Demand side resource scheduling method and system based on different scene adjustment of power grid

Abstract

The invention relates to the technical field of power dispatching, and discloses a demand side resource dispatching method and system based on different scene adjustment of a power grid, and the method comprises the steps: obtaining the historical data of a load abnormal quantity and an abnormal change duration index; classifying the power grid scenes according to the obtained historical data to obtain different power grid scenes; carrying out analysis according to the load curve characteristics in different power grid scenes, obtaining adjustment requirements of the power grid in different scenes, and establishing callable resource libraries matched with the power grid in different scenes; the scenes are trained and learned through an MLP model, and different scenes are identified in combination with corresponding evaluation indexes so as to verify the validity of scene identification; scene adjustment is realized by building scheduling models and constraint models of different scenes of a power grid. According to the method, the power grid scenes are effectively classified, the types of the demand side resources in the resource library, the response time period, the response speed and the response cost information are clarified, and rapid adjustment and accurate scheduling of the power grid are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of power dispatching, and in particular to a demand-side resource dispatching method and system based on power grid adjustment in different scenarios. Background Art

[0002] Due to the frequent occurrence of extreme events and the complex operation scenarios of the power grid, as my country actively builds a new power system based on new energy such as wind and solar power, the power system lacks sufficient flexibility in scheduling during certain periods, bringing new challenges to power grid planning and power supply security.

[0003] At present, a lot of research has been carried out on the generation methods of typical power grid scenarios, which can be summarized as follows: K-means clustering, hierarchical clustering, density clustering, fuzzy mean clustering and scenario reduction method, etc. Most of the existing studies use clustering methods to generate power grid scenarios. This type of method is simple in principle and widely used, but the clustering effect is easily affected by noise data. If the initial cluster center is not selected properly, the clustering effect may be poor.

[0004] Therefore, a scheduling method is needed that can effectively meet the regulation needs of different scenarios, better realize the classification of power grid scenarios, improve the flexibility of power system scheduling, and achieve the economy of power grid operation. Summary of the invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a demand-side resource scheduling method and system based on adjustment of different power grid scenarios to solve the problem that the existing method uses a clustering method to generate power grid scenarios, but the clustering effect is easily affected by noise data and the effect is poor. It mainly focuses on optimizing the scheduling of demand-side resources under normal power grid operating scenarios, and cannot consider other power grid scenarios at the same time, and lacks sufficient flexibility in scheduling in certain time periods.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides a demand-side resource scheduling method based on different power grid scenario regulation, comprising:

[0009] Obtain historical data of two indicators: abnormal load amount and duration of abnormal change;

[0010] Classify the power grid scenarios based on the acquired historical data to obtain different power grid scenarios;

[0011] Analyze the load curve characteristics under different power grid scenarios, obtain the regulation requirements of different power grid scenarios, and establish a callable resource library that matches different power grid scenarios;

[0012] The multi-layer perceptron model (MLP) is used to train and learn different power grid scenarios, and different scenarios are identified based on evaluation indicators. The effectiveness of scenario identification is verified based on the numerical values ​​of the evaluation indicators.

[0013] The verified different scenarios are matched and adjusted with the resource library, and the matching and adjustment of the scenarios are achieved by building dispatch models and constraint models for different scenarios of the power grid.

[0014] As a preferred solution of the demand-side resource scheduling method based on different power grid scenario regulation described in the present invention, the load abnormality and the abnormal change duration are selected based on the load change, expressed as:

[0015] Q adn (t)=Q(t)≥aQ avg (t)

[0016]

[0017] Among them, Q adn (t) is the load abnormality at time t, Q(t) is the load change at time t, α is the first abnormality coefficient, Q avg (t) is the average value of load change in one day, t du is the first change moment of load change, β is the second abnormal coefficient; T is the total number of moments in a day, is the duration, t adn The maximum value of the time allowed for drastic changes in a day.

[0018] As a preferred solution of the demand-side resource scheduling method based on different power grid scenario adjustment described in the present invention, the different power grid scenarios include: rigid supply guarantee scenarios, risk resistance scenarios and flexible adjustment scenarios.

