Generator Set Control Method and Device Based on Multi-Time-Scale Characteristics

Abstract

This paper relates to a generator set control method and device based on multi-timescale features. The higher-level scale provides guidance information to the lower-level scale; the weekly plan provides the unit start-up and shutdown status for the day-ahead plan, and the day-ahead plan provides the basic output points of the units for the intraday plan. At the weekly scale, the constraints of unit timing coupling can be ignored, and a dataset is generated by feeding back from the lower-level scale to the higher-level scale. This results in smaller data dimensionality and lower computational cost for data generation. Then, based on a neural network, the start-up and shutdown status of the units is predicted, and reliable start-up and shutdown statuses are selected and passed down, eliminating the need for feasibility remedial measures for start-up and shutdown statuses and reducing computational complexity. At the day-ahead and intraday scales, the start-up and shutdown statuses of the units determined by the higher-level time scale are used to identify more redundant constraints and accelerate the solution speed. This paper utilizes the features of different time scales to solve the unit combination problem, which can accelerate the solution speed while ensuring solution quality, and the solution speed is more stable.

Description

Technical Field

[0001] The embodiments in this specification relate to the field of power system dispatching, and in particular to a generator set control method and device based on multi-time scale characteristics. Background Technology

[0002] Currently, the penetration rate of renewable energy in the power system is continuously increasing. However, renewable energy power generation is characterized by strong randomness, volatility, and intermittency, posing new challenges to traditional dispatch strategies. Existing methods are mostly based on day-ahead dispatch and cannot cope with the long-term volatility of renewable energy. In addition, most power generation currently relies on thermal power units, which have complex start-up procedures and long start-up and shutdown times. How to control both thermal power units and clean energy power units to meet electricity load demand is a pressing technical problem that needs to be solved. Summary of the Invention

[0003] To address the problems existing in the prior art, this specification provides a generator set control method and device based on multi-timescale features. For the long-term sub-hour interval generator set combination problem, a multi-timescale coordinated optimization method is proposed, involving weekly, day-ahead, and intraday intervals. The higher-level scale provides guidance information to the lower-level scale; the weekly plan provides the generator set start-up and shutdown status for the day-ahead plan, and the day-ahead plan provides the basic output points for the intraday plan. At the weekly scale, the constraints of generator set timing coupling can be ignored. A dataset is generated by feeding back from the lower-level scale to the higher-level scale, resulting in smaller data dimensionality and lower computational cost. Then, based on a neural network, the generator set start-up and shutdown status is predicted, and reliable start-up and shutdown are selected and passed down, eliminating the need for feasibility remediation measures for start-up and shutdown status, thus reducing computational complexity. At the day-ahead and intraday scales, the generator set start-up and shutdown status determined by the higher-level time scale can be used to identify more redundant constraints, further accelerating the solution speed. This method utilizes the features of different time scales to solve the generator set combination problem, which can accelerate the solution speed while ensuring solution quality, and the solution speed is more stable.

[0004] The specific technical solutions of the embodiments in this specification are as follows:

[0005] On one hand, embodiments of this specification provide a generator control method based on multi-time-scale characteristics, applied to a target power grid including thermal power unit nodes and clean energy generator unit nodes. The method includes:

[0006] Obtain the load demand and clean energy generation information of the clean energy generator nodes for each predicted period under the multi-level time scale of the target power grid within the target control time range;

[0007] For the first time scale in the multi-level time scale, the thermal power generation information of the thermal power unit node corresponding to the multiple prediction periods under the first time scale is calculated based on the load demand and clean energy power generation information corresponding to each prediction period under the first time scale.

[0008] Based on the thermal power generation information corresponding to multiple prediction periods under the first-level time scale, determine the thermal power generation information corresponding to multiple prediction periods under the second-level time scale.

[0009] Based on the thermal power generation information corresponding to multiple forecast periods under the second-level time scale, the load demand and clean energy power generation information for each forecast period under the second-level time scale, the thermal power generation information for multiple forecast periods is calculated.

[0010] Determine whether the second-level time scale is the last time scale in the multi-level time scale;

[0011] If not, the second-level time scale is used as the first-level time scale, and the step of determining the thermal power generation information corresponding to multiple prediction periods under the second-level time scale based on the thermal power generation information corresponding to multiple prediction periods under the first-level time scale is repeated.

[0012] If so, the thermal power generation information of the thermal power unit nodes in the target control time range is controlled using the thermal power generation information of multiple predicted time periods at the last time scale.

[0013] Furthermore, the multi-level time scale includes a weekly scale, a day-ahead scale, and an intraday scale arranged in sequence.

[0014] Furthermore, the thermal power generation information includes the start-up and shutdown status and basic output points of the thermal power unit nodes;

[0015] The calculation of thermal power generation information for the thermal power unit nodes corresponding to multiple prediction periods under the first-level time scale, based on the load demand and clean energy power generation information for each prediction period under this first-level time scale, further includes:

[0016] The load demand and clean energy power generation information corresponding to each prediction period under the weekly scale are input into the pre-trained model for calculation to obtain the start-up and shutdown status scores of each thermal power unit node under each prediction period.

[0017] If the start-stop state score exceeds a threshold, the start-stop state of the thermal power unit node corresponding to the start-stop state score during the corresponding prediction period is "started". If not, it is determined whether the start-stop state score is less than the difference between 1 and the threshold. If so, the start-stop state of the thermal power unit node corresponding to the start-stop state score during the corresponding prediction period is "stopped".

[0018] Furthermore, the steps for training the model include:

[0019] Obtain the load demand and clean energy generation information of the target power grid for each historical period on a weekly scale within the historical control time range, and use it as training samples. Where X t N represents the training sample corresponding to the t-th historical time period on a weekly scale. D Indicates the total load. This represents the Nth historical period corresponding to the t-th historical period on a weekly scale. D The load requirement of N loads w This indicates the total number of clean energy generator sets. This represents the Nth historical period corresponding to the t-th historical period on a weekly scale. w The output power of a clean energy generator set;

[0020] Obtain the historical start-up and shutdown status of each thermal power unit node in each historical period within the intraday scale of the historical control time range;

[0021] Based on the proportional relationship between the intraday time granularity and the weekly time granularity, the historical start-up and shutdown status of each thermal power unit node in each historical period at the intraday scale is determined according to the formula. The data is aggregated to obtain the historical start-up and shutdown status scores of each thermal power unit node in each historical period at the weekly scale, which are then used as labels for the corresponding training samples. in, t represents the historical start-up and shutdown status score of the i-th thermal power unit node in the m-th historical period on a weekly scale. r t represents the time granularity on an intraday scale. w Indicates a time granularity on a weekly scale. Ω represents the historical start-up and shutdown status of the i-th thermal power unit node in the n-th historical period on an intraday scale. m ={n|t r,n ∈t w,m m=1,2,...,T W This represents the m-th time period t belonging to the week-scale. w,m The nth time period t on the intraday scale r,n The quantity, T W N represents the total number of historical periods on a weekly scale. G This represents the total number of nodes in a thermal power unit. This represents the Nth historical period corresponding to the t-th historical period on a weekly scale. G The start / stop status score of each thermal power unit node;

[0022] The training samples and labels of each historical period within the weekly scale of the historical control time range are input into the model for training, resulting in the trained model.

[0023] Furthermore, the formula for determining the thermal power generation information corresponding to multiple prediction periods at the second time scale based on the thermal power generation information corresponding to multiple prediction periods at the first time scale is as follows:

[0024]

[0025] in, This indicates the start-up / shutdown status of the i-th thermal power unit node in the n-th forecast period of the current day. N represents the start-up / shutdown status of the i-th thermal power unit node in the m-th prediction period on a weekly scale. G The total number of nodes in thermal power units, t d,n t represents the nth forecast period on a day-ahead scale. w,m This represents the m-th prediction period on a weekly scale.

