Method and system for evaluating peak load regulation capacity of electric power system of large field model, medium and processor

Abstract

The invention provides an electric power system peak load regulation capacity evaluation method for a large field model, and the method comprises the steps: S1, obtaining a historical data set, and calculating a corresponding index data set; s2, supplementing the index data set to a historical data set to form a peak regulation capacity evaluation case data set; s3, performing semantization on the data set to obtain a complete semantization data set # imgabs0 #; S4, inputting the complete semantization data set # imgabs0 # into a large language model to perform similar information search to obtain a richness-enhanced semantization data set # imgabs1 #; S5, decoding the richness-enhanced semantization data set # imgabs2 #, inputting the decoded richness-enhanced semantization data set # imgabs2 # into a domain model to perform data set domain conversion, and performing data set domain conversion to obtain a data set; s6, inputting the field-enhanced semantic data set # imgabs3 # into a large language model, and carrying out model enhanced training on the field-enhanced semantic data set # imgabs3 #; s7, inputting to-be-evaluated system data to obtain an appraisal evaluation result; and S8, performing semantic decoding on the result. The problems that a case prediction result of multi-source information training data is insufficient in accuracy, large in randomness and poor in professional integrating degree are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of peak load capacity assessment, and in particular to a method, system, medium and processor for assessing peak load capacity of a power system of a large domain model. Background Art

[0002] Peak load regulating capacity is the difference between the total maximum adjustable output of the peak-shaving power plant in the system and the total minimum technical output. Its main function is to maintain the balance of active power, keep the system frequency stable, and ensure that the power grid can meet the power demand during peak hours. For example, the Chinese invention application with application number CN202211631509.7 discloses a peak-shaving capacity market capacity demand assessment system and method, which includes a login authentication module, a data entry module, an uncertainty calculation module, a peak-shaving capacity calculation module and a data output module; the uncertainty calculation module calculates n load power curves, new energy curves and external power curves according to the relevant data of the historical curve; the peak-shaving capacity demand calculation module calculates the minimum output gear of the thermal power unit to achieve the expected new energy abandonment rate and the peak-shaving capacity demand of each gear according to the relevant parameters of the power system transmitted by each curve and the data entry module, and the data output module outputs the evaluation report. The peak-shaving capacity accurate assessment function of the peak-shaving capacity market is realized by using the random time series production simulation technology that considers the uncertainty of future load growth, the uncertainty of new energy power generation, and the uncertainty of external power.

[0003] However, the existing power system peak load capacity assessment methods are difficult to accurately predict and evaluate cases with complex data containing multi-source information, and the prediction results for cases containing multi-source information training data are not accurate enough. In addition, the existing large model prediction methods are highly random and poorly professional in their results of power system peak load capacity assessment. In addition, the existing peak load capacity assessment methods cannot effectively solve the problem of inaccurate data assessment in the absence of training cases.

[0004] In view of this, a method, system, medium and processor for evaluating the peak load capacity of a power system based on a large domain model are needed. Summary of the invention

[0005] In view of the problems of large randomness, poor professional fit and inaccurate evaluation in the prior art, the present invention provides a method, system, medium and processor for evaluating the peak load capacity of a power system based on a large domain model, which can solve the problems of large randomness and poor professional fit of the prediction results of a single large model; and solve the problem of inaccurate prediction evaluation in the absence of actual training cases by using a large model for case search and screening. The specific technical solution is as follows:

[0006] A method for evaluating peak load capacity of a power system based on a large domain model, comprising:

[0007] S1: Obtain a historical data set of the power system, calculate a corresponding peak load capacity evaluation index based on the historical data set, and obtain a corresponding index data set;

[0008] S2: Supplement the indicator data set to the historical data set to form a peak load capacity assessment case data set;

[0009] S3: Semantize the obtained data set of the peak-shaving capacity evaluation case and the data set of the system to be evaluated to obtain a complete semantic data set

[0010] S4: Complete semantic dataset Input into the large language model to search for similar information and screen cases to expand the number of cases and obtain a richness-enhanced semantic dataset

[0011] S5: Make the richness of the dataset semantically enhanced Perform semantic decoding, input the decoded data set into the domain model to domainize the data set, and then perform semantic encoding on the domainized data set to obtain a domain-enhanced semantic data set

[0012] S6: Enhance semantic datasets with semantic domains Input into a large language model for model reinforcement training;

[0013] S7: Input the semantic data of the system to be evaluated into the enhanced large language model to evaluate the peak demand results and obtain the evaluated results;

[0014] S8: semantically decode the evaluated evaluation results to obtain the peak-shaving capacity equivalent load curve data of the system to be evaluated as the result of the peak-shaving capacity evaluation.