[0019] As a preferred solution of the demand-side resource scheduling method based on different power grid scenarios described in the present invention, a resource library that can be called to match different power grid scenarios is established, including:

[0020] Based on the economic dispatch target in the flexible regulation scenario, a first resource library is randomly generated, and the first resource library includes the type of demand-side resources, the responsive time period, the response speed, the response capacity, and the response cost;

[0021] Based on the supply and demand balance goal in the rigid supply guarantee scenario, the response capacity of demand-side resources is expanded on the basis of the first resource pool, the response cost is changed, and the second resource pool is generated;

[0022] Based on the short-term load matching targets of the power generation side and the power consumption side in the risk resistance scenario, on the basis of the second resource library, the response speed of the demand-side resources is improved, the response cost is changed, and the third resource library is generated.

[0023] As a preferred solution of the demand-side resource scheduling method based on different power grid scenario regulation described in the present invention, wherein: building a scheduling model for different power grid scenarios to achieve scenario regulation includes:

[0024] Construct a dispatch model with the lowest grid operation cost in the flexible regulation scenario. Its specific expression is as follows:

[0025] min C1=f1+f2+f3

[0026]

[0027] Among them, C1 is the total operating cost of the power grid in the flexible scenario, f1 is the cost of purchasing electricity from the power grid, f2 is the total cost of calling demand-side resources, f3 is the total cost of abandoning wind and solar power, and C ele is the time-of-use electricity price, P grid The amount of electricity purchased from the power grid, is the cost of calling demand-side resources in different demand-side resource pools, j represents the demand-side resource pool, i represents the type of demand-side resources in the demand-side resource pool, C wind , C solar Represent the cost coefficients of wind and solar power abandonment, Represent the amount of wind power curtailment and solar power curtailment respectively.

[0028] As a preferred solution of the demand-side resource scheduling method based on different power grid scenarios according to the present invention, the scheduling model for different power grid scenarios also includes:

[0029] Construct a dispatch model under the rigid supply guarantee scenario to ensure that the power grid load shedding is minimized while meeting the minimum power grid operation cost. The specific expression is as follows:

[0030] min C2=f1+f2+f3+f4

[0031]

[0032] Among them, C2 is the total cost of grid operation under the rigid supply guarantee scenario, f4 is the load shedding cost, is the load shedding cost coefficient, The load shedding amount.

[0033] As a preferred solution of the demand-side resource scheduling method based on different power grid scenarios according to the present invention, the scheduling model for different power grid scenarios also includes:

[0034] Construct a dispatch model under risk resistance scenarios to reduce the peak-to-valley difference rate of the power grid and minimize the power grid operation cost. The specific expression is as follows:

[0035] min C3=f1+f2+f3+f4

[0036]

[0037] Among them, C3 is the total cost of grid operation under the risk resistance scenario, C is the peak-to-valley difference rate, P max and P min They are the maximum load and minimum load of the day respectively.

[0038] In a second aspect, the present invention provides a demand-side resource scheduling system based on different power grid scenario adjustments, including:

[0039] An acquisition module is used to obtain historical data of two indicators: load abnormality and abnormal change duration;

[0040] A classification module is used to classify power grid scenarios based on the acquired historical data to obtain different power grid scenarios;

[0041] The resource library establishment module is used to analyze the load curve characteristics under different power grid scenarios, obtain the regulation requirements of different power grid scenarios, and establish a callable resource library that matches different power grid scenarios;

[0042] The training module is used to train and learn different power grid scenarios through the multi-layer perceptron model MLP, identify different scenarios based on evaluation indicators, and verify the effectiveness of scenario identification based on the evaluation indicator values;

[0043] The adjustment module is used to match and adjust the verified different scenarios with the resource library, and achieve scenario matching and adjustment by building scheduling models and constraint models for different power grid scenarios.

[0044] Also includes,

[0045] The first building module is used to build a dispatching model with the lowest grid operation cost in a flexible regulation scenario;

[0046] The second building module is used to build a scheduling model in the rigid supply guarantee scenario;

[0047] The third building block is to construct a scheduling model for risk resistance scenarios.

[0048] In a third aspect, the present invention provides a computing device, comprising:

[0049] Memory and processor;

[0050] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the demand-side resource scheduling method based on different power grid scenarios are implemented.

[0051] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, can implement the steps of the demand-side resource scheduling method based on different power grid scenario adjustments.