[0026] Furthermore, the calculation of the thermal power generation information for multiple forecast periods based on the thermal power generation information corresponding to multiple forecast periods under the second-level time scale, the load demand for each forecast period under the second-level time scale, and the clean energy power generation information further includes:

[0027] The start-up and shutdown status of each thermal power unit node corresponding to multiple forecast periods under the day-ahead scale, the load demand and clean energy power generation information of each forecast period under the day-ahead scale, and the constraints corresponding to the day-ahead scale are input into a commercial solver for solving, so as to obtain the start-up and shutdown status of each thermal power unit corresponding to the remaining forecast periods under the day-ahead scale and the output of each thermal power unit in all forecast periods of the day-ahead scale.

[0028] Furthermore, the formula for determining the thermal power generation information corresponding to multiple prediction periods at the second time scale based on the thermal power generation information corresponding to multiple prediction periods at the first time scale is as follows:

[0029]

[0030] in, This represents the basic output point of the i-th thermal power unit node in the n-th prediction period at the intraday scale. (·).start utilizes the initial value function in the Gurobi solver to pass the thermal power unit output, providing an initial heuristic solution for the decision variables, and then further optimization is performed based on this. t represents the output of the i-th thermal power unit node in the m-th forecast period at the day-ahead scale. r,n This represents the nth forecast period on an intraday scale.

[0031] Furthermore, the calculation of the thermal power generation information for multiple forecast periods based on the thermal power generation information corresponding to multiple forecast periods under the second-level time scale, the load demand for each forecast period under the second-level time scale, and the clean energy power generation information further includes:

[0032] The basic output points of each thermal power unit node in each forecast period at the intraday scale, the load demand and clean energy power generation information in each forecast period at the intraday scale, and the corresponding constraints at the intraday scale are input into a commercial solver for solving to obtain the output of each thermal power unit node in each forecast period at the intraday scale.

[0033] Furthermore, after determining the thermal power generation information corresponding to multiple prediction periods at the second time scale based on the thermal power generation information corresponding to multiple prediction periods at the first time scale, the method further includes:

[0034] Based on the thermal power generation information corresponding to multiple prediction periods under the first-level time scale and / or the thermal power generation information corresponding to multiple prediction periods under the second-level time scale, redundant constraints of the second-level time scale are removed.

[0035] On the other hand, embodiments of this specification also provide a generator control device based on multi-time-scale characteristics, deployed in a target power grid including thermal power unit nodes and clean energy generator unit nodes, the device comprising:

[0036] The information acquisition unit is used to acquire the load demand and clean energy generation information of the clean energy generator nodes corresponding to each predicted period under the multi-level time scale of the target power grid within the target control time range.

[0037] The first-level time scale calculation unit is used to calculate the thermal power generation information of the thermal power unit node corresponding to multiple prediction periods under the first-level time scale, based on the load demand and clean energy power generation information corresponding to each prediction period under the first-level time scale.

[0038] The information transmission unit is used to determine the thermal power generation information corresponding to multiple prediction periods under the second time scale based on the thermal power generation information corresponding to multiple prediction periods under the first time scale.

[0039] The second-level time scale calculation unit is used to calculate the thermal power generation information for multiple prediction periods based on the thermal power generation information corresponding to multiple prediction periods under the second-level time scale, the load demand and clean energy power generation information for each prediction period under the second-level time scale.

[0040] An iterative calculation unit is used to determine whether the second-level time scale is the last time scale among the multi-level time scales; if not, the second-level time scale is used as the first-level time scale, and the step of determining the thermal power generation information corresponding to multiple prediction periods under the second-level time scale based on the thermal power generation information corresponding to multiple prediction periods under the first-level time scale is repeated.

[0041] The control unit is used to control the thermal power unit node within the target control time range by using the thermal power generation information of multiple predicted time periods of the last time scale when the judgment result of the iterative calculation unit is yes.

[0042] Using the embodiments in this specification, the long-term unit combination problem is decomposed into multi-level time scales for the target control time range. Each time scale is divided into multiple prediction time periods. The load demand and clean energy generation information of the clean energy generating unit nodes corresponding to each prediction time period of the target power grid at each time scale are obtained. Then, the thermal power generation information corresponding to each time scale is calculated sequentially from coarse-grained to fine-grained according to the time scale level. Simultaneously, the higher-level time scale provides guidance information for the lower-level time scale. Based on the thermal power generation information corresponding to multiple prediction time periods at the higher-level time scale, the thermal power generation information corresponding to multiple prediction time periods at the lower-level time scale is determined. This is used as a basis for calculating the thermal power generation information for multiple prediction time periods at the lower-level time scale until the thermal power generation information for multiple prediction time periods at the last time scale is calculated, obtaining the finest-grained thermal power generation information. Therefore, the thermal power generation information of the multiple prediction time periods at the last-level time scale is used to control the thermal power generating unit nodes within the target control time range, improving the control effect of the thermal power generating unit nodes and thus increasing resource utilization. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 The figure shown is a schematic diagram of an implementation system for a generator set control method based on multi-time-scale features in an embodiment of this specification.

[0045] Figure 2 The diagram shown is a flowchart of a generator set control method based on multi-time-scale features according to an embodiment of this specification.

[0046] Figure 3 The diagram shown is a flowchart illustrating the process of calculating the thermal power generation information of the thermal power unit nodes corresponding to multiple prediction periods under the first-level time scale based on the load demand and clean energy power generation information corresponding to each prediction period under the first-level time scale in an embodiment of this specification.

[0047] Figure 4 The diagram shown is a schematic diagram of weekday and dayday time periods in the embodiments of this specification;

[0048] Figure 5 The diagram shown is a schematic representation of a generator set control device based on multi-time-scale features in an embodiment of this specification.

[0049] Figure 6 The diagram shown is a structural schematic of the computer device in an embodiment of this specification.

[0050] [Explanation of Figure Markers]:

[0051] 101. Terminal;

[0052] 102. Server;

[0053] 501. Information Acquisition Unit;

[0054] 502. First-level time-scale calculation unit;

[0055] 503. Information transmission unit;

[0056] 504. Second-level time-scale calculation unit;

[0057] 505. Iterative calculation unit;

[0058] 506. Control unit;

[0059] 602. Computer equipment;

[0060] 604, Processor;

[0061] 606. Memory;

[0062] 608. Drive mechanism;

[0063] 610. Input / output module;

[0064] 612. Input devices;

[0065] 614. Output devices;

[0066] 616. Presentation equipment;

[0067] 618. Graphical User Interface;

[0068] 620. Network interface;

[0069] 622. Communication link;

[0070] 624. Communication bus. Detailed Implementation

[0071] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the embodiments of this specification, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the embodiments of this specification.

[0072] It should be noted that the terms "first," "second," etc., in the description, claims, and accompanying drawings of the embodiments herein are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, apparatus, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0073] It should be noted that the acquisition, storage, use, and processing of data in the technical solutions of the embodiments of this specification all comply with the relevant provisions of national laws and regulations.

[0074] It should be noted that in the embodiments of this specification, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, they do not mean that the applicant has used or necessarily used the solution.

[0075] like Figure 1 The diagram illustrates an implementation system of a generator set control method based on multi-timescale features, as described in this specification. The system includes a terminal 101 and a server 102. The terminal 101 and server 102 can communicate via a network, which may include a local area network (LAN), a wide area network (WAN), the Internet, or a combination thereof, and be connected to a website, user equipment (e.g., a computing device), and a backend system.

