Deep peak regulation intelligent control method based on AGC circuit

Abstract

This invention relates to the field of power system automatic control technology and discloses a deep peak-shaving intelligent control method based on AGC circuits. This invention collects power generation data and user demand data from the power grid, establishes a multi-objective optimization model, constrains power generation cost, power generation, and power output to obtain the ideal planned power generation sequence of the power grid, calculates the real-time adjusted power demand based on user demand data, and dynamically allocates weights based on the power generation performance and cost of adjustable frequency resources to obtain the dynamic weight of each adjustable frequency resource. This allows for the calculation of the power generation of each resource over a future period, constituting the ideal planned power generation sequence for each resource. This improves the utilization rate of new energy power generation, reduces the overall deep peak-shaving cost of the power grid, and enhances the economic efficiency and safety of operation.

Description

Technical Field

[0001] This invention relates to the field of power system automatic control technology, specifically to a deep peak-shaving intelligent control method based on AGC circuits. Background Technology

[0002] With the advancement of the dual-carbon strategy, the penetration rate of renewable energy, represented by wind power and photovoltaics, in the power system continues to rise. However, the output of renewable energy has significant intermittency, volatility, and anti-peak-shaving characteristics, posing a severe challenge to the traditional peak-shaving system dominated by thermal power.

[0003] Traditional AGC (Automatic Generation Control) systems face numerous challenges and limitations when dealing with complex scenarios involving the overlapping of high-proportion renewable energy connections and deep peak-shaving demands. Traditional systems employ a static allocation framework. For example, during periods of stable wind speed or ample sunlight, wind farms or photovoltaic power plants are fully capable of making limited power adjustments within seconds to minutes, serving as valuable regulation resources. However, the traditional allocation weighting mechanism fails to establish a corresponding dynamic allocation, resulting in renewable energy generation being assigned extremely low regulation priority and wasting valuable regulation resources. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a deep peak-shaving intelligent control method based on AGC circuits. This method has advantages such as power generation cost and performance based on adjustable resources, dynamic allocation of weights for adjustable resources, and thus obtaining an ideal planned power generation sequence, thereby improving the economy and accuracy of the allocation scheme and solving the aforementioned technical problems.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a deep peak-shaving intelligent control method based on AGC circuits, comprising the following steps:

[0006] S1: Collect power generation data and user demand data from the power grid;

[0007] S2: Establish a multi-objective optimization model;

[0008] S3: Input the data collected in S1 into the multi-objective optimization model constructed in S2 to calculate the ideal planned power generation sequence. ;

[0009] S4: Calculate the real-time adjustment demand based on the ideal planned power generation sequence obtained in S3. ;

[0010] S5: Calculate the dynamic allocation weights of each resource. ;

[0011] S6: Adjust the demand based on the real-time adjustment quantity obtained from S4 The dynamically allocated weights obtained from S5 The ideal planned power generation sequence for each resource was calculated. .

[0012] As a preferred embodiment of the present invention, the expression for the S1 power generation data is: ,in, to Representing all times up to the current time. The actual power generation of the power grid;

[0013] The S1 user demand data includes the user's real-time power supply demand.

[0014] As a preferred embodiment of the present invention, the overall objective function of the S2 multi-objective optimization model is: , in, Represent the overall objective function; This indicates taking the minimum value; Indicates the first time in the future rolling time domain Electricity price per unit time period; Indicates the first time in the future rolling time domain Planned power generation for each time period; Indicates the first time in the future rolling time domain The planned power generation of the previous period in the current period; and Indicates the weighting coefficient. ; Indicates the future rolling time domain Summation is performed on each time period; It represents the absolute value.

[0015] As a preferred technical solution of the present invention, the future rolling time domain in the first Planned power generation for each time period The expression is as follows: , in, Indicates time The actual power generation of the power grid; Indicates time The former Actual power generation from the power grid over a given period; Indicates the first The duration of each time period;

[0016] The first in the future rolling time domain Electricity price per time period According to the future The time periods are divided according to the time period type, which includes peak periods, flat periods, and valley periods.

[0017] As a preferred embodiment of the present invention, the constraints of the multi-objective optimization model are as follows: , in, Indicates the first time in the future rolling time domain Planned power generation for each time period; This indicates the user's real-time power demand; Indicates the maximum power generation of the power grid; Indicates the first The duration of each time period; Indicates the first time in the future rolling time domain The planned power generation of the previous period in the current period; Indicates the minimum safe generating capacity of the power grid; This indicates the maximum safe power generation capacity of the power grid.

