Construction method of multi-level dynamic linkage power purchase transaction evaluation model

Abstract

The invention discloses a method for constructing a multi-level dynamic linkage electricity purchase transaction evaluation model, which relates to the field of electric power technology and comprises the following steps: S1: dividing and processing the electricity purchase structure; S2: collecting historical electricity consumption data of electricity purchasers, and establishing risk indicators according to the historical electricity consumption data; S3: collecting real-time electricity purchase data in the current electricity purchase structure, normalizing the multi-level dynamic real-time electricity purchase data, and obtaining normalized electricity purchase vectors and data analysis instructions; establishing a transaction evaluation model through the historical electricity purchase data, dividing the transaction evaluation model into a training set and a test set, so that the transaction evaluation model obtained by training has the advantages of a support vector regression machine and a principal component analysis method, reduces prediction errors, enhances the generalization ability and stability of the electricity purchase transaction model, and obtains the electricity price according to the electricity purchase transaction model, thereby ensuring the economy of the power supply system.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and in particular to a method for constructing a multi-level dynamically linked electricity purchase transaction evaluation model. Background Art

[0002] The power purchasing business is part of the main business operation of the power grid company. The power purchasing cost is an important cost of the company, accounting for about 70% of the company's operating costs. In the process of carrying out the power purchasing business, the power purchasing cost will be affected by various internal and external factors such as changes in electricity demand, unit maintenance and adjustment, three public dispatches, electricity price adjustments, and implementation of energy-saving and emission reduction policies. The power grid company carries out the power purchasing business on the premise of ensuring reliable power supply, conducts transactions in accordance with the law, safeguards the legitimate rights and interests of both buyers and sellers, and ensures the company's reasonable profits.

[0003] When evaluating power purchase transactions in the power purchase market, due to the complex structure of the power purchase business, the company's profits are largely affected by the power purchase cost. When users purchase electricity, they mainly face the risk of price fluctuations, such as fluctuations in electricity prices caused by changes in market supply and demand, etc., which makes it impossible to guarantee the economy of the power supply system. Therefore, a method for constructing a multi-level dynamic linkage power purchase transaction evaluation model is provided to solve the problems raised in the above background technology. Summary of the invention

[0004] The purpose of the present invention is to provide a method for constructing a multi-level dynamically linked electricity purchase transaction evaluation model to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] 1. A method for constructing a multi-level dynamic linkage power purchase transaction evaluation model, comprising the following steps:

[0007] S1: Divide the electricity purchase structure;

[0008] S2: Collect historical electricity consumption data and multi-level dynamic real-time electricity purchase data of electricity purchasers, and establish constraints on independent variables and dependent variables based on the historical electricity consumption data and multi-level dynamic real-time electricity purchase data;

[0009] S3: normalize the historical electricity consumption data and the multi-level dynamic real-time electricity purchase data to obtain the normalized electricity purchase vector and data analysis instructions;

[0010] S4: Based on the data collected by S2 and S3, calculate the type of electricity purchased and summarize the benchmark values ​​of electricity purchase amount, electricity price and electricity fee;

[0011] S5: Determine the constraint conditions for adjusting the annual utilization hours of thermal power trading units;

[0012] S6: Input the adjustment amount of the independent variable;

[0013] S7: Adjust the purchased electricity quantity, electricity price and electricity fee according to the benchmark value, calculate the evaluation value of each type of electricity purchase, and determine whether the evaluation value meets all constraints. If the constraints are not met, return to S5.

[0014] As a further solution of the present invention: wherein, the historical electricity consumption data includes electricity consumption, electricity purchase volume, market supply and demand electricity cost and user participation; the multi-level dynamic real-time electricity purchase data includes electricity purchase volume data, electricity price data and electricity fee data, and the data is collected, entered, analyzed and processed.

[0015] As a further solution of the present invention: wherein, the specific parameters of the public opinion risk index, market supply and demand risk index, price fluctuation risk index, compliance risk index and supply guarantee risk index can be obtained through user questionnaires, and then through statistical quantification, the specific historical data of historical electricity consumption data can be obtained.