[0052] Compared with the prior art, the present invention has the following beneficial effects: the present invention realizes effective classification of power grid scenarios by establishing a callable resource library under different scenarios, establishes a callable resource library matching the regulation needs of different power grid scenarios, clarifies the types of demand-side resources in the resource library, the responsive time period, response speed and response cost information, realizes rapid regulation and precise scheduling of the power grid, and effectively improves the stability and economy of the power grid operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0054] Figure 1 A schematic diagram of the overall process logic of the demand-side resource scheduling method based on different power grid scenario regulation according to the first embodiment of the present invention;

[0055] Figure 2 This is a schematic diagram of the overall framework structure of the demand-side resource scheduling method based on different power grid scenario regulation according to the first embodiment of the present invention;

[0056] Figure 3 A schematic diagram of a load output curve of a certain region in the demand-side resource scheduling method based on power grid adjustment in different scenarios according to the second embodiment of the present invention;

[0057] Figure 4 This is a schematic diagram of various scenario curves in a certain area in the demand-side resource scheduling method based on different power grid scenario adjustment described in the second embodiment of the present invention. DETAILED DESCRIPTION

[0058] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0059] Example 1, reference Figure 1-Figure 2 , is an embodiment of the present invention, and provides a demand-side resource scheduling method based on different power grid scenario regulation, comprising:

[0060] S100: Acquire historical data of two indicators: load abnormality and abnormal change duration;

[0061] Specifically, two indicators, namely, the load abnormality amount and the duration of abnormal change, are proposed based on the daily average load.

[0062] It should be noted that in order to better capture the trend and amplitude of load fluctuations, analyze the changes in load, the dynamic changes in power grid operation, and timely discover potential abnormalities, it provides an important reference for the operation and management of the power system. The load change Q(t) represents the fluctuation of load in a specific time period. The details are as follows.

[0063]

[0064] Among them, Q(t) is the load change at time t, q(t) and q(t-1) are the load amounts at time t and t-1 respectively.

[0065] In order to further more effectively and accurately judge the abnormal situation of load and realize the division of power grid scenarios, an appropriate range is selected as the threshold based on the load change amount Q(t) to measure the normal range of load change, and two key indicators, load abnormality and duration, are introduced as the basis for scenario generation. adn (t) describes the situation where the load changes extremely drastically during the day, and the duration indicator Describes the duration of drastic load changes during a day.

[0066] In the embodiment of the present application, the load abnormality Q adn (t) and duration of abnormal change The selection is based on the load change, expressed as:

[0067] Q adn (t)=Q(t)≥aQ avg (t) (2)

[0068]

[0069] Among them, Q adn (t) is the load abnormality at time t, Q(t) is the load change at time t, α is the first abnormality coefficient, Q avg (t) is the average value of load change in one day, t du is the first change moment of load change, β is the second abnormal coefficient; T is the total number of moments in a day, is the duration, t adn The maximum value of the time allowed for drastic changes in a day.

[0070] Specifically, the first abnormality coefficient α is an extremely drastic abnormality coefficient, for example, α=115% or is set as required;

[0071] The first change time of load change t du For moments when the change is more drastic, for example: moments when the change amplitude exceeds 15% in the whole change or set as needed;

[0072] The second abnormality coefficient β is a more severe abnormality coefficient, for example, β=1.3 or set as required;

[0073] The maximum value t of the time that allows drastic changes in a day adn The maximum value of the time that allows for more drastic changes in a day, for example: t adn =2 or set as needed.

[0074] S200: Classifying power grid scenarios according to the acquired historical data to obtain different power grid scenarios;

[0075] In the embodiments of the present application, different power grid scenarios include: rigid supply guarantee scenario, risk resistance scenario and flexible regulation scenario.

[0076] Among them, in the flexible regulation scenario, the load curve changes smoothly, and since the power grid itself has a certain regulation capability, the overall supply and demand of electricity is balanced; in the rigid supply guarantee scenario, due to extreme weather or other factors, the demand for electricity may be at a high level for a long time, resulting in insufficient power supply; in the risk resistance scenario, the load fluctuates greatly at a certain moment, the load on the demand side changes dramatically in a short period of time, and the supply and demand of electricity is extremely unbalanced.

[0077] S300: Analyze the load curve characteristics under different power grid scenarios, obtain the regulation requirements of different power grid scenarios, and establish a callable resource library matching different power grid scenarios;

[0078] Specifically, the load curve characteristics under different power grid scenarios are analyzed to obtain the regulation requirements of different power grid scenarios, including:

[0079] Flexible regulation: If the load fluctuation rate of the load curve in the flexible regulation scenario is less than the set first threshold, the economy of the demand-side resources is adjusted. The first threshold can be 15% or adjusted according to actual demand;

[0080] This scenario shows that the load curve is basically stable, with a small peak-to-valley difference, requiring demand-side resources to have good economic efficiency.