[0018] As a preferred embodiment of the present invention, the expression for the S3 ideal planned power generation sequence is: ,in, , and Representing the future rolling domain, respectively The time period, the first The time period and the first Ideal planned power generation for a given period of time.

[0019] As a preferred embodiment of the present invention, the expression for the real-time adjustment of the demand in step S4 is as follows: , in, This indicates that demand will be adjusted in real time. Indicates the first term in the future rolling domain Ideal planned power generation for each time period; This indicates the user's real-time power demand.

[0020] As a preferred embodiment of the present invention, the expression for the dynamic weight allocation in S5 is as follows: , in, Indicates the first Dynamic allocation weights for class resources; Indicates the first Performance index of class resources; Indicates the first Resource type in the future rolling time domain The cost of generating electricity in a given time period; This represents the summation operation; Indicates the total number of resource categories; and These are the weighting coefficients. .

[0021] As a preferred technical solution of the present invention, the first Performance index of class resources The expression is as follows: , in, Indicates the first The average real-time response latency of the resource type; This indicates the maximum allowable response latency threshold; Indicates the first The number of times that resource-type tasks have successfully completed deep peak shaving; Indicates the power grid to the first Total number of deep peak-shaving commands issued for resource categories; This indicates taking the maximum value; and These are the weighting coefficients. .

[0022] As a preferred embodiment of the present invention, the calculation formula for obtaining the ideal planned power generation sequence of each resource in step S6 is as follows: , in, Indicates the first The first class of resources in the future scrolling domain Ideal planned power generation for each time period; This indicates that demand will be adjusted in real time. Indicates the first Dynamic allocation weights for class resources;

[0023] The expression for the ideal planned power generation sequence of each resource is: ,in, , and They represent the first Class resources in the future rolling domain The time period, the first The time period and the first Ideal planned power generation for a given period of time.

[0024] Compared with existing technologies, this invention provides a deep peak-shaving intelligent control method based on AGC circuits, which has the following beneficial effects:

[0025] This invention collects power generation data and user demand data from the power grid, establishes a multi-objective optimization model, constrains power generation cost, power generation, and power output to obtain the ideal planned power generation sequence of the power grid, calculates the real-time adjustment demand based on user demand data, and dynamically allocates weights based on the power generation performance and cost of adjustable frequency resources to obtain the dynamic weight of each adjustable frequency resource. In turn, the ideal planned power generation sequence of each resource is calculated, thereby improving the utilization rate of new energy power generation, reducing the overall deep peak shaving cost of the power grid, and improving the economic efficiency and safety of operation. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Please see Figure 1 The deep peak-shaving intelligent control method based on AGC circuit includes the following steps:

[0029] S1: Collect power generation data and user demand data from the power grid;

[0030] The expression for power generation data is: ,in, to Representing all times up to the current time. The actual power generation of the power grid at any time The value is determined by the data collection frequency; user demand data includes the user's real-time power supply demand.

[0031] S2: Establish a multi-objective optimization model with the following overall objective function: , in, Represent the overall objective function; This indicates taking the minimum value; Indicates the first time in the future rolling time domain Electricity price per unit time period; Indicates the first time in the future rolling time domain Planned power generation for each time period; Indicates the first time in the future rolling time domain The planned power generation of the previous period in the current period; and Indicates the weighting coefficient. ; Indicates the future rolling time domain Summation is performed on each time period;

[0032] The generation cost term in the overall objective function and power generation fluctuation items The combined effect of these factors, including the generation fluctuation term, aims to ensure smooth power generation. The overall objective function guides the optimization results to prioritize deep peak shaving during relatively low-price off-peak and flat periods while meeting regulation needs. This is because deep peak shaving during high-price peak periods incurs higher direct generation costs. Through the optimization of this objective function, the system can automatically balance regulation costs and generation fluctuations across different time periods, forming a deep peak shaving plan that balances economic efficiency and security. This effectively reduces the overall cost of deep peak shaving while meeting grid regulation needs and minimizing frequent interventions in unit operation during peak periods.

[0033] The first in the future rolling time domain Planned power generation for each time period The expression is as follows: , in, Indicates time The actual power generation of the power grid; Indicates time The former Actual power generation from the power grid over a given period; Indicates the first The duration of each time period;

[0034] The first in the future rolling time domain Electricity price per time period According to the future The time periods are divided into three types: peak, flat, and valley. Peak periods include 08:30-11:30 and 18:00-23:00. During these periods, due to concentrated industrial and commercial activities, the grid load increases significantly, resulting in relatively high electricity prices. Flat periods, which serve as a transition between peak and valley periods, include 07:00-08:30 and 11:30-18:00. During these periods, the electricity load is relatively stable, and the electricity price falls between that of the peak and valley periods. Valley periods are periods of low electricity demand, from 23:00 to 07:00. During this time, most commercial establishments are closed, residential electricity consumption decreases, the grid load is relatively low, and therefore the electricity price is also low.