[0076] Users can input the structure of the target power grid, load demand, and clean energy generation information in the target power grid into the server 102 through the terminal 101. The server 102 uses a stored commercial solver to calculate the load demand, clean energy generation information, and structure of the target power grid to obtain the thermal power generation information of the thermal power unit nodes in the target power grid. Then, the thermal power generation information is provided to the user through the terminal 101 so that the user can control the thermal power nodes according to the provided thermal power generation information.

[0077] Alternatively, server 102 may be a node of a cloud computing system (not shown in the figure), or each server may be a separate cloud computing system comprising multiple computers interconnected by a network and operating as a distributed processing system.

[0078] In addition, it should be noted that, Figure 1 The examples shown are merely one application environment provided by the embodiments in this specification. In practical applications, other application environments may also be included, and this specification does not impose any limitations.

[0079] To address the problems existing in the prior art, this specification provides a generator set control method based on multi-timescale features. For the long-term sub-hour interval generator set combination problem, a multi-timescale coordinated optimization method is proposed, involving weekly, day-ahead, and intraday intervals. The higher-level scale provides guidance information to the lower-level scale; the weekly plan provides the generator set start-up and shutdown status for the day-ahead plan, and the day-ahead plan provides the basic output points for the intraday plan. At the weekly scale, a dataset is generated by feeding back from the lower-level scale to the higher-level scale, resulting in smaller data dimensionality and lower computational cost. Then, a neural network is used to predict the generator set start-up and shutdown status, selecting reliable start-up and shutdown states for downward propagation, eliminating the need for feasibility remediation measures for start-up and shutdown statuses and reducing computational complexity. At the day-ahead and intraday scales, the generator set start-up and shutdown status determined by the higher-level time scale can be used to identify more redundant constraints, further accelerating the solution speed. This method utilizes the features of different time scales to solve the generator set combination problem, accelerating the solution speed while ensuring solution quality, and achieving more stable solution speed.

[0080] Figure 2 The diagram shows a schematic flow chart of a generator set control method based on multi-timescale features in an embodiment of this specification. This diagram illustrates the process of calculating the thermal power generation information of thermal power generation nodes in the target power grid and using this calculated information for control. The order of steps listed in the embodiment is merely one possible execution order among many and does not represent the only possible execution order. In actual system or device products, the methods shown in the embodiments or accompanying drawings can be executed sequentially or in parallel.

[0081] Specifically, the methods described in the embodiments of this specification can be applied to a target power grid that includes thermal power unit nodes and clean energy power unit nodes, such as... Figure 2 As shown, the method may include:

[0082] Step 201: Obtain the load demand and clean energy generation information of the clean energy generator nodes for each predicted period under the multi-level time scale of the target power grid within the target control time range;

[0083] Step 202: For the first time scale in the multi-level time scale, calculate the thermal power generation information of the thermal power unit node corresponding to the multiple prediction periods under the first time scale based on the load demand and clean energy power generation information corresponding to each prediction period under the first time scale.

[0084] Step 203: Determine the thermal power generation information corresponding to multiple prediction periods under the second time scale based on the thermal power generation information corresponding to multiple prediction periods under the first time scale;

[0085] Step 204: Calculate the thermal power generation information for multiple forecast periods based on the thermal power generation information corresponding to multiple forecast periods under the second-level time scale, the load demand and clean energy power generation information for each forecast period under the second-level time scale;

[0086] Step 205: Determine whether the second-level time scale is the last time scale in the multi-level time scale;

[0087] Step 206: If not, then use the second-level time scale as the first-level time scale, and repeat the step of determining the thermal power generation information corresponding to multiple prediction periods under the second-level time scale based on the thermal power generation information corresponding to multiple prediction periods under the first-level time scale.

[0088] Step 207: If so, the thermal power generation information of the thermal power unit nodes in the target control time range is controlled using the thermal power generation information of multiple predicted time periods of the last time scale.

[0089] Using the embodiments in this specification, the long-term unit combination problem is decomposed into multi-level time scales for the target control time range. Each time scale is divided into multiple prediction time periods. The load demand and clean energy generation information of the clean energy generating unit nodes corresponding to each prediction time period of the target power grid at each time scale are obtained. Then, the thermal power generation information corresponding to each time scale is calculated sequentially from coarse-grained to fine-grained according to the time scale level. Simultaneously, the higher-level time scale provides guidance information for the lower-level time scale. Based on the thermal power generation information corresponding to multiple prediction time periods at the higher-level time scale, the thermal power generation information corresponding to multiple prediction time periods at the lower-level time scale is determined. This is used as a basis for calculating the thermal power generation information for multiple prediction time periods at the lower-level time scale until the thermal power generation information for multiple prediction time periods at the last time scale is calculated, obtaining the finest-grained thermal power generation information. Therefore, the thermal power generation information of the multiple prediction time periods at the last-level time scale is used to control the thermal power generating unit nodes within the target control time range, improving the control effect of the thermal power generating unit nodes and thus increasing resource utilization.

[0090] In the embodiments described in this specification, the Unit Commitment (UC) problem is a mixed-integer optimization problem, involving a large number of integer and continuous decision variables, as well as numerous equality and inequality constraints. The computational complexity of the UC model increases exponentially with the time scale. Simply applying existing short-term methods to long-term problems significantly increases the solution time.

[0091] Currently, research on unit coordination problems mainly focuses on short-term day-ahead scales, with limited research on long-term scales. This is primarily because the number of decision variables and constraints increases significantly with the increase in scheduling cycles, leading to an exponential increase in the computational complexity of UC models. Existing methods for solving UC problems include Lagrange relaxation, mixed-integer linear programming, dynamic programming, Benders decomposition, heuristic algorithms, and data-driven approaches. Industrial-scale UC problems are typically solved using commercial solvers based on the branch and cut (B&C) framework, such as CPLEX or Gurobi. These methods accelerate the solution process through problem decomposition, reduction of decision variables, and reduction of redundant constraints.

[0092] Problem decomposition methods break down large-scale problems into easier subproblems. Existing methods primarily focus on spatial decomposition, such as using Lagrange relaxation to relax system-level constraints, decoupling large-scale UC problems into subproblems of individual units. However, research on time-scale decomposition is limited. Decision variable reduction methods mainly refer to reducing the number of integer variables, i.e., determining the start-up and shutdown states of units, to reduce computational complexity. Traditional methods are often heuristic, such as determining the start-up and shutdown states of units in the primal problem based on the integer solutions of the dual problem. In recent years, with the development of artificial intelligence, methods have emerged that determine integer variables based on historical power system dispatch data, such as using machine learning to predict corresponding unit start-up and shutdown based on daily load demand. The constraint directly related to the start-up and shutdown states of units is the minimum start-up and shutdown time constraint. Existing decision variable reduction methods cannot guarantee that the determined unit start-up and shutdown will satisfy the minimum start-up and shutdown time constraint, requiring feasibility remediation measures, which increases computational complexity. For data-driven decision variable reduction methods, a single data point is typically defined as the load demand across all time periods within a scheduling cycle and its corresponding generator start-up and shutdown. Solving a UC problem within a scheduling cycle yields only one data point, resulting in high computational costs for generating the data. Redundancy constraint reduction methods aim to accelerate the solution by reducing constraints that do not affect the feasible region of the problem. UC problems contain numerous redundant constraints, and existing redundancy constraint reduction methods primarily address them at a single scale, such as identifying redundant power flow safety constraints at the day-ahead scale by solving auxiliary problems. Utilizing guidance information from other time scales would allow for the identification of even more redundant constraints, further accelerating the solution process.

[0093] Therefore, the method in the embodiments of this specification decomposes the long-term UC problem into multiple time scales, using the higher time scale to provide forward guidance for the lower time scale.

