An automatic generation control method and system

Abstract

The application discloses an automatic power generation control method and system, and relates to the technical field of power control.The automatic power generation control method comprises the following steps: obtaining operation characteristic data of a generator set; identifying key factors that affect the automatic frequency regulation and load distribution performance of the generator set; using a regional load prediction model to predict the frequency and load demand of the generator set at a future time, obtaining a frequency deviation and a load change trend; using AGC closed-loop control logic to monitor the frequency deviation and load change in real time, and dynamically optimizing the output distribution of the generator set.The application can accurately identify the key factors that affect the automatic frequency regulation and load distribution performance through factor analysis, reduce the dependence on secondary factors, make the frequency regulation and load distribution more targeted, thereby ensuring that the load demand is met in time, and reducing the operation risk caused by frequency fluctuation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power control, in particular to an automatic generation control method and system. BACKGROUND

[0002] Automatic Generation Control, AGC for short, is an important part of Energy Management System, EMS for short, which sends instructions to relevant power plants or units according to the control target of the power grid dispatching center, and realizes the automatic control of the power of the generator through the automatic control adjustment device of the power plant or unit. In an interconnected power system, the frequency stability in the system is maintained by the automatic generation control AGC of each regional power grid. Generally speaking, frequency regulation will generate costs, so the regional power grid always wants to minimize its regulation cost, and economic optimization is generally required when automatic generation control is performed. When each region cooperates with each other, economic optimization can be achieved.

[0003] In the prior art, it is inconvenient to accurately identify the key factors affecting the automatic frequency regulation and load distribution performance, so that the frequency regulation and load distribution are not targeted, and it is inconvenient to accurately predict the frequency deviation and load change trend at the future time, so as to inconveniently realize the early warning of the possible load fluctuation and frequency change in the future, and at the same time, it is inconvenient to use the AGC closed-loop control logic to realize real-time optimization of the output distribution, so that the generator unit output is more close to the actual demand, thereby reducing the operation efficiency of the generator unit and increasing the energy consumption and operation cost.

[0004] At present, no effective solution has been proposed for the problems in the related art. SUMMARY

[0005] In view of the problems in the related art, the present application provides an automatic generation control method and system to overcome the above technical problems existing in the prior art.

[0006] To this end, the specific technical solutions adopted by the present application are as follows:

[0007] According to one aspect of the present application, an automatic generation control method is provided, which comprises the following steps:

[0008] S1, obtaining real-time operation data of a generator unit, and preprocessing the real-time operation data to obtain operation characteristic data of the generator unit;

[0009] S2, using a factor analysis algorithm to analyze the obtained operation characteristic data, and identifying the key factors affecting the automatic frequency regulation and load distribution performance of the generator unit;

[0010] S3, based on the identified key factors, a regional load forecasting model is constructed, and the frequency and load demand of the generating unit at future time are predicted by using the regional load forecasting model, to obtain the frequency deviation and load change trend;

[0011] S4, based on the frequency deviation and load change trend, combined with the real-time operation state of the generating unit, the operation parameters of the generating unit are dynamically adjusted through the AGC control strategy, and the frequency deviation and load change are monitored in real time by using the AGC closed-loop control logic, and the output distribution of the generating unit is dynamically optimized.

[0012] Further, the obtained operation characteristic data is analyzed by using a factor analysis algorithm, and the key factors affecting the automatic frequency regulation and load distribution performance of the generating unit are identified, including the following steps:

[0013] S21, obtaining the historical operation data of the generating unit, counting the occurrence times of various key factors affecting the automatic frequency regulation and load distribution performance, and calculating the distribution proportion of the key factors;

[0014] S22, based on a rule analysis algorithm, the distribution of the influencing factors in the operation characteristic data of the generating unit is analyzed in depth, and the distribution rule of the influencing factor proportion is generated;

[0015] S23, based on the generated distribution rule of the influencing factor proportion, an influencing factor identification model of the automatic frequency regulation and load distribution performance of the generating unit is constructed;

[0016] S24, if the effect degree of the detected influencing factor in the operation state exceeds the set threshold, the operation state of the current generating unit is analyzed by using the influencing factor identification model, and the key factors having significant influence on the automatic frequency regulation and load distribution performance are identified.