[0016] As a further solution of the present invention: wherein, when the electricity purchasing structure in S1 is divided, the MECE principle should be followed, which has the functions of being independent of each other and not duplicative.

[0017] As a further solution of the present invention: Among them, the MECE principle can ensure the comprehensiveness and accuracy of problem analysis. The MECE principle requires that when analyzing a problem, it should be split into several sub-problems that are neither overlapping nor comprehensive, so as to avoid omissions and duplications when dividing the electricity purchasing structure.

[0018] As a further solution of the present invention: wherein the amount of electricity purchased includes the electricity purchased by the municipal company, the electricity purchased by the non-municipal company and the electricity purchased by the county-level company, the electricity price benchmark value is divided by the amount of electricity purchased to obtain (the electricity purchased by the municipal company):

[0019] Q bas111 =CA 1k *t 1k *(1-β 1k );

[0020] Where: Q bas111 is the base value of electricity purchase for each type of self-generated power in service, CA 1K is the installed capacity of each thermal power trading unit, t 1k is the annual utilization hours (k = 1, 2, 3, m), β 1k It is the power consumption rate of the factory.

[0021] As a further solution of the present invention: wherein, the base price of electricity purchased by the out-of-city company is: Q bas2 =Q bas21 +Q bas22 ;

[0022] Where: Q bas2 is the benchmark value of electricity purchased from outside the city, Q bas21 is the benchmark value of long-term contractual electricity purchase, Q bas22 It is the accurate value of electricity purchase amount for short-term transactions.

[0023] As a further solution of the present invention: wherein the base value of the electricity purchase price of the county-level company is: F bas2 =Q bas3 +P bas3 ;

[0024] Where: Q bas3 is the benchmark value of electricity purchase by county-level companies, P bas3 is the average electricity price benchmark of county-level companies, Q bas22 It is the accurate value of electricity purchase amount for short-term transactions.

[0025] As a further solution of the present invention: wherein, the historical electricity purchase data in S2 is obtained to establish a transaction evaluation model, the transaction evaluation model is divided into a training set and a test set, and the training set and the test set are respectively normalized and preprocessed by dimensionality reduction.

[0026] As a further solution of the present invention: wherein, the ratio of training set to test set can be 7:3, and the training set and the test set can be allocated by the fold cross validation method. This method enables the trained transaction evaluation model to have the advantages of both support vector regression machine and principal component analysis method, reduce prediction errors, and enhance the generalization ability and stability of the model.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] 1. This method can ensure the comprehensiveness and accuracy of problem analysis by following the MECE principle, splitting it into several sub-problems that are neither overlapping nor comprehensive, avoiding omissions and duplications when dividing the power purchase structure, and thus establishing a multi-level power purchase transaction structure.

[0029] 2. By setting up public opinion risk indicators including user policy awareness indicators, we can fully understand the market public opinion involving the public's discussion on electricity prices, avoid excessive excess electricity, which leads to most fluctuations in electricity prices and increases in electricity risk costs. In addition, a transaction evaluation model is established through historical electricity purchase data, and the transaction evaluation model is divided into a training set and a test set. The trained transaction evaluation model has the advantages of both support vector regression and principal component analysis, reduces prediction errors, and enhances the generalization ability and stability of the electricity purchase transaction model. The electricity price is obtained according to the electricity purchase transaction model, thereby ensuring the economy of the power supply system. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a schematic diagram of the structure flow of the power purchase transaction evaluation model in the present invention;

[0031] Figure 2 This is a schematic diagram of historical electricity consumption data collection in the present invention;

[0032] Figure 3 This is a schematic diagram of the expected value of the electricity price of the out-of-city companies in the present invention;

[0033] Figure 4 A schematic diagram of the expected value of electricity price of the municipal company in the present invention;

[0034] Figure 5 This is a schematic diagram of the expected value of electricity prices for county-level companies in the present invention. DETAILED DESCRIPTION

[0035] See also Figures 1 to 5 :A method for constructing a multi-level dynamic linkage power purchase transaction evaluation model includes the following steps:

[0036] S1: Divide the electricity purchase structure;

[0037] S2: Collect historical electricity consumption data and multi-level dynamic real-time electricity purchase data of electricity purchasers, and establish constraints on independent variables and dependent variables based on the historical electricity consumption data and multi-level dynamic real-time electricity purchase data;

[0038] Among them, when performing the division processing, the historical electricity consumption data of electricity purchasing users and the multi-level dynamic real-time electricity purchasing data are first summarized and decomposed, and then the model parameters are obtained.