[0094] In the embodiments described in this specification, the target power grid includes thermal power units and clean energy power generation units, and the clean energy power generation units can be wind power generation, etc.

[0095] In the embodiments of this specification, the load demand of the target power grid can be learned in advance using a machine learning model. When it is necessary to solve the unit combination problem within the target control time range, the trained machine learning model is first used to predict the load demand corresponding to each prediction period under the multi-level time scale within the target control time range.

[0096] In addition, the power generation of wind turbines can be learned based on environmental information. When it is necessary to solve the unit combination problem within the target control time range, the environmental prediction information corresponding to the target control time range is input, thereby calculating the wind power generation information corresponding to each prediction period under the multi-level time scale of the target control time range.

[0097] In the embodiments of this specification, thermal power generation information may include the start-up and shutdown status of thermal power unit nodes and power generation capacity.

[0098] According to one embodiment of this specification, the multi-level time scale includes a weekly scale, a day-ahead scale, and an intraday scale arranged in sequence.

[0099] The method described in this specification decomposes long-term UC problems (such as annual and quarterly problems) into weekly units, establishing a multi-timescale coordinated optimization model that integrates week-day-intraday time scales. The time window is selected based on electricity demand, which is also based on a weekly cycle. Furthermore, by utilizing wind power information and other meteorological information, the accuracy of wind power generation predictions can reach 90% week in advance, effectively considering the volatility of new energy power generation.

[0100] Preferably, the weekly scale uses a one-week scheduling cycle, and the prediction period t at the weekly scale is... w Take the maximum value of the minimum start-stop time of all thermal power unit nodes. The start-stop state of the thermal power unit nodes obtained in this way will definitely meet the minimum start-stop time constraint. This means that the thermal power unit nodes can change their start-stop status in each time period without the need for feasibility repair for the minimum start-stop time constraint.

[0101] The day-ahead scale uses a one-day scheduling cycle, and the forecast period t under the day-ahead scale is... d Take the minimum start-up and shutdown time of all thermal power unit nodes, so that at t d During this period, the thermal power units will not be started or stopped, which will minimize the deviation between the day-ahead plan and the intraday plan for unit start-up and shutdown, and make the transmission of unit output to the downstream more reliable.

[0102] The intraday timescale uses a 4-hour scheduling cycle, and the forecast period t under the intraday timescale is... r Typically, the intervals are sub-hourly, such as 15 minutes.

[0103] like Figure 3 As shown, calculating the thermal power generation information of the thermal power unit nodes corresponding to multiple prediction periods under the first-level time scale based on the load demand and clean energy power generation information for each prediction period under the first-level time scale further includes:

[0104] Step 301: Input the load demand and clean energy power generation information corresponding to each prediction period under the weekly scale into the pre-trained model for calculation to obtain the start-stop status score of each thermal power unit node under each prediction period.

[0105] Step 302: Determine whether the start / stop status score exceeds the threshold;

[0106] Step 303: If yes, then the start-stop state of the thermal power unit node corresponding to the start-stop state score is start-up during the corresponding predicted time period;

[0107] Step 304: If not, determine whether the start / stop state score is less than the difference between 1 and the threshold;

[0108] Step 3041: If yes, then the start-stop state of the thermal power unit node corresponding to the start-stop state score is stopped during the corresponding predicted time period.

[0109] In the embodiments of this specification, the steps for training the model include:

[0110] Step 1: Obtain the load demand and clean energy generation information of the target power grid for each historical period on a weekly scale within the historical control time range, and use it as training samples. Where X t N represents the training sample corresponding to the t-th historical time period on a weekly scale. D Indicates the total load. This represents the Nth historical period corresponding to the t-th historical period on a weekly scale. D The load requirement of N loads w This indicates the total number of clean energy generator sets. This represents the Nth historical period corresponding to the t-th historical period on a weekly scale. w The output power of a clean energy generator set;

[0111] Step 2: Obtain the historical start-up and shutdown status of each thermal power unit node in each historical period within the intraday scale of the historical control time range;

[0112] Step 3: Based on the ratio between the intraday time granularity and the weekly time granularity, aggregate the historical start-up and shutdown status of each thermal power unit node in each historical period at the intraday scale according to formula (1) to obtain the historical start-up and shutdown status score of each thermal power unit node in each historical period at the weekly scale, which serves as the label for the corresponding training sample.

[0113]

[0114] in, t represents the historical start-up and shutdown status score of the i-th thermal power unit node in the m-th historical period on a weekly scale. r t represents the time granularity on an intraday scale. w Indicates a time granularity on a weekly scale. Ω represents the historical start-up and shutdown status of the i-th thermal power unit node in the n-th historical period on an intraday scale. m ={n|t r,n ∈t w,mm=1,2,...,T W This represents the m-th time period t belonging to the week-scale. w,m The nth time period t on the intraday scale r,n The quantity, T W N represents the total number of historical periods on a weekly scale. G This represents the total number of nodes in a thermal power unit. This represents the Nth historical period corresponding to the t-th historical period on a weekly scale. G The start / stop status score of each thermal power unit node;

[0115] Step 4: Input the training samples and labels of each historical period on a weekly scale within the historical control time range into the model for training, and obtain the trained model.

[0116] In the embodiments described in this specification, the load demand and wind power generation information X t Training samples X can be obtained from publicly available datasets from some independent power system operators or generated through Monte Carlo simulations. t The labels need to be obtained by solving the UC problem. Note that if the start-up and shutdown states of the thermal power unit nodes are obtained by directly solving the weekly scale model, they cannot be used as labels because these shutdown states will have large deviations or even no solution on small time scales.

[0117] Therefore, the embodiments in this specification solve the lowest-level time-scale UC problem using a commercial solver, and obtain the label of each training sample by feeding back from the lower-level scale to the higher-level scale.

[0118] Here, we introduce soft labels. The start / stop state scores obtained here are not integer variables of 0 or 1, but rather scores between 0 and 1. For each training sample, the soft label is a vector. Each dimension can be viewed as the probability of the corresponding unit starting up during that period. Using soft tags can solve the problem of not being able to obtain the correct start-up and shutdown status of the weekly plan, while providing richer and more continuous information and conveying more knowledge.

[0119] Since a single data point is defined as information within a specific time period, the data dimensionality is relatively small, and a simple fully connected neural network is sufficient for learning. The network learns the mapping from load demand and wind power generation information within a specific time period to the start-up and shutdown status scores for that period.

[0120]

[0121] The neural network structure includes: 1. Input layer: receives load and wind power generation information and transmits it to the network. 2. Hidden layer: performs feature extraction and representation learning on the input data, capturing complex patterns in the data by combining linear transformations and non-linear activation functions. 3. Output layer: receives the features extracted by the hidden layer and outputs the probability of each thermal power unit node starting using the sigmoid function. The model uses the binary cross-entropy loss function (BCE Loss). The loss function is calculated by averaging the BCE Loss for each dimension of the label, and training is performed through backpropagation. The model parameters are adjusted by minimizing the loss function.

[0122] During prediction, the model outputs a score of U for the predicted start-up and shutdown status of all thermal power unit nodes for a certain time period. t For the predicted start-up and shutdown status of thermal power unit node i in time period t on a weekly scale Set a threshold ξ to filter unit start-up and shutdown, and the filtering criteria are as shown in (3). If Then the start-up and shutdown status of thermal power unit node i in time period t on a weekly scale. Set to 1, if Then the start-up and shutdown status of thermal power unit node i in time period t on a weekly scale. Setting it to 0 indicates that, for other cases, the start-up and shutdown status of the thermal power unit nodes is undetermined and needs to be further determined on a day-ahead scale.