[0015] Furthermore, in step S1, obtaining a historical data set of the power system, calculating a corresponding peak load capacity evaluation index according to the historical data set, and obtaining a corresponding index data set includes the following steps:

[0016] S11: Obtain the load data set, conventional output data set and wind and solar output data set of the power system. The formulas of each data set are as follows:

[0017]

[0018] In the above formula, E1(x load ) is the load data set, x load represents the variable of all types of power loads; n is the total number of time scales of data division; x tem It is the tthi Temperature data at all times; It is the tth i Light intensity data at all times; It is the tth i Wind speed data at all times; It is the tth i Active power demand data of load at all times;

[0019] E2(x gen,nom ) is the conventional output data set; x gen,nom is a variable representing conventional output data; It is the tth i Conventional unit cost data at all times; Indicates the tth i Normal active output value at the time;

[0020] E3(x gen,new ) is the wind and solar power output data set; x gen,new is a variable representing wind and solar power output data; For the tth i The wind and solar active power output data at all times; For the tth i Photovoltaic active output data at all times; For the tth i Wind power active output data at all times;

[0021] S12: Calculate the time series characteristic index of the load and obtain the daily maximum peak-to-valley difference ΔP calculated on a weekly basis load,week The peak-to-valley difference rate of the day with the largest peak-to-valley difference in the week ρ load,week , forming the peak-to-valley difference data set E4 (ΔP load,week ) and the peak-to-valley difference rate dataset E5(ρ load,week ), the calculation formula is as follows:

[0022]

[0023] In the above formula, ΔP load,week Represents the maximum peak-to-valley difference of the day based on weekly statistics; P peak,week , P valley,week Represents the peak and valley values ​​of the day with the largest peak-to-valley difference in the week; ρ load,week Represents the peak-to-valley difference rate of the day with the largest peak-to-valley difference in the week; It is the conventional active power output value from the 1st hour to the 24th hour of the day with the largest peak-to-valley difference in a week. The day with the largest peak-to-valley difference is obtained by screening the day with the largest peak-to-valley difference.

[0024] S13: Calculate the anti-peak regulation characteristic index of wind and solar output, obtain the wind and solar anti-peak regulation rate, and form the wind and solar anti-peak regulation rate data set E6 (r repeak ), the calculation formula is as follows:

[0025]

[0026] In the above formula, r repeak is the wind and solar reverse peak regulation rate; m count is the number of wind and solar reverse peak regulation; d total The total number of days for statistics;

[0027] S14: Calculate the peak-shaving capacity equivalent load curve to form a peak-shaving capacity equivalent load curve data set E7[P equ-load (d day,t )]; the peak load capacity equivalent load curve reflects the peak load capacity of the power system as a result; the calculation formula is as follows:

[0028] P equ-load (d day,t )=P load (d day,t )-P newenergy (d day,t ),d day ∈[1,7],t∈[0,23];

[0029] In the above formula, P equ-load (d day,t ) is the equivalent load power of wind and solar power connected at time t on a certain day; P load (d day,t ) is the load power at time t on a certain day; P newenergy (d day,t ) is the wind and solar power output at time t on a certain day.

[0030] Furthermore, in S2, the addition of the indicator data set to the historical data set to form a peak-shaving capacity assessment case data set refers to adding the peak-to-valley difference data set E4 (ΔP load,week ), peak-to-valley difference rate data set E5(ρ load,week ), wind and solar anti-peak rate dataset E6 (r repeak ) and peak load capacity equivalent load curve data set E7[P equ-load (d day,t )] and the load data set E1(x load ), conventional output data set E2(x gen,nom ) and wind and solar output dataset E3(x gen,new ) are combined to form the peak load capacity assessment case data set.

[0031] Furthermore, in step S3, the semantic formula is as follows:

[0032]

[0033] In the above formula, represents the kth data set in the peak load capacity assessment case data set at time t, k∈[1,7], and the 7 data sets are load data sets E1(x load ), conventional output data set E2(x gen,nom )、wind and solar power output data set E3(x gen,new ), peak-to-valley difference data set E4 (ΔP load,week ), peak-to-valley difference rate data set E5(ρ load,week ), wind and solar anti-peak rate dataset E6 (r repeak ) and peak load capacity equivalent load curve data E7[P equ-load (d day,t )], the time scale and the total number of moments are determined according to the size of the data set; is the semantic statement corresponding to the kth data set at time t; Encode() is a semantic encoder that compiles by writing simple data type names and data into simple sentences.

[0034] Furthermore, in step S4, the case screening includes the following steps:

[0035] S41: Semantic information must include the existence of E1(x load )、E2(x gen,nom ) and E3(x gen,new ) of the corresponding data, where E1(x load ) must contain Data, E2(x gen,nom ) must contain Data, E3(x gen,new ) must contain Data of at least one of the three;

[0036] S42: The total number of moments T included in the semantic information is greater than or equal to 12.