[0017] Further, based on the rule analysis algorithm, the distribution of the influencing factors in the operation characteristic data of the generating unit is analyzed in depth, and the distribution rule of the influencing factor proportion is generated, including the following steps:

[0018] S221, setting the initial parameters and the maximum number of iterations of the rule analysis algorithm;

[0019] S222, in the feasible feature space of the operation characteristic data of the generating unit, based on the historical data and statistical characteristics, an initial influencing factor combination is randomly generated, and the distribution rule and fitness value of each combination are calculated by using the rule analysis algorithm;

[0020] S223, according to the fitness value, the influencing factor combination with the best operation performance is included in the best group, and the worst one is included in the optimization group;

[0021] S224, cross combination between the best group and the optimization group is performed by using the feature cross analysis method to generate a new candidate influence factor combination;

[0022] S225, the distribution rule and fitness value of the new candidate influence factor combination are calculated, and if the fitness value of the new influence factor combination on the influence factor distribution rule is greater than that of the original influence factor combination, the new influence factor combination is used to replace the original influence factor combination;

[0023] S226, it is checked whether the maximum iteration number is reached, if yes, the final distribution rule of the influence factor proportion is output, otherwise, the step S223 is returned to continue the optimization analysis.

[0024] Further, the cross combination between the best group and the optimization group is performed by using the feature cross analysis method to generate a new candidate influence factor combination, including the following steps:

[0025] S2241, the influence factor combination with the highest fitness value is selected from the best group as the parent candidate, and the influence factor combination is randomly selected from the optimization group to form a pair of basic influence factor combinations for cross analysis;

[0026] S2242, according to the cross formula, the transverse cross operation is performed on each pair of influence factor combinations in the best group and the optimization group to generate a new candidate influence factor combination.

[0027] Further, based on the identified key factors, a regional load prediction model is constructed, and the frequency and load demand of the generator unit at the future time are predicted by using the regional load prediction model to obtain the frequency deviation and load change trend, including the following steps:

[0028] S31, taking the minimization of the regional load prediction error as the objective function, combining the operating characteristics of the generator unit, the constraint conditions of the frequency and load prediction are established;

[0029] S32, the deviation between the frequency deviation and the load change distribution is used to measure the distance between the actual distribution of the frequency deviation and the load change and the reference distribution, and a load prediction uncertainty set is constructed;

[0030] S33, based on the distributed robust optimization algorithm, an opportunity constraint optimization model for the uncertainty set is constructed, and a mathematical solver is used to solve the opportunity constraint model to output the frequency change and load demand prediction plan value of the generator unit at the future time;

[0031] S34, combined with the frequency change and load demand prediction plan value, based on the current state, output increment and disturbance of the generator unit, a regional load prediction model is constructed to obtain the frequency deviation and load change trend.

[0032] Further, the deviation amount between the frequency deviation and the load change distribution is utilized, the relative entropy is used to measure the distance between the actual distribution and the reference distribution, and the load prediction uncertainty set is constructed including the following steps:

[0033] S321, set the relative entropy limit value of the frequency deviation and the load change distribution, utilize the deviation amount between the relative entropy of the actual distribution and the reference distribution, and define the distance between the actual distribution and the reference distribution;

[0034] S322, based on the deviation amount between the actual distribution and the reference distribution, construct the uncertainty set of the frequency deviation and the load change distribution.

[0035] According to another aspect of the present application, an automatic power generation control system is also provided, which comprises:

[0036] A data acquisition module is configured to acquire real-time operation data of a generator set, pre-process the real-time operation data, and obtain operation characteristic data of the generator set;

[0037] A factor analysis module is configured to analyze the obtained operation characteristic data by using a factor analysis algorithm, and identify key factors affecting the automatic frequency regulation and load distribution performance of the generator set;

[0038] A model construction module is configured to construct a regional load prediction model based on the identified key factors, and predict the frequency deviation and load change trend of the generator set at a future time by using the regional load prediction model;

[0039] A power generation control module is configured to dynamically adjust the operation parameters of the generator set by using an AGC control strategy based on the frequency deviation and load change trend and in combination with the real-time operation state of the generator set, and monitor the frequency deviation and load change in real time by using an AGC closed-loop control logic, and dynamically optimize the output distribution of the generator set.