[0123]

[0124] This step utilizes historical operating information of the power system and uses a neural network to predict the start-up and shutdown of thermal power unit nodes, replacing the solver to solve the periodic-scale model, thus accelerating the online solution speed and making the solution speed more stable.

[0125] Due to the carefully selected periodic model time period t w Value (forecast period t on a weekly scale) w By taking the maximum value of the minimum start-stop time of all thermal power unit nodes, the start-stop state of the obtained thermal power unit nodes will definitely satisfy the minimum start-stop time constraint, eliminating the need for feasibility repair for the minimum start-stop time constraint and reducing computational complexity.

[0126] Then the established start / stop status is transmitted to the daytime scale.

[0127] Specifically, the formula for determining the thermal power generation information corresponding to multiple prediction periods at the second time scale based on the thermal power generation information corresponding to multiple prediction periods at the first time scale is as follows:

[0128]

[0129] in, This indicates the start-up / shutdown status of the i-th thermal power unit node in the n-th forecast period of the current day. N represents the start-up / shutdown status of the i-th thermal power unit node in the m-th prediction period on a weekly scale. G The total number of nodes in thermal power units, t d,n t represents the nth forecast period on a day-ahead scale. w,m This represents the m-th prediction period on a weekly scale.

[0130] This can be understood as follows: if the nth time period on the day-ahead scale belongs to the mth time period on the week-ahead scale, then the start / stop status of the nth time period on the day-ahead scale is equal to the start / stop status of the mth time period on the week-ahead scale.

[0131] Then, calculations are performed on the day-ahead scale. Specifically, the start-up and shutdown status of each thermal power unit node corresponding to multiple forecast periods under the day-ahead scale, the load demand and clean energy power generation information of each forecast period under the day-ahead scale, and the constraints corresponding to the day-ahead scale are input into a commercial solver for solving, so as to obtain the start-up and shutdown status of each thermal power unit corresponding to the remaining forecast periods under the day-ahead scale and the output of each thermal power unit in all forecast periods of the day-ahead scale.

[0132] In the embodiments of this specification, the optimization objective of a single UC problem is Equation (5):

[0133]

[0134] Where T is the total number of scheduling periods, u i,t C represents the start / stop status of thermal power unit node i during time period t. i (·) represents the fuel cost of thermal power unit node i, which can represent how much carbon is emitted when thermal power unit node i is operating at a specified power, that is, the carbon emissions. v represents the output power of node i of the thermal power unit during time period t. i,t This indicates whether node i of the thermal power unit performs a start-up operation (1) or some other operation (0) during time period t. This represents the startup cost of thermal power unit node i, which indicates the amount of carbon emitted during the transition of generator unit node i from a shutdown state to a normal operating state; it is also the carbon emission amount. C w (·) represents the cost of wind curtailment at the wind turbine node. This represents the wind curtailment power of wind turbine node j during time period t.

[0135] The following constraints need to be met:

[0136]

[0137] in, P represents the output power of wind turbine node j during time period t. Dt F represents the total load demand of the target power grid during time period t, where T represents the total number of scheduling periods. l This represents the transmission capacity of line l in the target power grid. and D is the power distribution transfer factor for thermal power unit node i and load k with respect to line l. k,t This represents the demand of load k during time period t, where L represents the total number of transmission lines, and N represents the demand of load k during time period t. D Indicates the total load. and Let represent the lower and upper power limits of node i in a thermal power unit, respectively. These represent the maximum increase and the maximum decrease in power at node i of the thermal power unit when it is in the on state, respectively. T represents the maximum increase and decrease in power at node i of the thermal power unit during start-up and shutdown transitions, respectively. i U ,T i D w represents the minimum start-up and shutdown hold times for node i of a thermal power unit, respectively. i,t Indicates whether node i of the thermal power unit performs a shutdown operation (1) or other operation (0) during time period t.

[0138] The optimization objective is to determine the start-up, shutdown, output, and wind curtailment power of thermal power unit nodes and wind turbine nodes, so as to minimize the system's operating cost (carbon emissions). The above-mentioned constraint (6) is a power balance constraint, constraint (7) is a power transmission line flow safety constraint, constraint (8) is a thermal power unit node output constraint, constraints (9) and (10) are thermal power unit node output ramp-up constraints, constraints (11) and (12) are minimum start-up and shutdown time constraints for thermal power unit nodes, and constraints (13) and (14) are logical constraints between the start-up and shutdown states of thermal power unit nodes and their start-up and shutdown actions.

[0139] The weekly scale uses a one-week scheduling cycle, with t as the unit of time. w For a given time period, based on weekly forecasts of wind power and load demand, the start-up and shutdown status of most thermal power unit nodes is calculated using a model, and then the start-up and shutdown status is transmitted to the day-ahead scale.

[0140] The optimization objective at the weekly scale is (15):

[0141]

[0142] Among them, T w Let be the total number of scheduling periods on a weekly scale. The constraints that need to be satisfied are (6)-(14).

[0143] The current scale uses a one-day scheduling cycle, in t. d For a given time period, the start-up and shutdown status of each thermal power unit node corresponding to multiple forecast periods at the day-ahead scale, the load demand and clean energy power generation information for each forecast period at the day-ahead scale, and the constraints corresponding to the day-ahead scale are input into a commercial solver for solving, to obtain the start-up and shutdown status of each thermal power unit corresponding to the remaining forecast periods at the day-ahead scale, as well as the output of each thermal power unit in all forecast periods at that day-ahead scale.

[0144] The optimization objective at the current scale is (16), and the constraints are (6)-(14):

[0145]

[0146] Among them, T D This represents the total number of scheduling periods on a day-ahead scale.

[0147] If commercial solvers or other existing methods are used to solve the start-stop status of the weekly plan, and all of them are directly passed down, the deviation is large or even there is no solution in most cases. Therefore, in the embodiments of this specification, after determining the thermal power generation information corresponding to multiple prediction periods under the second time scale based on the thermal power generation information corresponding to multiple prediction periods under the first time scale, the method further includes: removing redundant constraints of the second time scale based on the thermal power generation information corresponding to multiple prediction periods under the first time scale and / or the thermal power generation information corresponding to multiple prediction periods under the second time scale.

[0148] Once the start / stop status is determined at the higher time scale, some related constraints in the lower time scale problem become redundant. This specification's embodiments improve upon existing methods for identifying redundant power flow safety constraints by proposing a method for identifying redundant minimum start / stop time constraints, thereby accelerating the problem-solving process.

[0149] First, the existing method for identifying redundancy power flow safety constraints involves continuously relaxing a mathematical programming problem. Finally, it can be directly determined whether the problem is redundant through (17) and (18), thus avoiding the need to solve the mathematical programming problem.

[0150]

[0151] Where j1, j2, ..., j NG It is 1, 2, ..., N G A rearrangement such that:

[0152] z l,t It is the line capacity F l During the extension of time period t,

[0153] This method is a sufficient condition; in the inequality, the minimum output of node i of the thermal power unit in each time period is 0, and the maximum output is... However, when performing calculations on a circumferential scale, the start-up and shutdown states of some thermal power unit nodes have already been determined, so the output range of the thermal power unit nodes can be dynamically adjusted:

[0154]

[0155] in, This represents the minimum output of thermal power unit node i during time period t. This represents the maximum output of thermal power unit node i during time period t.

[0156] Formulas (17) and (18) are modified by replacing the output range of the thermal power unit node with the dynamic output range (19), and the original method is extended to the connection of wind turbine unit nodes, i.e. in, For the power distribution transfer factor of wind turbine node j with respect to line l, the modified criteria (20) and (21) can identify more redundant constraints:

[0157]

[0158] Based on the established start and stop states at the week-scale, it is also possible to identify the redundant minimum start and stop time constraints of time-coupled timing at the day-ahead scale.