[0039] The collected multi-level dynamic real-time electricity purchase data and historical electricity consumption data contain relevant data such as power generation, electricity consumption, electricity purchase, market supply and demand, electricity cost, and user participation.

[0040] S3: normalize the historical electricity consumption data and the multi-level dynamic real-time electricity purchase data to obtain the normalized electricity purchase vector and data analysis instructions;

[0041] S4: Based on the data collected by S2 and S3, calculate the type of electricity purchased and summarize the benchmark values ​​of electricity purchase amount, electricity price and electricity fee;

[0042] Among them, the parameters used by the model will be obtained in the process of decomposing and summarizing the purchased electricity amount, price and fee. The parameters are divided into three categories: constants, independent variables and dependent variables. The constants are fixed values.

[0043] Constants, independent variables and dependent variables are usually determined based on annual plan values ​​or business status, thereby adjusting the three numerical parameters of the independent and dependent variables: baseline values, adjustment amounts and evaluation values.

[0044] S5: Determine the constraint conditions for adjusting the annual utilization hours of the thermal power trading unit;

[0045] Among them, the constraint conditions refer to the calculation and analysis of the number of days that the power generation is affected by the maintenance of the coal-fired units, the remaining adjustable power generation, and the upper and lower limits of the annual utilization hour adjustment.

[0046] S6: Input the adjustment amount of the independent variable;

[0047] Among them, the independent variable is the quantity that changes due to external factors during the power purchase transaction evaluation process; such as (policy adjustments, changes in market supply and demand, etc.), the change of the independent variable will affect the evaluation of the power purchase volume, electricity price and electricity bill.

[0048] S7: Adjust the purchased electricity quantity, electricity price and electricity fee according to the benchmark value, calculate the evaluation value of each type of electricity purchase, and determine whether all constraints are met.

[0049] Among them, the benchmark value adjustment is input according to certain constraints, and the evaluation value is the sum of the benchmark value and the adjustment amount. When the adjustment amount is the fluctuation range, the evaluation value is equal to the benchmark value plus the product of the benchmark value and the adjustment amount.

[0050] The baseline value of the above-mentioned dependent variable is calculated by using a fixed calculation formula from the baseline value of the constant and the independent variable, and the evaluation value is calculated by using a fixed calculation formula from the evaluation value of the constant and the independent variable.

[0051] The historical electricity consumption data includes electricity consumption, electricity purchase amount, market supply and demand electricity cost and user participation; the multi-level dynamic real-time electricity purchase data includes electricity purchase amount data, electricity price data and electricity fee data, and the data is collected, entered, analyzed and processed.

[0052] Among them, the electricity purchase data includes the electricity purchase amount of each type of electricity purchase, the total electricity purchase amount, and the proportion of each type of electricity purchase in the total electricity purchase amount. These data reflect the scale and structure of electricity purchases by the power grid company;

[0053] Electricity price data: including electricity prices of various types of electricity purchases, average electricity prices, and electricity price trends; these data reflect the power grid company's electricity purchase costs and market price levels;

[0054] Electricity fee data: including electricity fees for each type of electricity purchase, total electricity fees, and changes in electricity fees, etc. These data reflect the power purchase expenditure and economic benefits of the power grid company.

[0055] Data collection: By installing smart meters, using sensors and other technical means, the power grid company's electricity purchase amount, electricity price, electricity charges and other data are collected in real time.

[0056] Data entry: Enter the collected data into the information system of the power grid company to establish a complete power purchase transaction database;

[0057] Data analysis: Use data analysis tools and methods to sort, analyze and calculate the input data to obtain multi-level dynamic real-time electricity purchase data;

[0058] Through multi-level dynamic real-time electricity purchase data, power grid companies can improve the scientific nature of electricity purchase decisions, reduce electricity purchase costs and strengthen risk management during electricity purchase transactions.