[0081] Rigid supply guarantee: If the daily average load of the load curve in the rigid supply guarantee scenario is greater than the preset ratio of the annual daily average load, and the imbalance between power supply and demand in the power grid occurs within a certain period of time during the day, the demand-side resources can respond to the capacity, and the preset ratio is 1.2 times or adjusted according to actual demand;

[0082] This scenario shows that the load on the electricity consumption side has been at a high level, and the imbalance between electricity supply and demand in the power grid has been long-term throughout the day, accompanied by obvious seasonality, requiring the demand-side resources to have a large responsive capacity.

[0083] Risk mitigation: If the sudden fluctuation rate of the load curve in the risk mitigation scenario is greater than the set second threshold or a power outage occurs, the flexibility of demand-side resources will be adjusted. The second threshold can be 30% or adjusted according to actual demand;

[0084] This scenario describes a sudden load drop or increase, or a power outage, which requires demand-side resources to have strong flexibility.

[0085] In the embodiment of the present application, a resource library that can be called to match different scenarios of the power grid is established;

[0086] It should be noted that based on the analysis of grid regulation needs, taking into account the uncertainty and complexity of actual demand response business, a random simulation method is used for analysis, and the type, available time period, response speed and call cost of each resource in the resource library are randomly distributed and generated.

[0087] Specifically include:

[0088] Based on the economic dispatch target in the flexible regulation scenario, a first resource pool A is randomly generated, and the first resource pool A includes the type of demand-side resources, the responsive time period, the response speed, the response capacity, and the response cost;

[0089] It should be noted that the above settings can make the resource response cost in the first resource library A low.

[0090] Based on the supply and demand balance goal in the rigid supply guarantee scenario, on the basis of the first resource pool A, the response capacity of demand-side resources is expanded, the response cost is changed, and the second resource pool B is generated;

[0091] Based on the short-term load matching targets of the power generation side and the power consumption side in the risk resistance scenario, on the basis of the second resource pool B, the response speed of the demand-side resources is improved, the response cost is changed, and the third resource pool C is generated.

[0092] S400: Train and learn different power grid scenarios through the multi-layer perceptron model MLP, identify different scenarios based on evaluation indicators, and verify the effectiveness of scenario identification based on the evaluation indicator values;

[0093] It should be noted that when using the MLP model for training and learning, considering the large data set, choosing the ReLU function as the activation function can significantly reduce the amount of calculation, improve the learning efficiency of the neural network, and alleviate the overfitting problem.

[0094] Among them, the ReLU function is defined as:

[0095] ReLU(x)=map(x,0) (5)

[0096] The calculation equation of the hidden layer is:

[0097] H=ω T x+b (6)

[0098] h=ReLU(H) (7)

[0099] Among them, ω is the weight, the higher the weight, the more important the feature is, b is the bias; h is the activation value corresponding to H.

[0100] Specifically, the evaluation indicators include: Accuracy (A), Precision (P), Recall (R), and F1 Score (F1 Score, F1).

[0101] Respectively expressed as:

[0102]

[0103] Among them, t p The model correctly predicts the actual true sample as the true sample; f n The model mistakenly predicts the actual true sample as a false sample; f p The model mistakenly predicts a fake sample as a true sample; n The model correctly predicts a sample that is actually fake as a fake sample.

[0104] It should be noted that the four evaluation indicators can analyze the performance of the model in multiple dimensions and measure the effectiveness of MLP model training. Among them, accuracy describes the percentage of correctly predicted results in the total samples, judging the overall accuracy of the samples; precision describes the proportion of samples that are actually true among the samples predicted to be true, which is an evaluation of the accuracy of the prediction of true samples; recall describes the proportion of samples that are predicted to be true among all samples that are actually true; F1 score is the harmonic mean of precision and recall.

[0105] S500: Match and adjust the verified different scenarios with the resource library, and achieve scenario matching and adjustment by building a dispatch model and constraint model for different scenarios of the power grid.