[0040] The present application has the following advantages:

[0041] 1. The present application can accurately identify the key factors affecting the automatic frequency regulation and load distribution performance by factor analysis, reduce the dependence on secondary factors, make the frequency regulation and load distribution more targeted, construct a regional load prediction model based on the key factors, accurately predict the frequency deviation and load change trend at a future time, realize the early warning of possible load fluctuation and frequency change in the future, optimize the output distribution in real time by using the AGC closed-loop control logic, make the output of the generator set more close to the actual demand, improve the operation efficiency of the generator set, reduce the energy consumption and operation cost, ensure that the load demand is met in time, and reduce the operation risk caused by frequency fluctuation.

[0042] 2、The application is based on factor analysis algorithm, a scientific key factor identification model is constructed, automatic frequency modulation and load distribution are more accurate, through real-time identification and dynamic adjustment, the utilization efficiency of power generation resources is significantly improved, the operation cost is reduced, abnormal state is quickly responded, the stability and adaptability of the generator set under complex working conditions are enhanced, thereby the frequency regulation accuracy is improved, the frequency deviation is reduced, and the power grid operation stability is enhanced.

[0043] 3、The application reduces frequency deviation and load fluctuation through error minimization and constraint conditions, the introduction of relative entropy and uncertainty set can cope with complex working conditions and prediction error, improve robustness, dynamically adjust generator set output plan, reduce loss in frequency modulation and load distribution, reduce operation cost, dynamically update prediction model and frequency deviation trend, improve adaptability to real-time operation state, dynamically update prediction model and frequency deviation trend, improve adaptability to real-time operation state, thereby avoid frequency instability caused by prediction error and external disturbance, and ensure safe operation of power grid. BRIEF DESCRIPTION OF DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0045] Figure 1 is a flow chart of an automatic power generation control method according to an embodiment of the present application;

[0046] Figure 2 is a principle block diagram of an automatic power generation control system according to an embodiment of the present application.

[0047] In the drawings:

[0048] 1, data acquisition module; 2, factor analysis module; 3, model construction module; 4, power generation control module. DETAILED DESCRIPTION

[0049] To further illustrate the embodiments, the present application provides drawings, which are part of the disclosure of the present application, mainly used to illustrate the embodiments, and can explain the operating principle of the embodiments in conjunction with the related description of the specification. With reference to these contents, those skilled in the art should understand other possible embodiments and advantages of the present application.

[0050] According to the embodiments of the present application, an automatic power generation control method and system are provided.

[0051] The application will be further described in conjunction with the drawings and specific embodiments, as shown in the drawings Figure 1 The automatic power generation control method according to the embodiment of the application comprises the following steps:

[0052] S1, obtaining real-time operation data of the generator set, and preprocessing the real-time operation data to obtain operation characteristic data of the generator set;

[0053] Specifically, the real-time operation data includes electrical characteristic related data, mechanical characteristic related data, thermal characteristic related data, state monitoring data and external condition related data, etc.

[0054] Specifically, the operation characteristic data includes state characteristics, performance characteristics, health state characteristics, dynamic characteristics and environmental influence characteristics, etc.

[0055] S2, using a factor analysis algorithm to analyze the obtained operation characteristic data, and identifying key factors affecting the automatic frequency regulation and load distribution performance of the generator set;

[0056] Specifically, the key factors include electrical characteristic factors, mechanical characteristic factors, thermal characteristic factors, operation state factors, load characteristic factors and environmental factors, etc.

[0057] S3, based on the identified key factors, constructing a regional load prediction model, and using the regional load prediction model to predict the frequency and load demand of the generator set at a future time, to obtain a frequency deviation and a load change trend;

[0058] S4, based on the frequency deviation and the load change trend, combining the real-time operation state of the generator set, dynamically adjusting the operation parameters of the generator set through the AGC control strategy, and using the AGC closed-loop control logic to monitor the frequency deviation and the load change in real time, and dynamically optimizing the output distribution of the generator set.