[0059] When dividing the electricity purchase structure in S1, the MECE principle should be followed, which has independent and non-repetitive effects.

[0060] Among them, the MECE principle can ensure the comprehensiveness and accuracy of problem analysis. The MECE principle requires that when analyzing a problem, it should be split into several sub-problems that are neither overlapping nor comprehensive, so as to avoid omissions and duplications when dividing the electricity purchasing structure.

[0061] The amount of electricity purchased includes the electricity purchased by the municipal company, the electricity purchased by the company outside the city and the electricity purchased by the county-level company. The base value of the electricity price is available at the amount of electricity purchased (the electricity purchased by the municipal company):

[0062] Q bas111 =CA 1k *t 1k *(1-β 1k );

[0063] Where: Q bas111 is the base value of electricity purchase for each type of self-generated power in service, CA 1K is the installed capacity of each thermal power trading unit, t 1k is the annual utilization hours (k=1, 2, 3...m), β 1k It is the power consumption rate of the factory.

[0064] Q bas111 Including thermal power, hydropower, wind power, solar power, biomass, nuclear power, tidal, geothermal and other power generation equipment.

[0065] Among them, the specific electricity purchase amount benchmark value of the municipal company can be calculated according to the above formula.

[0066] The base price for electricity purchased by companies outside the city is:

[0067] Q bas2 =Q bas21 +Q bas22 ;

[0068] Where: Q bas2 is the benchmark value of electricity purchased from outside the city, Q bas21 is the benchmark value of long-term contractual electricity purchase, Q bas22 It is the accurate value of electricity purchase amount for short-term transactions.

[0069] Among them, the benchmark value of electricity purchase amount for companies outside the city can be calculated according to the above formula.

[0070] The base price for electricity purchased by county-level companies is:

[0071] F bas3 =Q bas3 +P bas3 ;

[0072] Where: Q bas3 is the benchmark value of electricity purchase by county-level companies, P bas3 is the average electricity price benchmark of county-level companies, Q bas22 It is the accurate value of electricity purchase amount for short-term transactions.

[0073] Among them, the benchmark value of electricity purchase amount for county-level companies can be calculated according to the above formula.

[0074] Among them, the installed capacity, plant power consumption rate, annual utilization hours and average power purchase price benchmark of the above-mentioned power generation types can be used to determine the long-term and short-term contract power purchase amount, capacity electricity fee, power quantity and electricity price benchmark values, as well as the short-term transaction power purchase amount and average power purchase price benchmark values.

[0075] Through the benchmark values ​​of direct power purchases and average power purchase prices by city (inside and outside the city) and county companies, we can obtain the annual utilization hours of hydropower, wind power, and nuclear power, the unit power purchase prices of thermal power trading units, the average power purchase prices of hydropower, wind power, and nuclear power, the power purchase prices of long-term contracts, and the average power purchase prices of short-term transactions. Then, based on the above electricity prices and market sentiment, we can infer the public's expectations for electricity prices.

[0076] The historical electricity purchase data in S2 is obtained to establish a transaction evaluation model, and the transaction evaluation model is divided into a training set and a test set. The training set and the test set are normalized and preprocessed by dimensionality reduction.

[0077] Among them, when the historical electricity purchase data is divided into a training set and a test set, the ratio of the training set to the test set can be 7:3, and the training set and the test set can be allocated by the fold cross validation method. This method enables the trained transaction evaluation model to have the advantages of both support vector regression machine and principal component analysis, reduce prediction errors, and enhance the generalization ability and stability of the model.