[0106] In the embodiment of the present application, building a dispatching model for different power grid scenarios to achieve scenario adjustment includes:

[0107] Construct a dispatch model with the lowest grid operation cost in the flexible regulation scenario. Its specific expression is as follows:

[0108] min C1=f1+f2+f3 (12)

[0109]

[0110] Among them, C1 is the total operating cost of the power grid in the flexible scenario, f1 is the cost of purchasing electricity from the power grid, f2 is the total cost of calling demand-side resources, f3 is the total cost of abandoning wind and solar power, and C ele is the time-of-use electricity price, P grid The amount of electricity purchased from the power grid, is the cost of calling demand-side resources in different demand-side resource pools, j represents the demand-side resource pool, i represents the type of demand-side resources in the demand-side resource pool, C wind , C solar Represent the cost coefficients of wind and solar power abandonment, Represent the amount of wind power curtailment and solar power curtailment respectively.

[0111] In the embodiment of the present application, the dispatching model for building different power grid scenarios also includes:

[0112] Construct a dispatch model under the rigid supply guarantee scenario to ensure that the power grid load shedding is minimized while meeting the minimum power grid operation cost. The specific expression is as follows:

[0113] min C2=f1+f2+f3+f4 (16)

[0114]

[0115] Among them, C2 is the total cost of grid operation under the rigid supply guarantee scenario, f4 is the load shedding cost, is the load shedding cost coefficient, The load shedding amount.

[0116] In the embodiment of the present application, the dispatching model for building different power grid scenarios also includes:

[0117] Construct a dispatch model under risk resistance scenarios to reduce the peak-to-valley difference rate of the power grid and minimize the power grid operation cost. The specific expression is as follows:

[0118] min C3=f1+f2+f3+f4 (18)

[0119]

[0120] Among them, C3 is the total cost of grid operation under the risk resistance scenario, C is the peak-to-valley difference rate, P max and P min They are the maximum load and minimum load of the day respectively.

[0121] Specifically, it also includes constraints, and the constraints in different scenarios are the same, which are specifically expressed as:

[0122] Construct an electric power balance constraint model. Its specific expression is as follows:

[0123]

[0124] Among them, P solar (t) is the photovoltaic power generation power at time t, P wind (t) is the wind power output at time t, P grid (t) is the power purchased from the power grid at time t, p load (t) is the electric load power at time t, is the increased electric load power of demand-side resources at time t, is the electric load power reduced by the demand-side resources at time t, is the load shedding power of the power system.

[0125] Construct a demand-side resource response capacity constraint model. Its specific expression is as follows:

[0126]

[0127] in, is the maximum available capacity of demand-side resources at time t.

[0128] Construct a demand-side resource response time constraint model. Its specific expression is as follows:

[0129]

[0130] Among them, 0 means that the demand-side resources cannot be called at the current moment, and 1 means that the demand-side resources can be called at the current moment.

[0131] Construct a wind and solar output constraint model. Its specific expression is as follows:

[0132]

[0133] in, is the maximum output power of photovoltaic power generation at time t, is the maximum output power of wind power at time t.

[0134] Construct the output and climbing constraint model of thermal power units. Its specific expression is as follows:

[0135]

[0136] in, is the output power of unit k at time t, and is the upper and lower limits of the output of thermal power units, MW; and is the up and down ramp rate of unit k, MW / h.

[0137] In summary, in terms of the coordinated optimization and scheduling of multiple types of demand-side resources, the current focus is on optimizing the scheduling of demand-side resources under normal power grid operating conditions, and other power grid scenarios are rarely considered at the same time. The invention is mainly to meet the regulation needs of different power grid scenarios, to achieve rapid call on resources, and to improve the stability and economy of power grid operation.

[0138] The above is a schematic scheme of a demand-side resource scheduling method based on different power grid scenarios in this embodiment. It should be noted that the technical solution of the system for demand-side resource scheduling based on different power grid scenarios belongs to the same concept as the technical solution of the above-mentioned demand-side resource scheduling method based on different power grid scenarios. For details not described in detail in the technical solution of the demand-side resource scheduling system based on different power grid scenarios in this embodiment, please refer to the description of the technical solution of the above-mentioned demand-side resource scheduling method based on different power grid scenarios.