[0059] It needs to be explained that the deviation of the current power grid frequency from the target frequency (such as 50Hz or 60Hz) is monitored; the load change trend is monitored in real time through the load prediction model or actual demand; the active power, reactive power, speed, fuel consumption rate and unit health status of the generator set are collected, and the deviation of the real-time frequency of the power grid from the target frequency is calculated; the increase and decrease rate of the load demand is predicted through trend analysis to determine the future load demand change; the response capability of the generator set to frequency adjustment and load change is evaluated according to the current operating state of the generator set; the frequency deviation value and the load change trend data; according to the dynamic response capability, operating state and health status of the generator set, the adjustment priority of different units is allocated; the adjustment target is decomposed into multiple sub-tasks, including adjusting the active power according to the frequency deviation; load balancing task; adjusting the unit output according to the load demand; using the optimal load distribution algorithm (such as economic load distribution algorithm), while meeting the adjustment demand, reducing fuel consumption and power generation cost, outputting the adjustment target (such as power increase and decrease, output distribution) of each generator set; adjust the active power output of the generator set to restore the frequency deviation; adjust the reactive power output to ensure voltage stability; adjust the main shaft speed of the generator set according to the load demand; adjust the fuel input to meet the power generation demand and optimize fuel consumption; the adjusted operating parameters, the generator set enters a new state; continuously collect the adjusted operating state, including frequency change, load change and power output; if the adjusted frequency deviation and load demand do not reach the target, trigger further adjustment; according to the logic of the AGC controller, automatically calculate the new adjustment instruction; the closed-loop control logic iteratively optimizes the operating parameters of the generator set based on real-time data until the target frequency and load distribution are reached; according to the dynamic adjustment capability of the generator set and the load demand of the power grid, the output distribution of multiple units is coordinated; while meeting the frequency regulation and load distribution, the economic load distribution (ELD) algorithm is used to reduce the total power generation cost; reserve part of the redundancy of the generator set to cope with sudden load increase or emergencies; monitor the operating state of the generator set (such as overload, abnormal vibration, temperature rise exceeding limit); enable the standby unit or the fast-start unit to make up for the output loss caused by abnormalities; if the AGC system cannot complete automatic adjustment, an alarm signal is sent and manual intervention is switched to; save real-time operating data, adjustment process data and output distribution results; evaluate the frequency modulation performance, load distribution accuracy and economy; optimize the AGC algorithm and load prediction model according to the evaluation results.

[0060] In this optional embodiment, real-time operating data of the generator set is acquired, and the real-time operating data is preprocessed to obtain operating feature data of the generator set, including the following steps:

[0061] S11, constructing an operating feature curve according to the real-time data, calculating the change amplitude difference value of the adjacent time feature curve, and extracting the feature index of the operating state change;

[0062] S12, a preset amplitude threshold is set, if the amplitude difference value exceeds the threshold, a key runtime is added; otherwise, the existing key time set is retained;

[0063] S13, checking the distribution of the key runtime, if the distribution is too dense, merging adjacent time points, otherwise, retaining the current key time set as the operation characteristic data of the generator set.

[0064] In this optional embodiment, the obtained operation characteristic data is analyzed by using the factor analysis algorithm to identify the key factors affecting the automatic frequency regulation and load distribution performance of the generator set, including the following steps:

[0065] S21, obtaining the historical operation data of the generator set, counting the occurrence times of various key factors affecting the automatic frequency regulation and load distribution performance, and calculating the distribution proportion of the key factors;

[0066] S22, based on the rule analysis algorithm, the distribution of the influencing factors in the operation characteristic data of the generator set is analyzed in depth to generate the distribution rule of the influencing factor proportion;

[0067] S23, based on the generated distribution rule of the influencing factor proportion, an influencing factor identification model of the automatic frequency regulation and load distribution performance of the generator set is constructed;

[0068] S24, if the effect degree of the detected influencing factor in the operation state exceeds the set threshold, the operation state of the current generator set is analyzed by using the influencing factor identification model to identify the key factors that have a significant impact on the automatic frequency regulation and load distribution performance.

[0069] Specifically, the factor analysis algorithm is an improved NSGA-II algorithm, which is a classical multi-objective optimization algorithm, and is used to process multi-objective optimization problems, and can effectively find a solution set that meets the multi-objective requirements. The basic NSGA-II algorithm is improved, and an effective integer coding method is designed, and a deduplication operation is introduced.