[0159] For node i of a thermal power unit, assuming t in the periodic scale w,m and t w,n The start and stop states for the two predicted time periods have been determined, time period [t] w,m+1 ,t w,n-1 The start-up and shutdown status of the units within the specified range remains undetermined. (Note: t) w,n The forecast period begins with the period number t on a day-ahead scale. * , t w,m The forecast period ends with the period numbered on a day-ahead scale, denoted as t. 0 ,like Figure 4 As shown.

[0160] First t w,m and t w,n The start-up and shutdown statuses for these two prediction periods have been determined. In these two prediction periods, the minimum start-up and shutdown time constraints (11) and (12) are obviously redundant. That is, the minimum start-up and shutdown time constraints corresponding to the thermal power unit nodes with determined start-up and shutdown statuses and the corresponding prediction periods are removed.

[0161] For the predicted time period [t] * -T iU +1,t * -1], where T i U This represents the minimum start-up hold time for node i of a thermal power unit, when u i,t* =0 (i.e., thermal power unit node i at t) * (The predicted time period is during a shutdown state). Since this predicted time period is less than the minimum start-up time of unit i, no thermal power unit node start-up (u) will occur within this predicted time period. i,t The case where it changes from 0 to 1, otherwise in t * During the predicted time period [t], the thermal power unit nodes should also be in the operating state. Therefore, for the predicted time period [t] * -T i U +1,t * -1], the minimum startup time constraint (11) coupled across multiple prediction periods becomes:

[0162]

[0163] For a predicted period whose start / stop status has been determined to be stopped, the minimum downtime constraint (12) for that predicted period is redundant and can be deleted.

[0164] Similarly, for the prediction period [t], * -T i U +1,t * -1], when u i,t* =1 (i.e., thermal power unit node i at t) * (When the time period is the power-on state), the minimum downtime constraint (12) coupled across multiple time periods becomes:

[0165]

[0166] And the predicted time period [t] * -T i U +1,t * -1], when u i,t* When =1, the minimum startup time constraint (11) can be deleted.

[0167] After removing redundant constraints from the second-level time scale, the second-level time scale is then calculated.

[0168] By performing calculations at the day-ahead scale, we can obtain the start-up and shutdown status of each thermal power unit node corresponding to the remaining forecast period at the day-ahead scale, as well as the output of each thermal power unit node in all forecast periods at that day-ahead scale.

[0169] The formula for determining thermal power generation information for multiple forecast periods at the intraday scale based on thermal power generation information for multiple forecast periods at the day-ahead scale is as follows:

[0170]

[0171] in, This represents the basic output point of the i-th thermal power unit node in the n-th prediction period at the intraday scale. (·).start utilizes the initial value function in the Gurobi solver to pass the thermal power unit output, providing an initial heuristic solution for the decision variables, and then further optimization is performed based on this. t represents the output of the i-th thermal power unit node in the m-th forecast period at the day-ahead scale. r,n This represents the nth forecast period on an intraday scale.

[0172] Furthermore, based on the thermal power generation information corresponding to multiple forecast periods at an intraday scale, the load demand and clean energy power generation information for each forecast period at that intraday scale, the calculation of the thermal power generation information for multiple forecast periods further includes:

[0173] The basic output points of each thermal power unit node in each forecast period at the intraday scale, the load demand and clean energy power generation information in each forecast period at the intraday scale, and the corresponding constraints at the intraday scale are input into a commercial solver for solving to obtain the output of each thermal power unit node in each forecast period at the intraday scale.

[0174] In the embodiments described in this specification, the intraday scale uses a 4-hour scheduling cycle, and t r For a time period, t r Typically, the interval is sub-hourly, such as 15 minutes. Based on the basic output point of the unit provided on a day-ahead scale, and the ultra-short-term forecasts of wind power and load demand, the precise output of the thermal power unit is calculated. The optimization objective is:

[0175]

[0176] Among them, T R This represents the total number of scheduling periods on a daily scale. Since the start-stop status is completely determined at this point, there is no need to consider constraints (11) and (12).

[0177] Then, redundant constraints are removed, and the output of each thermal unit node in each forecast period at the intraday scale can be calculated using a commercial solver.

[0178] Finally, the output of each thermal unit node in each forecast period at the intraday scale is used to control each thermal unit node.

[0179] Based on the same inventive concept, embodiments of this specification also provide a generator set control device based on multi-time-scale characteristics, deployed in a target power grid including thermal power unit nodes and clean energy generator set nodes, such as... Figure 5 As shown, the device includes:

[0180] The information acquisition unit 501 is used to acquire the load demand and clean energy generation information of the clean energy generator nodes corresponding to each predicted period under the multi-level time scale of the target power grid within the target control time range.

[0181] The first-level time scale calculation unit 502 is used to calculate the thermal power generation information of the thermal power unit node corresponding to multiple prediction periods under the first-level time scale, based on the load demand and clean energy power generation information corresponding to each prediction period under the first-level time scale.

[0182] The information transmission unit 503 is used to determine the thermal power generation information corresponding to multiple prediction periods under the second time scale based on the thermal power generation information corresponding to multiple prediction periods under the first time scale.

[0183] The second-level time scale calculation unit 504 is used to calculate the thermal power generation information for multiple prediction periods based on the thermal power generation information corresponding to multiple prediction periods under the second-level time scale, the load demand and clean energy power generation information for each prediction period under the second-level time scale.

[0184] The iterative calculation unit 505 is used to determine whether the second-level time scale is the last time scale in the multi-level time scale; if not, the second-level time scale is used as the first-level time scale, and the step of determining the thermal power generation information corresponding to multiple prediction periods under the second-level time scale based on the thermal power generation information corresponding to multiple prediction periods under the first-level time scale is repeated.

[0185] The control unit 506 is used to control the thermal power unit node within the target control time range by using the thermal power generation information of multiple predicted time periods of the last time scale when the judgment result of the iterative calculation unit 505 is yes.

[0186] Furthermore, the multi-level time scale includes a weekly scale, a day-ahead scale, and an intraday scale arranged in sequence.

[0187] Furthermore, the thermal power generation information includes the start-up and shutdown status and basic output points of the thermal power unit nodes;

[0188] The calculation of thermal power generation information for the thermal power unit nodes corresponding to multiple prediction periods under the first-level time scale, based on the load demand and clean energy power generation information for each prediction period under this first-level time scale, further includes:

[0189] The load demand and clean energy power generation information corresponding to each prediction period under the weekly scale are input into the pre-trained model for calculation to obtain the start-up and shutdown status scores of each thermal power unit node under each prediction period.

[0190] If the start-stop state score exceeds a threshold, the start-stop state of the thermal power unit node corresponding to the start-stop state score during the corresponding prediction period is "started". If not, it is determined whether the start-stop state score is less than the difference between 1 and the threshold. If so, the start-stop state of the thermal power unit node corresponding to the start-stop state score during the corresponding prediction period is "stopped".