[0139] The demand-side resource scheduling system based on different power grid scenarios in this embodiment includes:

[0140] An acquisition module is used to obtain historical data of two indicators: load abnormality and abnormal change duration;

[0141] A classification module is used to classify power grid scenarios based on the acquired historical data to obtain different power grid scenarios;

[0142] The resource library establishment module is used to analyze the load curve characteristics under different power grid scenarios, obtain the regulation requirements of different power grid scenarios, and establish a callable resource library that matches the different power grid scenarios;

[0143] The training module is used to train and learn the scenes through the MLP model, and identify different scenes in combination with corresponding evaluation indicators to verify the effectiveness of scene identification;

[0144] The adjustment module is used to match and adjust the verified different scenarios with the resource library. The matching and adjustment are achieved by building scheduling models and constraint models for different power grid scenarios.

[0145] This embodiment further provides a computing device, which is applicable to the case of demand-side resource scheduling based on different scenarios of power grid regulation, including:

[0146] A memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the demand-side resource scheduling method based on different power grid scenarios as proposed in the above embodiment.

[0147] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for implementing demand-side resource scheduling based on different power grid scenarios as proposed in the above embodiment is implemented.

[0148] The storage medium proposed in this embodiment and the demand-side resource scheduling method for realizing regulation based on different power grid scenarios proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0149] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course can also be implemented by hardware. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (Random Access Memory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0150] Example 2, refer to Table 1 and Figure 3-Figure 4Based on the previous embodiment, this embodiment provides an application case of a demand-side resource scheduling method based on different power grid scenario adjustments to illustrate the feasibility of our solution.

[0151] This embodiment is based on the GAMS platform, and performs simulation and optimization analysis in the WIN11 operating system and the AMD Ryzen 7 5800H 3.20 GHz processor environment.

[0152] In this embodiment, in order to verify the effectiveness and efficiency of the proposed demand-side resource scheduling method based on the regulation demand of different power grid scenarios, the load output and new energy output data of a certain region from January 1, 2023 to December 30, 2023 are selected for case analysis to verify the effectiveness of the proposed model, namely S400.

[0153] Select a region’s one-year load output and renewable energy output data to generate scenarios. The load output curve is as follows: Figure 3 , the sampling point granularity of the horizontal axis in the figure is 1h. Based on the proposed load abnormality and duration indicators, the load output curve generates three scenarios, namely, rigid supply guarantee scenario, risk resistance scenario and flexible adjustment scenario. The curves of each scenario are as follows Figure 4 .

[0154] Based on the scene generation results obtained by the proposed method, the scene is identified and the MLP model is used for training and learning. The four evaluation indicators of accuracy, precision, recall rate and F1 score are combined to provide effectiveness verification for scene identification. Each scene is divided into training set, validation set and test set in a ratio of 7:1.5:1.5. The MLP model is optimized for multiple iterations to effectively analyze the performance of scene identification. The analysis results are shown in Table 1.

[0155] Table 1 Evaluation index values

[0156]

[0157] It can be seen from Table 1 that the accuracy, precision and recall rate of the validation set model are all above 0.9, and the F1 score is also close to 0.9, which shows that the proposed scene generation method is feasible and can well distinguish the characteristics of each scene, providing support for the establishment and matching of subsequent resource libraries; the accuracy, precision and recall rate of the test set model are slightly lower than those of the validation set model, but can also be maintained at around 0.9, and its F1 score is not much different from that of the validation set model, indicating that the MLP model can effectively identify the generated scenes, verifying the effectiveness and rationality of the scene generation method.

[0158] Example 3, referring to Tables 2-6, based on the previous example, this example provides a comparative case of a demand-side resource scheduling method based on different power grid scenario adjustments to illustrate the feasibility and beneficial effects of our solution.

[0159] Based on the above scenario identification and scheduling model, the CONOPT solver in the GAMS software under the Win11 operating system is used to solve and carry out case analysis. This example uses the load output and renewable energy output data of a certain region from January 1, 2023 to December 30, 2023 for analysis. There are three thermal power stations in the region, where i is 10, j is 3, and k is 3. The relevant parameters of the thermal power units are shown in Table 2.

[0160] Table 2 Relevant parameters of thermal power units

[0161]

[0162] The grid load shedding cost coefficient is 1, and the cost coefficient of wind and solar power abandonment in the flexible scenario is 0.7. In order to further reduce the phenomenon of wind and solar power abandonment in the rigid supply guarantee scenario and risk resistance scenario, the cost coefficient of wind and solar power abandonment is set to 0.73. The grid purchase price parameters are shown in Table 3.

