to a variable fixing mechanism based on a confidence threshold; and recovering the initial solution for the unit start-up and shutdown state to obtain the optimal unit start-up and shutdown scheme. The memory may include main memory, such as high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device. The processor, network interface, and memory are interconnected via an internal bus, which may be an industry standard architecture bus, a peripheral component interconnection standard bus, an extended industry standard architecture bus, etc., and the bus may be divided into an address bus, a data bus, a control bus, etc. The memory is used to store the program; specifically, the program may include program code, and the program code includes computer operation instructions. Memory can include main memory and non-volatile memory, and provides instructions and data to the processor.

[0164] Example 4

[0165] A computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the accelerated solution method for stochastic unit combination. For example, the method includes: acquiring power system input characteristics for the current scheduling period; inputting the power system input characteristics for the current scheduling period into a trained unit combination start-up and shutdown variable prediction model to obtain an initial solution for the unit start-up and shutdown state, wherein the unit combination start-up and shutdown variable prediction model is trained according to a variable fixing mechanism based on a confidence threshold; and recovering the initial solution for the unit start-up and shutdown state to obtain the optimal unit start-up and shutdown scheme. Specifically, the computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. The volatile memory may include random access memory (RAM) and / or cache memory, etc. The non-volatile memory may include read-only memory (ROM), hard disk, flash memory, optical disk, magnetic disk, etc.

[0166] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0167] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0168] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0169] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0170] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and disclosure of the invention. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the following claims.

[0171] It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

[0172] The above description is merely a preferred embodiment of the present invention and does not constitute any limitation on the present invention. Any simple modifications, alterations, or equivalent structural changes made to the above embodiments based on the technical essence of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for accelerating the solution of random unit combinations, characterized in that, include: Obtain the power system input characteristics for the current scheduling period; The power system input features of the current scheduling cycle are input into the trained unit combination start-up and shutdown variable prediction model to obtain the initial solution of the unit start-up and shutdown state. The unit combination start-up and shutdown variable prediction model is trained according to the variable fixing mechanism of confidence threshold. The initial solution of the unit start-up and shutdown state is recovered to obtain the optimal unit start-up and shutdown scheme.

2. The method for accelerating the solution of randomized unit combinations according to claim 1, characterized in that, Before inputting the power system input features of the current scheduling period into the trained unit combination start-up and shutdown variable prediction model, the following steps are also included: Build an optimized solution database; Construct a predictive model for unit start-up and shutdown variables; Based on the confidence threshold variable fixing mechanism, the optimized solution database is used to train the unit combination start-up and shutdown variable prediction model to obtain the trained unit combination start-up and shutdown variable prediction model.

3. The method for accelerating the solution of randomized unit combinations according to claim 2, characterized in that, The process of constructing the optimized solution database is as follows: Constructing a randomized set combination problem; Several uncertainties in renewable energy scenarios are generated; Under various renewable energy uncertainty scenarios, the stochastic unit combination problem is solved to obtain the unit start-up and shutdown schemes under each renewable energy uncertainty scenario; An optimized solution database is constructed based on the unit start-up and shutdown schemes under the aforementioned uncertainties of renewable energy.

4. The method for accelerating the solution of randomized unit combinations according to claim 3, characterized in that, The objective function of the randomized grouping problem is: Where t, s, g and e represent the indexes of scheduling time step, scheduling scenario, thermal power unit and energy storage system, respectively; And ε represents the set of scheduling time steps, the set of scenarios, the set of thermal power units, and the set of energy storage systems, respectively; π s π represents the probability of scenario s occurring; s for u g,t and d g,t These represent whether thermal power unit g starts or stops at time step t; p g,t,s p represents the active power output of thermal power unit g in scenario s at time step t; e,t,s C represents the charging and discharging power of energy storage system e in scenario s at time step t; g (·) represents the power generation cost function of thermal power unit g; and C represents the start-up cost coefficient and shutdown cost coefficient of thermal power unit g, respectively; e This represents the charging and discharging cost coefficient of the energy storage system e.