[0070] In this optional embodiment, based on the rule analysis algorithm, the distribution of the influencing factors in the operation characteristic data of the generator set is analyzed in depth to generate the distribution rule of the influencing factor proportion, including the following steps:

[0071] S221, setting the initial parameters and the maximum number of iterations of the rule analysis algorithm;

[0072] S222, in the feasible feature space of the operation characteristic data of the generator set, based on the historical data and statistical characteristics, an initial influencing factor combination is randomly generated, and the distribution rule and fitness value of each combination are calculated by the rule analysis algorithm;

[0073] S223, according to the fitness value, the performance of the best combination of influencing factors is included in the best group, and the worst is included in the optimization group;

[0074] S224, using the feature cross analysis method, combination cross is performed between the best group and the optimization group to generate a new candidate influencing factor combination;

[0075] S225, for the new candidate influencing factor combination, calculate its distribution rule and fitness value; if the fitness of the new influencing factor combination on the influencing factor distribution rule is greater than that of the original influencing factor combination, the new influencing factor combination is used to replace the original influencing factor combination;

[0076] S226, check whether the maximum iteration number is reached, if reached, output the final distribution rule of the influencing factor proportion, otherwise, return to step S223 and continue optimization analysis.

[0077] Specifically, the rule analysis algorithm is a sorting cross optimization algorithm, which is an optimization algorithm based on cross and sorting mechanism, and is usually used to solve complex combination optimization problems, such as scheduling problem, path planning problem, etc. It operates on the arrangement order of the solution, combines cross, mutation and selection mechanism, and constantly searches for global optimal solution in the search space.

[0078] In this optional embodiment, using the feature cross analysis method, combination cross is performed between the best group and the optimization group to generate a new candidate influencing factor combination, including the following steps:

[0079] S2241, select the influencing factor combination with the highest fitness value from the best group as the parent candidate, and randomly select the influencing factor combination from the optimization group to form a pair of basic influencing factor combinations for cross analysis;

[0080] S2242, according to the cross formula, perform horizontal cross operation on each pair of influencing factor combinations in the best group and the optimization group to generate a new candidate influencing factor combination.

[0081] In this optional embodiment, the cross formula is:

[0082] ;

[0083] In the formula, Z new represents the newly generated candidate influencing factor combination; Z best represents the influencing factor combination selected from the best group; Z opt represents the influencing factor combination selected from the optimization group; delta represents the weight coefficient; epsilon represents random disturbance.

[0084] In this optional embodiment, based on the identified key factors, a regional load prediction model is constructed, and the frequency and load demand of the generating unit at the future time are predicted using the regional load prediction model, to obtain the frequency deviation and load change trend, including the following steps:

[0085] S31, taking the minimization of the regional load prediction error as the objective function, combining the operating characteristics of the generating unit, and establishing the constraint condition of the frequency and load prediction;

[0086] S32, using the deviation amount between the frequency deviation and the load change distribution, using relative entropy to measure the distance between the actual distribution and the reference distribution of the frequency deviation and the load change, and constructing a load prediction uncertainty set;

[0087] S33, based on the distributed robust optimization algorithm, an opportunity constraint optimization model for the uncertainty set is constructed, and a mathematical solver is used to solve the opportunity constraint model, to output the frequency change and load demand prediction plan value of the generating unit at the future time;

[0088] Specifically, the distributed robust optimization algorithm is an optimization method for solving problems under uncertainty, especially in the presence of an uncertainty set, which provides a robust solution to the constraint condition. The core goal is to ensure that the constraint condition and the optimization target can meet a certain confidence level under the influence of uncertainty.

[0089] Specifically, the mathematical solver includes gurobi (high-efficiency solver) or other mathematical solvers.

[0090] S34, combining the frequency change and the load demand prediction plan value, based on the current state, output increment and disturbance of the generating unit, a regional load prediction model is constructed, to obtain the frequency deviation and load change trend.

[0091] In this optional embodiment, using the deviation amount between the frequency deviation and the load change distribution, using relative entropy to measure the distance between the actual distribution and the reference distribution, and constructing a load prediction uncertainty set includes the following steps:

[0092] S321, set the relative entropy limit value of the frequency deviation and the load change distribution, use the relative entropy to measure the deviation amount between the actual distribution and the reference distribution, and define the distance between the actual distribution and the reference distribution;

[0093] Specifically, relative entropy, also known as Kullback-Leibler (KL) divergence, is an important concept in information theory, which is used to measure the difference between two probability distributions. It represents the amount of information lost when one distribution (usually the actual distribution) is given another distribution (usually the reference distribution).