[0037] Furthermore, in step S5, the process of semantic decoding is the inverse process of semantic encoding, and the expression is as follows:

[0038]

[0039] In the above formula, Decode() is a semantic decoder, and the units of data generated by Encode() and recognized by Decode() are unified by the large language model, and the corresponding unit conversion is also performed by the large model.

[0040] Furthermore, in step S5, the prediction training principle formula of the domain model is:

[0041]

[0042] In the above formula, and is the training input data of the lth and l+1th layers; Q(l) and Q(l+1) are the adaptive training weight coefficients of the lth and l+1th layers respectively; b(l) and b(l+1) are the adaptive training bias coefficients of the lth and l+1th layers respectively; is the input peak-shaving capacity equivalent load curve training data, that is, the training output data; Softmax() is the activation function.

[0043] A power system peak-shaving capacity assessment system for a large domain model is applied to the power system peak-shaving capacity assessment method for a large domain model described above, comprising:

[0044] An index calculation module is used to obtain a historical data set of the power system, calculate a corresponding peak load capacity evaluation index based on the historical data set, and obtain a corresponding index data set;

[0045] A construction module, which is used to supplement the indicator data set to the historical data set to jointly form a peak load capacity assessment case data set;

[0046] The semantic module is used to semanticize the data set of the peak-shaving capacity assessment case and the data set of the system to be assessed to obtain a complete semantic data set.

[0047] The expansion module is used to complete the semantic dataset Input into the large language model to search for similar information and screen cases to expand the number of cases and obtain a richness-enhanced semantic dataset

[0048] Domain modules are used to enrich and semantically enhance datasets Perform semantic decoding, input the decoded data set into the domain model to domainize the data set, and then perform semantic encoding on the domainized data set to obtain a domain-enhanced semantic data set

[0049] Training module, which is used to enhance semantic datasets with semantic domains Input into a large language model for model reinforcement training;

[0050] An evaluation module is used to input semantic system data to be evaluated into the enhanced large language model to evaluate the peak demand results and obtain a reviewed evaluation result;

[0051] The decoding module is used to perform semantic decoding on the reviewed evaluation results to obtain the peak-shaving capacity equivalent load curve data of the system to be evaluated as the result of the peak-shaving capacity evaluation.

[0052] A computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the above-mentioned method for evaluating the peak-shaving capacity of a power system based on a large domain model.

[0053] A processor is used to run a program, wherein the program executes the power system peak-shaving capacity assessment method of the above-mentioned domain large model when running.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] 1. A method for evaluating the peak-shaving capacity of a power system based on a domain big model is proposed to solve the problem of insufficient accuracy of case prediction results of multi-source information training data; the language big model is combined with the domain model to solve the problem of large randomness and poor professional fit of the prediction results of a single big model; a big model is used for case search and screening to solve the problem of inaccurate prediction and evaluation in the absence of actual training cases. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0057] Figure 1 It is a flowchart of the power system peak load capacity assessment method for the first field large model;

[0058] Figure 2 The figure is a flow chart of the power system peak load capacity assessment method for the second domain large model. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0061] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0062] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0063] Embodiment 1

[0064] Using large language models for enriching training data sets and predictive evaluation tasks can quickly predict high-dimensional data and effectively deal with situations where training data is lacking. At the same time, it can simplify the difficulty of data processing and eliminate the need for manual parameter adjustment when dealing with cases with high training data complexity. It can also use domain models to assist in the training of large models, thereby improving the professional fit of large model training data and the accuracy of predictions.

[0065] like Figure 1 , Figure 2 The figure shows a flow chart of a method for evaluating the peak load capacity of a power system based on a large domain model, including:

[0066] S1: Obtain a historical data set of the power system, calculate a corresponding peak-shaving capacity evaluation index based on the historical data set, and obtain a corresponding index data set.

[0067] Furthermore, in step S1, obtaining a historical data set of the power system, calculating a corresponding peak load capacity evaluation index according to the historical data set, and obtaining a corresponding index data set includes the following steps:

[0068] S11: Obtain the load data set, conventional output data set and wind and solar output data set of the power system. The formulas of each data set are as follows:

[0069]

[0070] In the above formula, E1(x load ) is the load data set, x load represents the variable of all types of power loads; n is the total number of time scales of data division; x tem It is the tth i Temperature data at all times; It is the tth i Light intensity data at all times; It is the tth i Wind speed data at all times; It is the tth iActive power demand data of load at all times;

[0071] E2(x gen,nom ) is the conventional output data set; x gen,nom is a variable representing conventional output data; It is the tth i Conventional unit cost data at all times; Indicates the tth i Normal active output value at the time;

[0072] E3(x gen,new ) is the wind and solar power output data set; x gen,new is a variable representing wind and solar power output data; For the tth i The wind and solar active power output data at all times; For the tth i Photovoltaic active output data at all times; For the tth i Wind power active output data at all times.