[0035] The constraints of the multi-objective optimization model are as follows: , in, Indicates the first time in the future rolling time domain Planned power generation for each time period; This indicates the user's real-time power demand; Indicates the maximum power generation of the power grid; Indicates the first The duration of each time period; Indicates the first time in the future rolling time domain The planned power generation of the previous period in the current period; Indicates the minimum safe generating capacity of the power grid; Indicates the maximum safe generating capacity of the power grid;

[0036] Key constraints of the multi-objective optimization model: To ensure the permissible security of the power grid, firstly, the planned power generation... The power generation must be strictly within the permissible range, meaning it must not be lower than the minimum power generation required to meet user demand, nor exceed the maximum power generation limit of the power grid. Secondly, the variation in power generation between adjacent time periods must be limited, and its value should be within the boundary determined by the preset minimum and maximum safe power change rates. This constraint ensures smooth changes in power generation in the power grid and strictly adheres to the actual ramp-up capability and operational safety requirements of thermal power units under deep peak-shaving conditions. These constraints work together in the optimization process to ensure that the generated power generation plan, while pursuing economic efficiency, can deeply fit the actual physical operation laws and safety criteria of the power grid.

[0037] S3: Input the data collected in S1 into the multi-objective optimization model constructed in S2, and obtain the ideal planned power generation sequence by solving the overall objective function based on the genetic algorithm. ,in, , and Representing the future rolling domain, respectively The time period, the first The time period and the first Ideal planned power generation for each time period;

[0038] S4: Calculate the real-time adjustment demand based on the ideal planned power generation sequence obtained in S3. Its expression is as follows: , in, This indicates that demand will be adjusted in real time. Indicates the first term in the future rolling domain Ideal planned power generation for each time period; This indicates the user's real-time power demand;

[0039] S5: Calculate the dynamic allocation weights of each resource. Its expression is as follows: , in, Indicates the first Dynamic allocation weights for class resources; Indicates the first Performance index of class resources; Indicates the first Resource type in the future rolling time domain Among the power generation costs for different time periods, renewable energy power generation has the lowest cost. This represents the summation operation; This indicates the total number of resource categories, including thermal power generation, wind power generation, gas-fired power generation, hydropower generation, etc. and These are the weighting coefficients. This formula ensures that within a second-level control cycle, the demand for regulation is intelligently and economically allocated to various peak-shaving resources, including thermal power, wind power, gas power, and hydropower. This ensures the stability of the power grid frequency while optimizing the overall economic benefits of deep peak shaving, and ensures that peak-shaving resources with the lowest power generation cost and the best power generation performance can receive a higher allocation weight.

[0040] No. Performance index of class resources The expression is as follows: , in, Indicates the first The average real-time response latency of the resource type; This indicates the maximum allowable response latency threshold; Indicates the first The number of times that resource-type tasks have successfully completed deep peak shaving; Indicates the power grid to the first Total number of deep peak-shaving commands issued for resource categories; This indicates taking the maximum value; and These are the weighting coefficients. This formula is used to accurately quantify and dynamically evaluate the comprehensive regulation capacity of various peak-shaving resources. Used to reflect resource response speed, by comparing its real-time average response latency. With the maximum allowed response latency threshold To calculate, the shorter the delay, the larger the value. Used to reflect the reliability of resource execution, with the success rate of historical deep peak shaving commands. This formula serves as the core basis for dynamic resource assessment, enabling precise identification and priority allocation of high-quality peak-shaving resources that offer rapid response and high reliability.

[0041] S6: Adjust the demand based on the real-time adjustment quantity obtained from S4 The dynamically allocated weights obtained from S5 The ideal planned power generation sequence for each resource was calculated. ;

[0042] The formulas for calculating the ideal planned power generation sequence for each resource are as follows: , in, Indicates the first The first class of resources in the future scrolling domain Ideal planned power generation for each time period; This indicates that demand will be adjusted in real time. Indicates the first Dynamic allocation weights for class resources;

[0043] The expression for the ideal planned power generation sequence of each resource is: ,in, , and They represent the first Class resources in the future rolling domain The time period, the first The time period and the first Ideal planned power generation for a given period of time.