[0163] Table 3 Grid power purchase price parameters

[0164]

[0165] There are four dispatching methods. Not calling the resource library means that only the power generation resources on the power generation side are used for dispatching, and the demand side resources do not participate in the dispatching; calling the first resource library A, calling the second resource library B, and calling the third resource library C means using the resources in the corresponding resource library for optimized dispatching to achieve source-load interaction.

[0166] (1) Flexible adjustment scenario.

[0167] In the flexible adjustment scenario, the resources in resource library A are scheduled. Table 4 shows the scheduling cost comparison of different scheduling methods in the flexible adjustment scenario.

[0168] Table 4 Comparison of dispatching costs of different dispatching methods under flexible regulation scenarios

[0169]

[0170] (2) Rigid supply guarantee scenario

[0171] 1) Call the first resource library A

[0172] In the rigid supply guarantee scenario, the resources in the first resource pool A are dispatched. After the dispatch, the original load level is effectively reduced. At the same time, the total dispatch cost is 9.1742 million yuan, the cost of wind and solar abandonment is 270,100 yuan, and the cost of load shedding is 426,000 yuan.

[0173] 2) Call the second resource library B

[0174] In the rigid supply guarantee scenario, the resources in the second resource pool B are used. After scheduling, the original load level is effectively reduced, the scheduling cost is 8.8333 million yuan, the cost of wind and solar abandonment is 55,500 yuan, and the cost of load shedding is 257,000 yuan. Table 5 compares the scheduling costs of different scheduling methods in the rigid supply guarantee scenario.

[0175] Table 5 Comparison of dispatching costs of different dispatching methods under rigid supply guarantee scenario

[0176]

[0177] (3) Risk mitigation scenarios

[0178] 1) Call the second resource library B

[0179] After the dispatch, the load curve decreased during the period when the renewable energy generation was low, and increased at noon, which effectively improved the wind and solar power consumption rate. In the evening, renewable energy generation decreased, the output of thermal power units was restricted, and the residential electricity load increased, resulting in source-load imbalance, and load shedding was forced. The dispatch cost was 6.5088 million yuan, the cost of wind and solar abandonment was 220,500 yuan, and the cost of load shedding was 468,000 yuan.

[0180] 2) Call the third resource library C

[0181] After dispatching, the load shedding was significantly reduced compared with calling the second resource pool B. The dispatching cost was 6.4487 million yuan, the cost of wind and solar abandonment was 111,000 yuan, and the cost of load shedding was 103,000 yuan. Table 6 compares the dispatching costs of different dispatching methods under risk mitigation scenarios.

[0182] Table 6 Comparison of dispatching costs of different dispatching methods under risk resistance scenarios

[0183]

[0184] Based on the above analysis, the demand-side resource scheduling method proposed in the present invention shows good economy in different typical scenarios of power systems, and can enable the power system to maintain safe and stable operation in a complex and changeable environment; and by comparing the operation results of calling different resource libraries in different scenarios, it is explained that the flexible adjustment scenario tends to call economical demand-side resources, the rigid supply guarantee scenario tends to call demand-side resources with larger callable capacity, and the risk resistance scenario tends to call demand-side resources with stronger flexibility, which provides guiding suggestions for subsequent power system scheduling.

[0185] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A demand-side resource scheduling method based on different power grid scenario regulation, characterized in that: include: Obtain historical data of two indicators: abnormal load amount and abnormal change duration; Classify the power grid scenarios based on the acquired historical data to obtain different power grid scenarios; Analyze the load curve characteristics under different power grid scenarios, obtain the regulation requirements of different power grid scenarios, and establish a callable resource library that matches different power grid scenarios; The multi-layer perceptron model (MLP) is used to train and learn different power grid scenarios, and different scenarios are identified based on evaluation indicators. The effectiveness of scenario identification is verified based on the numerical values ​​of the evaluation indicators. The verified different scenarios are matched and adjusted with the resource library, and the matching and adjustment of the scenarios are achieved by building dispatch models and constraint models for different scenarios of the power grid.

2. The demand-side resource scheduling method based on different power grid scenario regulation according to claim 1 is characterized in that: The load abnormality and abnormal change duration are selected based on the load change, and are expressed as: Q adn (t)=Q(t)≥aQ avg (t) Among them, Q adn (t) is the load abnormality at time t, Q(t) is the load change at time t, α is the first abnormality coefficient, Q avg (t) is the average value of load change in one day, t du is the first change moment of load change, β is the second abnormal coefficient; T is the total number of moments in a day, is the duration, t adn The maximum value of the time allowed for drastic changes in a day.