[0073] S12: Calculate the time series characteristic index of the load. The time series characteristic of the load includes two indicators: peak-to-valley difference and peak-to-valley difference rate. The peak-to-valley difference and peak-to-valley difference rate of the maximum load peak-to-valley difference day in each week are used to measure the maximum peak-to-valley difference of the load in each week and the proportion of the peak-to-valley difference of the day to the load peak value, and the maximum peak-to-valley difference of the day in the weekly statistics is obtained. load,week The peak-to-valley difference rate of the day with the largest peak-to-valley difference in the week ρ load,week , forming the peak-to-valley difference data set E4 (ΔP load,week ) and the peak-to-valley difference rate dataset E5(ρ load,week ), the calculation formula is as follows:

[0074]

[0075] In the above formula, ΔP load,week Represents the maximum peak-to-valley difference of the day based on weekly statistics; P peak,week , P valley,week Represents the peak and valley values ​​of the day with the largest peak-to-valley difference in the week; ρ load,week Represents the peak-to-valley difference rate of the day with the largest peak-to-valley difference in the week; It is the conventional active power output value from the 1st hour to the 24th hour on the day with the largest peak-to-valley difference in a week. The maximum peak-to-valley difference day is obtained by screening the day with the largest peak-to-valley difference.

[0076] S13: Calculate the anti-peak regulation characteristic index of wind and solar output, obtain the wind and solar anti-peak regulation rate, and form the wind and solar anti-peak regulation rate data set E6 (r repeak ), the calculation formula is as follows:

[0077]

[0078] In the above formula, r repeak is the wind and solar reverse peak regulation rate; m count is the number of wind and solar reverse peak regulation; d total is the total number of days counted; the number of wind and solar reverse peak regulation times m count The calculation of the conventional output peak-shaving rate and the wind and solar output peak-shaving rate must be calculated first; the expression of the conventional output peak-shaving rate is:

[0079]

[0080] Among them, r load (d day ) is the normal output peak rate for 24 hours a day; P load,max (d day ) is the maximum power of the load day; P load,min (d day ) is the minimum power of the load day; is the normal active power output value from the 1st hour to the 24th hour of the day; the expression of wind and solar power output peak regulation rate is:

[0081]

[0082] Among them, r ne (d day ) is the peak load rate of wind and solar power output in 24 hours a day; P ne,max (d day ) is the maximum power of wind and solar power output in one day; P ne,min (d day ) is the minimum power of wind and solar power output in one day; and They are the photovoltaic active output and wind power active output data from the 1st hour to the 24th hour of the day;

[0083] After calculating the conventional output peak-shaving rate and wind and solar output peak-shaving rate of each day, if the wind and solar output peak-shaving rate is greater than the conventional output peak-shaving rate, the day is an anti-peak-shaving day, that is, r equ-load (d day )>r load (d day ), then the number of wind and solar reverse peak regulation times m count The value of is increased by one.

[0084] S14: Calculate the peak-shaving capacity equivalent load curve to form a peak-shaving capacity equivalent load curve data set E7[P equ-load (d day,t)]; The peak-shaving capacity equivalent load curve reflects the peak-shaving capacity of the power system as a result. Among all the output curves, the curve with the largest wind-solar anti-peak-shaving rate index value obtained by S13 is selected as the time series process curve with the most obvious anti-peak-shaving characteristics, and the curve with the largest daily maximum peak-to-valley difference index value obtained by S12 is selected as the output curve with the maximum load peak-to-valley difference day; the time series process curve with the most obvious anti-peak-shaving characteristics is selected and combined with the output curve with the maximum load peak-to-valley difference day, and the equivalent load time series process curve obtained is the peak-shaving capacity equivalent load curve of the power grid; the calculation formula for the equivalent load of the peak-shaving capacity with one hour as a period for seven days is as follows:

[0085] P equ-load (d day,t )=P load (d day,t )-P newenergy (d day,t ),d day ∈[1,7],t∈[0,23];

[0086] In the above formula, P equ-load (d day,t ) is the equivalent load power of wind and solar power connected at time t on a certain day; P load (d day,t ) is the load power at time t on a certain day; P newenergy (d day,t ) is the wind and solar power output at time t on a certain day.

[0087] The most obvious anti-peaking characteristic is r repeak The maximum time series process curve and the maximum peak-to-valley difference of the load is ΔP load,week The maximum daily output curve corresponds to the normal output data E2 (x gen,nom ) and wind and solar output data E3(x gen,new ) can be calculated to obtain P equ-load (d day,t ) data composed of equivalent load timing process curve.