[0078] Working principle: This method can ensure the comprehensiveness and accuracy of problem analysis by following the MECE principle, splitting it into several non-overlapping and fully covered sub-problems, avoiding omissions and duplications when dividing the power purchase structure, thereby establishing a multi-level power purchase transaction structure. By setting up public opinion risk indicators including user policy awareness indicators, it can fully understand the market public opinion involving the public's discussion on electricity prices, avoid excessive excess electricity, which leads to most fluctuations in electricity prices and increases in electricity risk costs, and establish a transaction evaluation model through historical power purchase data. The transaction evaluation model is divided into a training set and a test set, so that the trained transaction evaluation model has the advantages of both support vector regression and principal component analysis, reduces prediction errors, and enhances the generalization ability and stability of the model. The electricity price is obtained according to the power purchase transaction model, thereby ensuring the economy of the power supply system.

[0079] What is described above is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for constructing a multi-level dynamic linkage power purchase transaction evaluation model, characterized in that: The following steps are involved: S1: Divide the electricity purchase structure; S2: Collect historical electricity consumption data and multi-level dynamic real-time electricity purchase data of electricity purchasers, and set constraints on independent variables and dependent variables based on the historical electricity consumption data and multi-level dynamic real-time electricity purchase data; S3: normalize the historical electricity consumption data and the multi-level dynamic real-time electricity purchase data to obtain the normalized electricity purchase vector and data analysis instructions; S4: Based on the data collected by S2 and S3, calculate the type of electricity purchased and summarize the benchmark values ​​of electricity purchase amount, electricity price and electricity fee; S5: Determine the constraint conditions for adjusting the annual utilization hours of thermal power trading units; S6: Input the adjustment amount of the independent variable; S7: Adjust the purchased electricity volume, electricity price and electricity fee according to the benchmark values ​​of purchased electricity volume, electricity price and electricity fee, calculate the evaluation value of each type of purchased electricity, and determine whether the evaluation value meets all constraints. If the constraints are not met, return to S5.

2. The method for constructing a multi-level dynamic linkage electricity purchase transaction evaluation model according to claim 1 is characterized in that: The historical electricity consumption data includes electricity consumption, electricity purchase volume, market supply and demand electricity cost and user participation; the multi-level dynamic real-time electricity purchase data includes electricity purchase volume data, electricity price data and electricity fee data, and the data is collected, entered, analyzed and processed.

3. The method for constructing a multi-level dynamic linkage power purchase transaction evaluation model according to claim 1 is characterized in that: When dividing the electricity purchase structure in S1, the MECE principle should be followed, which has the functions of being independent of each other and not duplicative.

4. The method for constructing a multi-level dynamic linkage electricity purchase transaction evaluation model according to claim 1 is characterized in that: The amount of electricity purchased includes the electricity purchased by the municipal company, the electricity purchased by companies outside the municipality, and the electricity purchased by county-level companies. The electricity purchase by the municipal company can be obtained by dividing the electricity price benchmark by the amount of electricity purchased: Q bas111 =CA 1k *t 1k *(1-β 1k ); Where: Q bas111 is the base value of electricity purchase for each type of self-generated power in service, CA 1K is the installed capacity of each thermal power trading unit, t 1k is the annual utilization hours (k=1, 2, 3...m), β 1k It is the power consumption rate of the factory.

5. The method for constructing a multi-level dynamic linkage power purchase transaction evaluation model according to claim 5 is characterized in that: The base price of electricity purchased by companies outside the city is: Q bas2 =Q bas21 +Q bas22 ; Where: Q bas2 is the benchmark value of electricity purchased from outside the city, Q bas21 is the benchmark value of long-term contractual electricity purchase, Q bas22 It is the accurate value of electricity purchase amount for short-term transactions.

6. The method for constructing a multi-level dynamic linkage power purchase transaction evaluation model according to claim 5 is characterized in that: The base price of electricity purchased by county-level companies is: F bas3 =Q bas3 +P bas3 ; Where: Q bas3 is the benchmark value of electricity purchase by county-level companies, P bas3 is the average electricity price benchmark of county-level companies, Q bas22 It is the accurate value of electricity purchase amount for short-term transactions.

7. The method for constructing a multi-level dynamic linkage power purchase transaction evaluation model according to claim 1 is characterized in that: The historical electricity purchase data in S2 is obtained to establish a transaction evaluation model, the transaction evaluation model is divided into a training set and a test set, and the training set and the test set are respectively preprocessed by normalization and dimensionality reduction.

Images