3. The demand-side resource scheduling method based on different power grid scenario adjustment according to claim 2 is characterized in that: The different power grid scenarios include: rigid supply guarantee scenario, risk resistance scenario and flexible regulation scenario; Establish a callable resource library that matches different power grid scenarios, including: Based on the economic dispatch target in the flexible regulation scenario, a first resource library is randomly generated, and the first resource library includes the type of demand-side resources, the responsive time period, the response speed, the response capacity, and the response cost; Based on the supply and demand balance goal in the rigid supply guarantee scenario, the response capacity of demand-side resources is expanded on the basis of the first resource pool, the response cost is changed, and the second resource pool is generated; Based on the short-term load matching targets of the power generation side and the power consumption side in the risk resistance scenario, on the basis of the second resource library, the response speed of the demand-side resources is improved, the response cost is changed, and the third resource library is generated.

4. The demand-side resource scheduling method based on different power grid scenario regulation according to claim 3 is characterized in that: Building dispatch models for different power grid scenarios to achieve scenario adjustment includes: A dispatching model with the lowest grid operation cost under the flexible regulation scenario is constructed, which is specifically expressed as follows: min C1=f1+f2+f3 Among them, C1 is the total operating cost of the power grid in the flexible scenario, f1 is the cost of purchasing electricity from the power grid, f2 is the total cost of calling demand-side resources, f3 is the total cost of abandoning wind and solar power, and C ele is the time-of-use electricity price, P grid The amount of electricity purchased from the power grid, is the cost of calling demand-side resources in different demand-side resource pools, j represents the demand-side resource pool, i represents the type of demand-side resources in the demand-side resource pool, C wind , C solar Represent the cost coefficients of wind and solar power abandonment, Represent the amount of wind power curtailment and solar power curtailment respectively.

5. The demand-side resource scheduling method based on different power grid scenario regulation according to claim 4 is characterized in that: The dispatching models for different power grid scenarios also include: Construct a dispatch model under the rigid supply guarantee scenario to ensure that the power grid load shedding is minimized while meeting the minimum power grid operation cost. The specific expression is as follows: min C2=f1+f2+f3+f4 Among them, C2 is the total cost of grid operation under the rigid supply guarantee scenario, f4 is the load shedding cost, is the load shedding cost coefficient, The load shedding amount.

6. The demand-side resource scheduling method based on different power grid scenario regulation according to claim 5 is characterized in that: The dispatching models for different power grid scenarios also include: Construct a dispatch model under risk resistance scenarios to reduce the peak-to-valley difference rate of the power grid and minimize the power grid operation cost. The specific expression is as follows: min C3=f1+f2+f3+f4 Among them, C3 is the total cost of grid operation under the risk resistance scenario, C is the peak-to-valley difference rate, P max and P min They are the maximum load and minimum load of the day respectively.

7. A system using the demand-side resource scheduling method based on different power grid scenario regulation as described in any one of claims 1 to 6, characterized in that: include: An acquisition module is used to obtain historical data of two indicators: load abnormality and abnormal change duration; A classification module is used to classify power grid scenarios based on the acquired historical data to obtain different power grid scenarios; The resource library establishment module is used to analyze the load curve characteristics under different power grid scenarios, obtain the regulation requirements of different power grid scenarios, and establish a callable resource library that matches different power grid scenarios; The training module is used to train and learn different power grid scenarios through the multi-layer perceptron model MLP, identify different scenarios based on evaluation indicators, and verify the effectiveness of scenario identification based on the evaluation indicator values; The adjustment module is used to match and adjust the verified different scenarios with the resource library, and achieve scenario matching and adjustment by building scheduling models and constraint models for different power grid scenarios.

8. The demand-side resource scheduling system based on different power grid scenario regulation according to claim 7 is characterized in that: The first building module is used to build a dispatching model with the lowest grid operation cost in a flexible regulation scenario; The second building module is used to build a scheduling model in the rigid supply guarantee scenario; The third building block is to construct a scheduling model for risk resistance scenarios.

9. An electronic device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the demand-side resource scheduling method based on different power grid scenarios adjustment as described in any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the demand-side resource scheduling method based on different power grid scenario adjustment as described in any one of claims 1 to 6.

Images