[0088] S2: Supplement the indicator dataset to the historical dataset to form a peak-shaving capacity assessment case dataset.

[0089] Furthermore, in S2, the addition of the indicator data set to the historical data set to form a peak-shaving capacity assessment case data set refers to adding the peak-to-valley difference data set E4 (ΔP load,week ), peak-to-valley difference rate data set E5(ρ load,week ), wind and solar anti-peak rate dataset E6 (r repeak ) and peak load capacity equivalent load curve data set E7[P equ-load (d day,t )] and the load data set E1(xload ), conventional output data set E2(x gen,nom ) and wind and solar output dataset E3(x gen,new ) are combined to form the peak load capacity assessment case data set.

[0090] The data set of the system to be evaluated is the input data for peak load capacity evaluation. The data set types include load data set E1(x load ), conventional output data set E2(x gen,nom ) and wind and solar output dataset E3(x gen,new ).

[0091] S3: Semantize the obtained data set of the peak-shaving capacity evaluation case and the data set of the system to be evaluated to obtain a complete semantic data set

[0092] Furthermore, in step S3, the semantic formula is as follows:

[0093]

[0094] In the above formula, represents the kth data set in the peak load capacity assessment case data set at time t, k∈[1,7], and the 7 data sets are load data sets E1(x load ), conventional output data set E2(x gen,nom )、wind and solar power output data set E3(x gen,new ), peak-to-valley difference data set E4 (ΔP load,week ), peak-to-valley difference rate data set E5(ρ load,week ), wind and solar anti-peak rate dataset E6 (r repeak ) and peak load capacity equivalent load curve data E7[P equ-load (d day,t )], the time scale and the total number of moments are determined according to the size of the data set; is the semantic statement corresponding to the kth data set at time t; Encode() is a semantic encoder that compiles by writing simple data type names and data into simple sentences.

[0095] Load data Let's take an example to explain. Can be semantically expressed as "the light intensity at the second moment is x solar,2 Lux, the wind speed at the second moment is x wind,2 m / s, the temperature at the second moment is x tem,2 ℃, the load active demand at the second moment is x load,2 kW"; the semanticized data is input into the large language model; the entire case data set After semanticization, we get the complete semantic data set of this case. Where T is the total number of moments.

[0096] S4: Complete semantic dataset Input into the large language model to search for similar information and screen cases to expand the number of cases and obtain a richness-enhanced semantic dataset

[0097] This step uses the advantages of large language models trained on massive data to improve the accuracy of training data for power system peak capacity assessment.

[0098] Furthermore, in step S4, the case screening includes the following steps:

[0099] S41: Semantic information must include the existence of E1(x load )、E2(x gen,nom ) and E3(x gen,new ) of the corresponding data, where E1(x load ) must contain Data, E2(x gen,nom ) must contain Data, E3(x gen,new ) must contain Data of at least one of the three;

[0100] S42: The total number of moments T included in the semantic information is greater than or equal to 12.

[0101] S5: Make the richness of the dataset semantically enhanced Perform semantic decoding, input the decoded data set into the domain model to domainize the data set, and then perform semantic encoding on the domainized data set to obtain a domain-enhanced semantic data set

[0102] Furthermore, in step S5, the process of semantic decoding is the inverse process of semantic encoding, and the expression is as follows:

[0103]

[0104] In the above formula, Decode() is a semantic decoder. The units of data generated by Encode() and recognized by Decode() are unified by the large language model. The corresponding unit conversion is also performed by the large model, which can convert "the cost of conventional units at the third moment is x cost,3 Ten thousand yuan, the conventional active power output at the third moment is kW" is converted to At the same time, data sets of the same type at different times are integrated.

[0105] Furthermore, in step S5, the data set is domainized, that is, the peak load capacity equivalent load curve data E7[P equ-load (d day,t )] to make predictions; the dataset after semantic decoding of the original case is input into the domain model for training. The prediction training principle formula of the domain model is:

[0106]

[0107] In the above formula, and is the training input data of the lth and l+1th layers; Q(l) and Q(l+1) are the adaptive training weight coefficients of the lth and l+1th layers respectively; b(l) and b(l+1) are the adaptive training bias coefficients of the lth and l+1th layers respectively; is the input peak-shaving capacity equivalent load curve training data, that is, the training output data; Softmax() is the activation function.

[0108] After the domain model training is completed, the new case dataset after semantic decoding in the richness enhanced semantic dataset is input. For each new case, the peak load capacity equivalent load curve data E7[P equ-load (d day,t )] to make predictions;

[0109] After the equivalent load curve data is predicted, the non-semantic domain-enhanced semantic data set can be integrated. After that, the semantic formula in step S3 is used to Perform semanticization to obtain semantic domain-enhanced semantic datasets

[0110] S6: Enhance semantic datasets with semantic domains Input into the large language model for model reinforcement training.