[0191] Furthermore, the steps for training the model include:

[0192] Obtain the load demand and clean energy generation information of the target power grid for each historical period on a weekly scale within the historical control time range, and use it as training samples. Where X t N represents the training sample corresponding to the t-th historical time period on a weekly scale. D Indicates the total load. This represents the Nth historical period corresponding to the t-th historical period on a weekly scale. D The load requirement of N loads w This indicates the total number of clean energy generator sets. This represents the Nth historical period corresponding to the t-th historical period on a weekly scale. w The output power of a clean energy generator set;

[0193] Obtain the historical start-up and shutdown status of each thermal power unit node in each historical period within the intraday scale of the historical control time range;

[0194] Based on the proportional relationship between the intraday time granularity and the weekly time granularity, the historical start-up and shutdown status of each thermal power unit node in each historical period at the intraday scale is determined according to the formula. The data is aggregated to obtain the historical start-up and shutdown status scores of each thermal power unit node in each historical period at the weekly scale, which are then used as labels for the corresponding training samples. in, t represents the historical start-up and shutdown status score of the i-th thermal power unit node in the m-th historical period on a weekly scale. r t represents the time granularity on an intraday scale. w Indicates a time granularity on a weekly scale. Ω represents the historical start-up and shutdown status of the i-th thermal power unit node in the n-th historical period on an intraday scale. m ={n|t r,n ∈t w,m m=1,2,...,T W This represents the m-th time period t belonging to the week-scale. w,m The nth time period t on the intraday scale r,n The quantity, T W N represents the total number of historical periods on a weekly scale. G This represents the total number of nodes in a thermal power unit. This represents the Nth historical period corresponding to the t-th historical period on a weekly scale. G The start / stop status score of each thermal power unit node;

[0195] The training samples and labels of each historical period within the weekly scale of the historical control time range are input into the model for training, resulting in the trained model.

[0196] Furthermore, the formula for determining the thermal power generation information corresponding to multiple prediction periods at the second time scale based on the thermal power generation information corresponding to multiple prediction periods at the first time scale is as follows:

[0197]

[0198] in, This indicates the start-up / shutdown status of the i-th thermal power unit node in the n-th forecast period of the current day. N represents the start-up / shutdown status of the i-th thermal power unit node in the m-th prediction period on a weekly scale. G The total number of nodes in thermal power units, t d,n t represents the nth forecast period on a day-ahead scale. w,m This represents the m-th prediction period on a weekly scale.

[0199] Furthermore, the calculation of the thermal power generation information for multiple forecast periods based on the thermal power generation information corresponding to multiple forecast periods under the second-level time scale, the load demand for each forecast period under the second-level time scale, and the clean energy power generation information further includes:

[0200] The start-up and shutdown status of each thermal power unit node corresponding to multiple forecast periods at the day-ahead scale, the load demand and clean energy power generation information of each forecast period at the day-ahead scale, and the constraints corresponding to the day-ahead scale are input into a commercial solver for solving, so as to obtain the start-up and shutdown status of each thermal power unit node corresponding to the remaining forecast periods at the day-ahead scale and the output of each thermal power unit node in all forecast periods at the day-ahead scale.

[0201] Furthermore, the formula for determining the thermal power generation information corresponding to multiple prediction periods at the second time scale based on the thermal power generation information corresponding to multiple prediction periods at the first time scale is as follows:

[0202]

[0203] in, This represents the basic output point of the i-th thermal power unit node in the n-th prediction period at the intraday scale. (·).start utilizes the initial value function in the Gurobi solver to pass the thermal power unit output, providing an initial heuristic solution for the decision variables, and then further optimization is performed based on this. t represents the output of the i-th thermal power unit node in the m-th forecast period at the day-ahead scale. r,n This represents the nth forecast period on an intraday scale.

[0204] Furthermore, the calculation of the thermal power generation information for multiple forecast periods based on the thermal power generation information corresponding to multiple forecast periods under the second-level time scale, the load demand for each forecast period under the second-level time scale, and the clean energy power generation information further includes:

[0205] The basic output points of each thermal power unit node in each forecast period at the intraday scale, the load demand and clean energy power generation information in each forecast period at the intraday scale, and the corresponding constraints at the intraday scale are input into a commercial solver for solving to obtain the output of each thermal power unit node in each forecast period at the intraday scale.

[0206] Furthermore, after determining the thermal power generation information corresponding to multiple prediction periods at the second time scale based on the thermal power generation information corresponding to multiple prediction periods at the first time scale, the device is further configured to:

[0207] Based on the thermal power generation information corresponding to multiple prediction periods under the first-level time scale and / or the thermal power generation information corresponding to multiple prediction periods under the second-level time scale, redundant constraints of the second-level time scale are removed.

[0208] Since the principle of the above-mentioned device in solving the problem is similar to that of the above-mentioned method, the implementation of the above-mentioned system can refer to the implementation of the above-mentioned method, and the repeated parts will not be described again.

[0209] like Figure 6 The diagram shown is a structural schematic of a computer device according to an embodiment of this specification. The methods described in this specification can be applied to the computer device of this embodiment.

[0210] Computer device 602 may include one or more processors 604, such as one or more central processing units (CPUs), each of which may implement one or more hardware threads. Computer device 602 may also include any memory 606 for storing information of any kind, such as code, settings, data, etc. Non-limitingly, for example, memory 606 may include any type of RAM, any type of ROM, flash memory, hard disk, optical disk, etc. More generally, any storage resource can be used to store information using any technology.

[0211] Furthermore, any storage resource can provide volatile or non-volatile retention of information.

[0212] Furthermore, any storage resource can represent a fixed or removable component of the computer device 602. In one case, when the processor 604 executes associated instructions stored in any storage resource or combination of storage resources, the computer device 602 can perform any operation of the associated instructions. The computer device 602 also includes one or more drive mechanisms 608 for interacting with any storage resource, such as a hard disk drive system, an optical disk drive system, etc.

[0213] Computer device 602 may also include an input / output module 610 (I / O) for receiving various inputs (via input device 612) and providing various outputs (via output device 614). A specific output mechanism may include a presentation device 616 and an associated graphical user interface (GUI) 618. In other embodiments, the input / output module 610 (I / O), input device 612, and output device 614 may be omitted, and the device may function solely as a computer device within a network. Computer device 602 may also include one or more network interfaces 620 for exchanging data with other devices via one or more communication links 622. One or more communication buses 624 couple the components described above together.

[0214] Communication link 622 can be implemented in any way, such as via a local area network, a wide area network (e.g., the Internet), a point-to-point connection, or any combination thereof. Communication link 622 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc., governed by any protocol or combination of protocols.

[0215] This specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0216] This specification also provides computer-readable instructions, wherein when a processor executes the instructions, the program therein causes the processor to perform the above-described method.

[0217] It should be understood that in the various embodiments of this specification, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this specification.

[0218] It should also be understood that, in the embodiments of this specification, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, in the embodiments of this specification, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0219] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this specification can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the embodiments in this specification.

[0220] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0221] In the embodiments provided in this specification, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the couplings or direct couplings or communication connections shown or discussed may be indirect couplings or communication connections through some interfaces, devices, or units, or they may be electrical, mechanical, or other forms of connection.

[0222] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments described in this specification, depending on actual needs.

[0223] Furthermore, the functional units in the various embodiments of this specification can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0224] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this specification, in essence, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this specification. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0225] This specification describes the principles and implementation methods of the embodiments using specific examples. The above descriptions of the embodiments are only for the purpose of helping to understand the methods and core ideas of the embodiments in this specification. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the embodiments in this specification. Therefore, the content of this specification should not be construed as a limitation on the embodiments in this specification.

Claims

1. A generator set control method based on multi-time-scale characteristics, characterized in that, Applied to a target power grid including thermal power unit nodes and clean energy power unit nodes, the method includes: Obtain the load demand and clean energy generation information of the clean energy generator nodes for each predicted period under the multi-level time scale of the target power grid within the target control time range; For the first time scale in the multi-level time scale, the thermal power generation information of the thermal power unit node corresponding to the multiple prediction periods under the first time scale is calculated based on the load demand and clean energy power generation information corresponding to each prediction period under the first time scale. Based on the thermal power generation information corresponding to multiple prediction periods under the first-level time scale, determine the thermal power generation information corresponding to multiple prediction periods under the second-level time scale. Based on the thermal power generation information corresponding to multiple forecast periods under the second-level time scale, the load demand and clean energy power generation information for each forecast period under the second-level time scale, the thermal power generation information for multiple forecast periods is calculated. Determine whether the second-level time scale is the last time scale in the multi-level time scale; If not, the second-level time scale is used as the first-level time scale, and the step of determining the thermal power generation information corresponding to multiple prediction periods under the second-level time scale based on the thermal power generation information corresponding to multiple prediction periods under the first-level time scale is repeated. If so, the thermal power generation information of the thermal power unit nodes in the target control time range is controlled using the thermal power generation information of multiple predicted time periods at the last time scale.