[0094] S322. Based on the deviation between the actual distribution and the reference distribution, construct an uncertain set of frequency deviation and load change distribution.

[0095] In this optional embodiment, the formula for defining the distance between the actual distribution and the reference distribution is:

[0096] ;

[0097] In the formula, This indicates that the actual distribution is measured using relative entropy. M Compared with the reference distribution N The distance between them; M ( x This represents the actual distribution of frequency deviation and load variation; N ( x This represents the reference distribution of frequency deviation and load variation; x This represents a combined variable of frequency deviation and load variation; d x It represents a small change in the combined variables; KL Represents relative entropy.

[0098] According to another embodiment of the invention, such as Figure 2 As shown, an automatic power generation control system is also provided, which includes:

[0099] Data acquisition module 1 is used to acquire real-time operating data of the generator set and preprocess the real-time operating data to obtain operating characteristic data of the generator set.

[0100] Factor analysis module 2 is used to analyze the obtained operating characteristic data using factor analysis algorithms to identify key factors affecting the automatic frequency regulation and load distribution performance of generator sets;

[0101] Model building module 3 is used to build a regional load forecasting model based on the identified key factors, and to use the regional load forecasting model to predict the frequency and load demand of generator units in the future, and obtain the frequency deviation and load change trend.

[0102] The power generation control module 4 is used to dynamically adjust the operating parameters of the generator set based on the frequency deviation and load change trend, combined with the real-time operating status of the generator set, through the AGC control strategy, and to dynamically optimize the power output distribution of the generator set by using the AGC closed-loop control logic to monitor the frequency deviation and load change in real time.

[0103] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An automatic generation control method characterized by, The automatic power generation control method comprises the following steps: S1, obtaining real-time operation data of the generator set, and preprocessing the real-time operation data to obtain operation characteristic data of the generator set; S2, using a factor analysis algorithm to analyze the obtained operation characteristic data to identify key factors affecting the automatic frequency regulation and load distribution performance of the generator set; including: S21, obtaining historical operation data of the generator set, counting the number of occurrences of various key factors affecting the automatic frequency regulation and load distribution performance, and calculating the distribution proportion of the key factors; S22, based on a rule analysis algorithm, in-depth analyzing the distribution of the influencing factors in the operation characteristic data of the generator set to generate a distribution rule of the influencing factor proportion; including: S221, setting initial parameters and a maximum number of iterations of the rule analysis algorithm; S222, in the feasible characteristic space of the operation characteristic data of the generator set, based on historical data and statistical characteristics, randomly generating an initial influencing factor combination, and calculating the distribution rule and fitness value of each combination through the rule analysis algorithm; S223, according to the fitness value, the influencing factor combination with the best performance is included in the best group, and the worst one is included in the optimization group; S224, using a feature crossover analysis method, performing combination crossover between the best group and the optimization group to generate a new candidate influencing factor combination; S225, calculating the distribution rule and fitness value of the new candidate influencing factor combination; if the fitness of the new influencing factor combination on the influencing factor distribution rule is greater than that of the original influencing factor combination, the new influencing factor combination is used to replace the original one; S226, checking whether the maximum number of iterations is reached, if yes, outputting the final distribution rule of the influencing factor proportion, otherwise, returning to step S223 for further optimization and analysis; S23, based on the generated distribution rule of the influencing factor proportion, constructing an influencing factor identification model of the automatic frequency regulation and load distribution performance of the generator set; S24, if the detected influencing factor in the operation state exceeds the set threshold, using the influencing factor identification model to analyze the operation state of the current generator set to identify key factors that have a significant impact on the automatic frequency regulation and load distribution performance; S3, based on the identified key factors, constructing a regional load prediction model, and using the regional load prediction model to predict the frequency and load demand of the generator set at a future time to obtain the frequency deviation and load change trend; including: S31, taking the minimization of the regional load prediction error as the objective function, combining the operating characteristics of the generator set, and establishing the constraint conditions for frequency and load prediction; S32, using the deviation between the frequency deviation and the load change distribution, using relative entropy to measure the distance between the actual distribution and the reference distribution of the frequency deviation and the load change, and constructing an uncertain set of load prediction; including: S321, setting the relative entropy limit of the frequency deviation and the load change distribution, using relative entropy to measure the deviation between the actual distribution and the reference distribution, and defining the distance between the actual distribution and the reference distribution; S322, based on the deviation between the actual distribution and the reference distribution, constructing an uncertain set of the frequency deviation and the load change distribution; S33, based on the distributed robust optimization algorithm, an opportunity constraint optimization model for the uncertain set is constructed, and a mathematical solver is used to solve the opportunity constraint model, and the frequency change and load demand prediction plan value of the generator unit at the future time are output; S34, combined with the frequency change and load demand prediction plan value, based on the current state, output increment and disturbance of the generator unit, a regional load prediction model is constructed, and the frequency deviation and load change trend are obtained; S4, based on the frequency deviation and load change trend, combined with the real-time running state of the generator unit, the operating parameters of the generator unit are dynamically adjusted through the AGC control strategy, and the AGC closed-loop control logic is used to monitor the frequency deviation and load change in real time, and the output distribution of the generator unit is dynamically optimized.