[0111] S7: Input the semantic system data to be evaluated into the enhanced large language model to evaluate the peak demand results and obtain the evaluated results; the semantic peak-to-valley difference data set W4 (ΔP load,week ), semantic peak-to-valley difference rate dataset W5(ρ load,week ), semantic wind and solar anti-peak rate dataset W6 (r repeak ) and semantic peak load capacity equivalent load curve data W7[P equ-load (d day,t )] as the evaluation result of the large language model output.

[0112] S8: Using the semantic decoding formula in step S5, semantic decoding is performed on the evaluated evaluation results to obtain the peak-shaving capacity equivalent load curve data of the system to be evaluated as the result of the peak-shaving capacity evaluation.

[0113] Compared with the prior art, this solution has the following advantages and effects:

[0114] (1) For the case where the prediction results of cases containing multi-source information training data are not accurate enough, the domain large model method can effectively capture the relationship between multiple factors, and at the same time can use semantic encoders and decoders to convert effective time series data into semantic information for training.

[0115] (2) In view of the fact that the results of power system peak-shaving capacity assessment using a single large model prediction method are highly random and have poor professional fit, the domain model is used to assist the large model training data to process the training data from the large model, thereby improving the effectiveness and professional fit of the training data set. At the same time, the domain model can adapt to cases containing various types of data, solving the problem that some cases cannot directly solve the peak-shaving capacity assessment indicators.

[0116] (3) For data sets that lack actual training cases, a large model is used to search and screen training case data to effectively extract prior information and improve the richness of the model training data set.

[0117] Embodiment 2

[0118] A power system peak-shaving capacity assessment system for a large domain model is applied to the power system peak-shaving capacity assessment method for a large domain model described above, comprising:

[0119] An index calculation module is used to obtain a historical data set of the power system, calculate a corresponding peak load capacity evaluation index based on the historical data set, and obtain a corresponding index data set;

[0120] A construction module, which is used to supplement the indicator data set to the historical data set to jointly form a peak load capacity assessment case data set;

[0121] The semantic module is used to semanticize the data set of the peak-shaving capacity assessment case and the data set of the system to be assessed to obtain a complete semantic data set.

[0122] The expansion module is used to complete the semantic dataset Input into the large language model to search for similar information and screen cases to expand the number of cases and obtain a richness-enhanced semantic dataset

[0123] Domain modules are used to enrich and semantically enhance datasets Perform semantic decoding, input the decoded data set into the domain model to domainize the data set, and then perform semantic encoding on the domainized data set to obtain a domain-enhanced semantic data set

[0124] Training module, which is used to enhance semantic datasets with semantic domains Input into a large language model for model reinforcement training;

[0125] An evaluation module is used to input semantic system data to be evaluated into the enhanced large language model to evaluate the peak demand results and obtain a reviewed evaluation result;

[0126] The decoding module is used to perform semantic decoding on the reviewed evaluation results to obtain the peak-shaving capacity equivalent load curve data of the system to be evaluated as the result of the peak-shaving capacity evaluation.

[0127] Embodiment 3

[0128] A computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the above-mentioned method for evaluating the peak-shaving capacity of a power system based on a large domain model.

[0129] Embodiment 4

[0130] A processor is used to run a program, wherein the program executes the power system peak-shaving capacity assessment method of the above-mentioned domain large model when running.

[0131] The present application proposes a method for evaluating the peak-shaving capacity of a power system based on a large domain model, including: S1: obtaining a historical data set of the power system, calculating corresponding peak-shaving capacity evaluation indicators based on the historical data set, and obtaining a corresponding indicator data set; S2: adding the indicator data set to the historical data set to jointly constitute a peak-shaving capacity evaluation case data set; S3: semantically translating the obtained data set of the peak-shaving capacity evaluation case and the data set of the system to be evaluated to obtain a complete semantic data set S4: Complete semantic dataset Input into the large language model to search for similar information and screen cases to expand the number of cases and obtain a richness-enhanced semantic dataset S5: Make the richness of the dataset semantically enhanced Perform semantic decoding, input the decoded data set into the domain model to domainize the data set, and then perform semantic encoding on the domainized data set to obtain a domain-enhanced semantic data set S6: Enhance semantic datasets with semantic domains Input into the big language model for model reinforcement training; S7: Input semantic data of the system to be evaluated into the enhanced big language model to evaluate the peak demand results and obtain the evaluated evaluation results; S8: Semantically decode the evaluated evaluation results to obtain the peak capacity equivalent load curve data of the system to be evaluated as the result of the peak capacity evaluation. Solve the problem of insufficient accuracy of case prediction results of multi-source information training data; Apply the method of combining language big model with domain model to solve the problem of large randomness and poor professional fit of the prediction results of a single big model; Use big model to search and screen cases to solve the problem of inaccurate prediction and evaluation in the absence of actual training cases.