5. The method for accelerating the solution of randomized unit combinations according to claim 3, characterized in that, The constraints of the stochastic unit combination problem include unit start-up and shutdown logic constraints, output upper and lower limits and ramping constraints, minimum start-up and shutdown time constraints, energy storage operation constraints, system supply and demand balance constraints, reserve constraints, and renewable energy acceptance constraints.

6. The method for accelerating the solution of randomized unit combinations according to claim 1, characterized in that, The process of restoring the initial solution of the unit start-up and shutdown state to obtain the optimal unit start-up and shutdown scheme is as follows: The initial solution of the unit start-up and shutdown state is restored by the feasibility recovery mechanism to obtain the optimal unit start-up and shutdown scheme.

7. A stochastic unit combination accelerated solution system, characterized in that, include: The acquisition module is used to acquire the power system input characteristics for the current scheduling period; The solution module is used to input the power system input features of the current scheduling cycle into the trained unit combination start-up and shutdown variable prediction model to obtain the initial solution of the unit start-up and shutdown state. The unit combination start-up and shutdown variable prediction model is trained according to the variable fixing mechanism of confidence threshold. The recovery module is used to recover the initial solution of the unit start-up and shutdown state to obtain the optimal unit start-up and shutdown scheme.

8. The stochastic unit combination accelerated solution system according to claim 7, characterized in that, Also includes: The first building module is used to build the optimized solution database; The second building module is used to build a predictive model for unit combination start-up and shutdown variables; The training module is used to train the unit combination start-up and shutdown variable prediction model using the optimized solution database based on the variable fixing mechanism of confidence threshold, so as to obtain the trained unit combination start-up and shutdown variable prediction model.

9. The stochastic unit combination accelerated solution system according to claim 8, characterized in that, The first building module includes: The first building unit is used to construct the randomized grouping problem; A generation unit is used to generate several renewable energy uncertainty scenarios; The solution unit is used to solve the stochastic unit combination problem under various renewable energy uncertainty scenarios, and obtain the unit start-up and shutdown schemes under various renewable energy uncertainty scenarios; The second construction unit is used to construct an optimized solution database based on the unit start-up and shutdown schemes under the various renewable energy uncertainty scenarios.

10. The stochastic unit combination accelerated solution system according to claim 9, characterized in that, The objective function of the randomized grouping problem is: Where t, s, g and e represent the indexes of scheduling time step, scheduling scenario, thermal power unit and energy storage system, respectively; And ε represents the set of scheduling time steps, the set of scenarios, the set of thermal power units, and the set of energy storage systems, respectively; π s π represents the probability of scenario s occurring; s for u g,t and d g,t These represent whether thermal power unit g starts or stops at time step t; p g,t,s p represents the active power output of thermal power unit g in scenario s at time step t; e,t,s C represents the charging and discharging power of energy storage system e in scenario s at time step t; g (·) represents the power generation cost function of thermal power unit g; and C represents the start-up cost coefficient and shutdown cost coefficient of thermal power unit g, respectively; e This represents the charging and discharging cost coefficient of the energy storage system e.

11. The stochastic unit combination accelerated solution system according to claim 9, characterized in that, The constraints of the stochastic unit combination problem include unit start-up and shutdown logic constraints, output upper and lower limits and ramping constraints, minimum start-up and shutdown time constraints, energy storage operation constraints, system supply and demand balance constraints, reserve constraints, and renewable energy acceptance constraints.

12. The stochastic unit combination accelerated solution system according to claim 7, characterized in that, The process of restoring the initial solution of the unit start-up and shutdown state to obtain the optimal unit start-up and shutdown scheme is as follows: The initial solution of the unit start-up and shutdown state is restored by the feasibility recovery mechanism to obtain the optimal unit start-up and shutdown scheme.

13. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the accelerated solution method for random grouping as described in any one of claims 1-6.

14. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the accelerated solution method for random grouping as described in any one of claims 1-6.

Images