[0044] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A deep peak-shaving intelligent control method based on AGC circuit, characterized in that: Includes the following steps: S1: Collect power generation data and user demand data from the power grid; S2: Establish a multi-objective optimization model; S3: Input the data collected in S1 into the multi-objective optimization model constructed in S2 to calculate the ideal planned power generation sequence. ; S4: Calculate the real-time adjustment demand based on the ideal planned power generation sequence obtained in S3. ; S5: Calculate the dynamic allocation weights of each resource. ; S6: Adjust the demand based on the real-time demand obtained from S4 The dynamically allocated weights obtained from S5 The ideal planned power generation sequence for each resource was calculated. .

2. The deep peak-shaving intelligent control method based on AGC circuit according to claim 1, characterized in that: The expression for the S1 power generation data is: ,in, to Representing all times up to the current time. The actual power generation of the power grid; The S1 user demand data includes the user's real-time power supply demand.

3. The deep peak-shaving intelligent control method based on AGC circuit according to claim 1, characterized in that: The overall objective function of the S2 multi-objective optimization model is: ， in, Represent the overall objective function; This indicates taking the minimum value; Indicates the th time in the future rolling time domain Electricity price per unit time period; Indicates the th time in the future rolling time domain Planned power generation for each time period; Indicates the th time in the future rolling time domain The planned power generation of the previous period in the current period; and Indicates the weighting coefficient. ; Indicates the future rolling time domain Summation is performed on each time period; It represents the absolute value.

4. The deep peak-shaving intelligent control method based on AGC circuit according to claim 3, characterized in that: The first in the future rolling time domain Planned power generation for each time period The expression is as follows: ， in, Indicates time The actual power generation of the power grid; Indicates time The former Actual power generation from the power grid over a given period; Indicates the first The duration of each time period; The first in the future rolling time domain Electricity price per time period According to the future The time periods are divided into time periods of different types, including peak periods, flat periods, and valley periods.

5. The deep peak-shaving intelligent control method based on AGC circuit according to claim 3, characterized in that: The constraints of the multi-objective optimization model are as follows: ， in, Indicates the th time in the future rolling time domain Planned power generation for each time period; This indicates the user's real-time power demand; Indicates the maximum power generation of the power grid; Indicates the first The duration of each time period; Indicates the th time in the future rolling time domain The planned power generation of the previous period in the current period; Indicates the minimum safe generating capacity of the power grid; This indicates the maximum safe power generation capacity of the power grid.

6. The deep peak-shaving intelligent control method based on AGC circuit according to claim 1, characterized in that: The expression for the S3 ideal planned power generation sequence is: ,in, , and Representing the future rolling domain, respectively The time period, the first The time period and the first Ideal planned power generation for a given period of time.

7. The deep peak-shaving intelligent control method based on AGC circuit according to claim 1, characterized in that: The expression for the real-time adjustment of demand by S4 is as follows: ， in, This indicates that demand will be adjusted in real time. Indicates the first term in the future rolling domain Ideal planned power generation for each time period; This indicates the user's real-time power demand.

8. The deep peak-shaving intelligent control method based on AGC circuit according to claim 1, characterized in that: The expression for the S5 dynamic weight allocation is as follows: ， in, Indicates the first Dynamic allocation weights for class resources; Indicates the first Performance index of class resources; Indicates the first Resource type in the future rolling time domain The cost of generating electricity in a given time period; This represents the summation operation; Indicates the total number of resource categories; and These are the weighting coefficients. .

9. The deep peak-shaving intelligent control method based on AGC circuit according to claim 8, characterized in that: No. Performance index of class resources The expression is as follows: ， in, Indicates the first The average real-time response latency of the resource type; This indicates the maximum allowable response latency threshold; Indicates the first The number of times that resource-type operations have successfully completed deep peak shaving; Indicates the power grid to the first Total number of deep peak-shaving commands issued for resource categories; This indicates taking the maximum value; and These are the weighting coefficients. .

10. The deep peak-shaving intelligent control method based on AGC circuit according to claim 1, characterized in that: The calculation formula for the ideal planned power generation sequence of each resource obtained by S6 is as follows: ， in, Indicates the first The first class of resources in the future scrolling domain Ideal planned power generation for each time period; This indicates that demand will be adjusted in real time. Indicates the first Dynamic allocation weights for class resources; The expression for the ideal planned power generation sequence of each resource is: ,in, , and They represent the first Class resources in the future rolling domain The time period, the first The time period and the first Ideal planned power generation for a given period of time.

Images