[0132] Those of ordinary skill in the art will appreciate that the units of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition of each example has been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0133] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored, etc.

[0134] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0135] If the integrated unit is implemented in the form of 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 solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-0nlyMemory), random access memory (RAM, RandomAccessMemory), mobile hard disk, magnetic disk or optical disk, etc., which can store program code.

[0136] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A method for evaluating the peak load capacity of a power system based on a large domain model, characterized in that: include: S1: Obtain a historical data set of the power system, calculate a corresponding peak load capacity evaluation index based on the historical data set, and obtain a corresponding index data set; S2: Supplement the indicator data set to the historical data set to form a peak load capacity assessment case data set; S3: Semantize the obtained data set of the peak-shaving capacity evaluation case and the data set of the system to be evaluated to obtain a complete semantic data set S4: Complete semantic dataset Input into the large language model to search for similar information and screen cases to expand the number of cases and obtain a richness-enhanced semantic dataset S5: Make the richness of the dataset semantically enhanced Perform semantic decoding, input the decoded data set into the domain model to domainize the data set, and then perform semantic encoding on the domainized data set to obtain a domain-enhanced semantic data set S6: Enhance semantic datasets with semantic domains Input into a large language model for model reinforcement training; S7: Input the semantic data of the system to be evaluated into the enhanced large language model to evaluate the peak demand results and obtain the evaluated results; S8: semantically decode the evaluated evaluation results to obtain the peak-shaving capacity equivalent load curve data of the system to be evaluated as the result of the peak-shaving capacity evaluation.

2. The method for evaluating peak load capacity of a power system based on a large domain model according to claim 1, characterized in that: In step S1, the process of obtaining a historical data set of the power system, calculating a corresponding peak load capacity evaluation index according to the historical data set, and obtaining a corresponding index data set includes the following steps: S11: Obtain the load data set, conventional output data set and wind and solar output data set of the power system. The formulas of each data set are as follows: In the above formula, E1(x load ) is the load data set, x load represents the variable of all types of power loads; n is the total number of time scales of data division; x tem It is the tth i Temperature data at all times; It is the tth i Light intensity data at all times; It is the tth i Wind speed data at all times; It is the tth i Active power demand data of load at all times; E2(x gen,nom ) is the conventional output data set; x gen,nom is a variable representing conventional output data; It is the tth i Conventional unit cost data at all times; Indicates the tth i Normal active output value at the time; E3(x gen,new ) is the wind and solar power output data set; x gen,new is a variable representing wind and solar power output data; For the tth i The wind and solar active power output data at all times; For the tth i Photovoltaic active output data at all times; For the tth i Wind power active output data at all times; S12: Calculate the time series characteristic index of the load and obtain the daily maximum peak-to-valley difference ΔP calculated on a weekly basis load,week The peak-to-valley difference rate of the day with the largest peak-to-valley difference in the week ρ load,week , forming the peak-to-valley difference data set E4 (ΔP load,week ) and the peak-to-valley difference rate dataset E5(ρ load,week ), the calculation formula is as follows: In the above formula, ΔP load,week Represents the maximum peak-to-valley difference of the day based on weekly statistics; P peak,week , P valley,week Represents the peak and valley values ​​of the day with the largest peak-to-valley difference in the week; ρ load,week Represents the peak-to-valley difference rate of the day with the largest peak-to-valley difference in the week; It is the conventional active power output value from the 1st hour to the 24th hour of the day with the largest peak-to-valley difference in a week. The day with the largest peak-to-valley difference is obtained by screening the day with the largest peak-to-valley difference. S13: Calculate the anti-peak regulation characteristic index of wind and solar output, obtain the wind and solar anti-peak regulation rate, and form the wind and solar anti-peak regulation rate data set E6 (r repeak ), the calculation formula is as follows: In the above formula, r repeak is the wind and solar reverse peak regulation rate; m count is the number of wind and solar reverse peak regulation; d total The total number of days for statistics; S14: Calculate the peak-shaving capacity equivalent load curve to form a peak-shaving capacity equivalent load curve data set E7[P equ-load (d day,t )]; the peak load capacity equivalent load curve reflects the peak load capacity of the power system as a result; the calculation formula is as follows: P equ-load (d day,t )=P load (d day,t )-P newenergy (d day,t ),d day ∈[1,7],t∈[0,23]; In the above formula, P equ-load (d day,t ) is the equivalent load power of wind and solar power connected at time t on a certain day; P load (d day,t ) is the load power at time t on a certain day; P newenergy (d day,t ) is the wind and solar power output at time t on a certain day.