2. The method according to claim 1, characterized in that, The multi-level time scales include weekly, day-ahead, and intraday scales arranged in sequence.

3. The method according to claim 2, characterized in that, The thermal power generation information includes the start-up and shutdown status and basic output points of the thermal power unit nodes; The calculation of thermal power generation information for the thermal power unit nodes corresponding to multiple prediction periods under the first-level time scale, based on the load demand and clean energy power generation information for each prediction period under this first-level time scale, further includes: The load demand and clean energy power generation information corresponding to each prediction period under the weekly scale are input into the pre-trained model for calculation to obtain the start-up and shutdown status scores of each thermal power unit node under each prediction period. If the start-stop state score exceeds a threshold, the start-stop state of the thermal power unit node corresponding to the start-stop state score during the corresponding prediction period is "started". If not, it is determined whether the start-stop state score is less than the difference between 1 and the threshold. If so, the start-stop state of the thermal power unit node corresponding to the start-stop state score during the corresponding prediction period is "stopped".

4. The method according to claim 3, characterized in that, The steps for training the model include: Obtain the load demand and clean energy generation information of the target power grid for each historical period on a weekly scale within the historical control time range, and use it as training samples. Where X t N represents the training sample corresponding to the t-th historical time period on a weekly scale. D Indicates the total load. This represents the Nth historical period corresponding to the t-th historical period on a weekly scale. D The load requirement of N loads w This indicates the total number of clean energy generator sets. This represents the Nth historical period corresponding to the t-th historical period on a weekly scale. w The output power of a clean energy generator set; Obtain the historical start-up and shutdown status of each thermal power unit node in each historical period within the intraday scale of the historical control time range; Based on the proportional relationship between the intraday time granularity and the weekly time granularity, the historical start-up and shutdown status of each thermal power unit node in each historical period at the intraday scale is determined according to the formula. The data is aggregated to obtain the historical start-up and shutdown status scores of each thermal power unit node in each historical period at the weekly scale, which are then used as labels for the corresponding training samples. in, t represents the historical start-up and shutdown status score of the i-th thermal power unit node in the m-th historical period on a weekly scale. r t represents the time granularity on an intraday scale. w Indicates a time granularity on a weekly scale. Ω represents the historical start-up and shutdown status of the i-th thermal power unit node in the n-th historical period on an intraday scale. m ={n|t r,n ∈t w,m m = 1, 2, ... ,T W This represents the m-th time period t belonging to the week-scale. w,m The nth time period t on the intraday scale r,n The quantity, T W N represents the total number of historical periods on a weekly scale. G This represents the total number of nodes in a thermal power unit. This represents the Nth historical period corresponding to the t-th historical period on a weekly scale. G The start / stop status score of each thermal power unit node; The training samples and labels of each historical period within the weekly scale of the historical control time range are input into the model for training, resulting in the trained model.

5. The method according to claim 3, characterized in that, The formula for determining the thermal power generation information corresponding to multiple prediction periods at the second time scale based on the thermal power generation information corresponding to multiple prediction periods at the first time scale is as follows: in, This indicates the start-up / shutdown status of the i-th thermal power unit node in the n-th forecast period of the current day. N represents the start-up / shutdown status of the i-th thermal power unit node in the m-th prediction period on a weekly scale. G The total number of nodes in thermal power units, t d,n t represents the nth forecast period on a day-ahead scale. w,m This represents the m-th prediction period on a weekly scale.

6. The method according to claim 5, characterized in that, Based on the thermal power generation information corresponding to multiple forecast periods under the second-level time scale, the load demand and clean energy power generation information for each forecast period under the second-level time scale, the calculation of the thermal power generation information for multiple forecast periods further includes: The start-up and shutdown status of each thermal power unit node corresponding to multiple forecast periods at the day-ahead scale, the load demand and clean energy power generation information of each forecast period at the day-ahead scale, and the constraints corresponding to the day-ahead scale are input into a commercial solver for solving, so as to obtain the start-up and shutdown status of each thermal power unit node corresponding to the remaining forecast periods at the day-ahead scale and the output of each thermal power unit node in all forecast periods at the day-ahead scale.

7. The method according to claim 6, characterized in that, The formula for determining the thermal power generation information corresponding to multiple prediction periods at the second time scale based on the thermal power generation information corresponding to multiple prediction periods at the first time scale is as follows: in, This represents the basic output point of the i-th thermal power unit node in the n-th prediction period at the intraday scale. (·).start utilizes the initial value function in the Gurobi solver to pass the thermal power unit output, providing an initial heuristic solution for the decision variables, and then further optimization is performed based on this. t represents the output of the i-th thermal power unit node in the m-th forecast period at the day-ahead scale. r,n This represents the nth forecast period on an intraday scale.

8. The method according to claim 7, characterized in that, Based on the thermal power generation information corresponding to multiple forecast periods under the second-level time scale, the load demand and clean energy power generation information for each forecast period under the second-level time scale, the calculation of the thermal power generation information for multiple forecast periods further includes: The basic output points of each thermal power unit node in each forecast period at the intraday scale, the load demand and clean energy power generation information in each forecast period at the intraday scale, and the corresponding constraints at the intraday scale are input into a commercial solver for solving to obtain the output of each thermal power unit node in each forecast period at the intraday scale.

9. The method according to claim 1, characterized in that, After determining the thermal power generation information corresponding to multiple prediction periods at the second time scale based on the thermal power generation information corresponding to multiple prediction periods at the first time scale, the method further includes: Based on the thermal power generation information corresponding to multiple prediction periods under the first-level time scale and / or the thermal power generation information corresponding to multiple prediction periods under the second-level time scale, redundant constraints of the second-level time scale are removed.

10. A generator set control device based on multi-time-scale characteristics, characterized in that, Deployed in a target power grid including thermal power unit nodes and clean energy power unit nodes, the device includes: The information acquisition unit is used to acquire the load demand and clean energy generation information of the clean energy generator nodes corresponding to each predicted period under the multi-level time scale of the target power grid within the target control time range. The first-level time scale calculation unit is used to calculate the thermal power generation information of the thermal power unit node corresponding to multiple prediction periods under the first-level time scale, based on the load demand and clean energy power generation information corresponding to each prediction period under the first-level time scale. The information transmission unit is used to determine the thermal power generation information corresponding to multiple prediction periods under the second time scale based on the thermal power generation information corresponding to multiple prediction periods under the first time scale. The second-level time scale calculation unit is used to calculate the thermal power generation information for multiple prediction periods based on the thermal power generation information corresponding to multiple prediction periods under the second-level time scale, the load demand and clean energy power generation information for each prediction period under the second-level time scale. An iterative calculation unit is used to determine whether the second-level time scale is the last time scale among the multi-level time scales; if not, the second-level time scale is used as the first-level time scale, and the step of determining the thermal power generation information corresponding to multiple prediction periods under the second-level time scale based on the thermal power generation information corresponding to multiple prediction periods under the first-level time scale is repeated. The control unit is used to control the thermal power unit node within the target control time range by using the thermal power generation information of multiple predicted time periods of the last time scale when the judgment result of the iterative calculation unit is yes.

Images