2. An automatic generation control method according to claim 1, characterized by, The real-time running data of the generator unit is obtained, and the real-time running data is preprocessed to obtain the operating characteristic data of the generator unit, including the following steps: S11, the operating characteristic curve is constructed according to the real-time data, the change amplitude difference value of the adjacent time characteristic curve is calculated, and the characteristic index of the operating state change is extracted; S12, a preset amplitude threshold is set, if the change amplitude difference value exceeds the threshold, a new key operating time is added, otherwise, the existing key time set is retained; S13, check the key operating time distribution, if the distribution is too dense, merge the adjacent time, otherwise, retain the current key time set as the operating characteristic data of the generator unit.

3. The automatic generation control method according to claim 1, wherein The characteristic cross analysis method is used to combine and cross between the best group and the optimization group to generate new candidate influence factor combinations, including the following steps: S2241, select the influence factor combination with the highest fitness value from the best group as the parent candidate, and randomly select the influence factor combination from the optimization group to form a pair of basic influence factor combinations for cross analysis; S2242, according to the cross formula, each pair of influence factor combinations in the best group and the optimization group is subjected to horizontal cross operation to generate new candidate influence factor combinations.

4. An automatic generation control method according to claim 3, characterized by, The cross formula is: ； wherein, Z new denotes a newly generated candidate combination of influence factors; Z best denotes a combination of influence factors selected from the best set; Z opt denotes a combination of influence factors selected from the optimized set; The formula for defining the distance between the actual distribution and the reference distribution is: denotes a weight coefficient; The automatic generation control system comprises: denotes a random perturbation.

5. The automatic generation control method according to claim 1, wherein A data acquisition module is configured to acquire real-time running data of a generator unit, and preprocess the real-time running data to obtain operating characteristic data of the generator unit; ； wherein represents the distance between the actual distribution M and the reference distribution N using relative entropy; M x represents the actual distribution of frequency deviation and load variation; N x represents the reference distribution of frequency deviation and load variation; x represents the combined variable of frequency deviation and load variation; d x represents the small change amount of the combined variable; A factor analysis module is configured to analyze the obtained operating characteristic data using a factor analysis algorithm to identify key factors affecting the automatic frequency regulation and load distribution performance of the generator unit; represents the relative entropy.​​ 6. An automatic generation control system for implementing the automatic generation control method according to any one of claims 1 to 5, characterized by A model construction module is configured to construct a regional load prediction model based on the identified key factors, and predict the frequency and load demand of the generator unit at a future time using the regional load prediction model to obtain a frequency deviation and a load change trend; A generation control module is configured to dynamically adjust operating parameters of the generator unit based on the frequency deviation and the load change trend, in combination with a real-time running state of the generator unit, through an AGC control strategy, and to monitor the frequency deviation and the load change in real time using an AGC closed-loop control logic, and to dynamically optimize output distribution of the generator unit. ​ ​ ​

Images