3. The method for evaluating peak load capacity of a power system based on a large domain model according to claim 2, characterized in that: In S2, the indicator data set is added to the historical data set to form a peak-shaving capacity assessment case data set, which means that the peak-to-valley difference data set E4 (ΔP load,week ), peak-to-valley difference rate data set E5(ρ load,week ), wind and solar anti-peak rate dataset E6 (r repeak ) and peak load capacity equivalent load curve data set E7[P equ-load (d day,t )] and the load data set E1(x load ), conventional output data set E2(x gen ,nom) and wind and solar output dataset E3(x gen,new ) are combined to form the peak load capacity assessment case data set.

4. The method for evaluating peak load capacity of a power system based on a large domain model according to claim 2, characterized in that: In step S3, the semantic formula is as follows: In the above formula, represents the kth data set in the peak load capacity assessment case data set at time t, k∈[1,7], and the 7 data sets are load data sets E1(x load ), conventional output data set E2(x gen,nom )、wind and solar power output data set E3(x gen,new ), peak-to-valley difference data set E4 (ΔP load,week ), peak-to-valley difference rate data set E5(ρ load,week ), wind and solar anti-peak rate dataset E6 (r repeak ) and peak load capacity equivalent load curve data E7[P equ-load (d day,t )], the time scale and the total number of moments are determined according to the size of the data set; is the semantic statement corresponding to the kth data set at time t; Encode() is a semantic encoder that compiles by writing simple data type names and data into simple sentences.

5. The method for evaluating peak load capacity of a power system based on a large domain model according to claim 2, characterized in that: In step S4, the case screening includes the following steps: S41: Semantic information must include the existence of E1(x load )、E2(x gen,nom ) and E3(x gen,new ) of the corresponding data, where E1(x load ) must contain Data, E2(x gen,nom ) must contain Data, E3(x gen,new ) must contain Data of at least one of the three; S42: The total number of moments T included in the semantic information is greater than or equal to 12.

6. The method for evaluating peak load capacity of a power system based on a large domain model according to claim 4, characterized in that: In step S5, the process of semantic decoding is the inverse process of semantic encoding, and the expression is as follows: In the above formula, Decode() is a semantic decoder, and the units of data generated by Encode() and recognized by Decode() are unified by the large language model, and the corresponding unit conversion is also performed by the large model.

7. The method for evaluating peak load capacity of a power system based on a large domain model according to claim 4, characterized in that: In step S5, the prediction training principle formula of the domain model is: In the above formula, and is the training input data for the lth and l+1th layers; Q(l) and Q(l+1) are the adaptive training weight coefficients for the lth and l+1th layers respectively; b(l) and b(l+1) are the adaptive training bias coefficients of the lth and l+1th layers respectively; is the input peak-shaving capacity equivalent load curve training data, that is, the training output data; Softmax() is the activation function.

8. A power system peak load capacity assessment system for a large domain model, characterized in that: A method for evaluating peak load capacity of a power system applied to a large domain model as described in any one of claims 1 to 7, comprising: An index calculation module is used to obtain a historical data set of the power system, calculate a corresponding peak load capacity evaluation index based on the historical data set, and obtain a corresponding index data set; A construction module, which is used to supplement the indicator data set to the historical data set to jointly form a peak load capacity assessment case data set; The semantic module is used to semanticize the data set of the peak-shaving capacity assessment case and the data set of the system to be assessed to obtain a complete semantic data set. The expansion module is used to complete the semantic dataset Input into the large language model to search for similar information and screen cases to expand the number of cases and obtain a richness-enhanced semantic dataset Domain modules are used to enrich and semantically enhance datasets Perform semantic decoding, input the decoded data set into the domain model to domainize the data set, and then perform semantic encoding on the domainized data set to obtain a domain-enhanced semantic data set Training module, which is used to enhance semantic datasets with semantic domains Input into a large language model for model reinforcement training; An evaluation module is used to input semantic system data to be evaluated into the enhanced large language model to evaluate the peak demand results and obtain a reviewed evaluation result; The decoding module is used to perform semantic decoding on the reviewed evaluation results to obtain the peak-shaving capacity equivalent load curve data of the system to be evaluated as the result of the peak-shaving capacity evaluation.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method for evaluating the peak-shaving capacity of a power system based on a domain large model as described in any one of claims 1 to 7.

10. A processor, characterized in that: The processor is used to run a program, wherein the program, when running, executes the method for evaluating the peak-shaving capacity of a power system based on a large domain model as described in any one of claims 1 to 7.

Images