Deep peak regulation control system for thermal power generating unit

Abstract

The invention discloses a thermal power generating unit deep peak regulation control system, and relates to the technical field of thermal power generating unit deep peak regulation, the thermal power generating unit deep peak regulation control system comprises a deep peak regulation control center, and the deep peak regulation control center is in communication connection with a data acquisition module, a load prediction module, a working condition monitoring module, a distributed control module and a safety monitoring early warning module. By introducing a distributed control architecture and bat algorithm optimization, each subsystem of the thermal power generating unit is divided into a plurality of local control units, independent and cooperative control is realized, the flexibility and the response speed of the thermal power generating unit in the deep peak regulation process are remarkably improved, and the reliability of the thermal power generating unit is improved. Distributed control ensures that each subsystem can quickly adjust operation parameters according to actual load requirements, and the global optimization capability of the bat algorithm further improves the self-adaptability and decision-making efficiency of the system in a complex working condition, so that the problem of response lag in a traditional centralized control mode is effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of deep peak regulation of thermal power generation units, and in particular to a deep peak regulation control system for thermal power generation units. Background Art

[0002] In recent years, new energy power generation such as wind power and photovoltaic power has developed rapidly, and the proportion of new energy installed capacity has increased significantly, which has changed the structure of the power market. In the traditional power system, thermal power units are mainly responsible for providing electricity and power. With the rise of new energy power generation, the role of thermal power units has gradually changed to providing reliable electricity and basic power supply with peak and frequency regulation capabilities. Thermal power units need to respond more flexibly to changes in grid load to adapt to the instability of new energy power generation.

[0003] For example, the deep peak-shaving thermal power unit control method and control system based on big data drive disclosed in Chinese patent publication number: CN118353092A uses climate data, electricity consumption data, power generation data and historical operation big data of the unit to determine the optimal control amount of the unit under deep peak-shaving conditions, thereby improving the efficiency and accuracy of the control.

[0004] In the prior art, the optimal control quantity of the unit under deep peak-shaving conditions is determined through climate data, electricity consumption data, power generation data and historical operation big data of the unit, so as to obtain the corresponding optimal control quantity under the corresponding peak-shaving instruction, thereby solving the problem that the method for determining the optimal control quantity is relatively single and has low adaptability to complex deep peak-shaving conditions. However, since the deep peak-shaving of thermal power units involves the coordinated control among multiple subsystems, it is difficult to achieve the optimal control effect when facing different deep peak-shaving conditions, resulting in slow load response of the unit and insufficient peak-shaving depth. Therefore, how to adopt a combination of distributed control and bat algorithm to adapt to complex conditions and speed up the response speed is the problem to be solved by the present invention. To this end, a deep peak-shaving control system for thermal power units is proposed. Summary of the Invention

[0005] The object of the present invention is to provide a deep peak regulation control system for a thermal power unit to solve the problems raised in the above background technology.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A deep peak shaving control system for a thermal power unit includes a deep peak shaving control center, wherein the deep peak shaving control center is communicatively connected to a data acquisition module, a load forecasting module, an operating condition monitoring module, a distributed control module, and a safety monitoring and early warning module, wherein electrical signals are connected between the modules;

[0008] The deep peak shaving control center, as the command center of the entire control system, is responsible for communication and coordination between modules, integrating information from each module and issuing instructions to achieve centralized management and unified control of the deep peak shaving process of thermal power units, ensuring the coordinated operation of each module;

[0009] The data acquisition module is used to collect and pre-process the operating data of each subsystem of the thermal power unit in real time, including key parameters such as temperature, pressure, flow, power, speed and load. The distributed arrangement of data collection points ensures the comprehensiveness and real-time nature of the data.

[0010] The load forecasting module is used to analyze the load change trend of the thermal power unit using historical operation data;

[0011] The operating condition monitoring module is used to monitor the operating condition of the thermal power unit and identify abnormalities in the current operating condition;

[0012] The distributed control module is used to divide the various subsystems of the thermal power unit into multiple local control units, and use the bat algorithm to analyze the load changes and operating condition monitoring results of the thermal power unit to find the optimal control parameters and scheduling plan, and then generate and execute control instructions;

[0013] The safety monitoring and early warning module is used to evaluate the operating effect of the thermal power unit and continuously monitor various key parameters during the deep peak regulation process of the thermal power unit. Once it is found that the key parameters exceed the safety threshold, the alarm mechanism is triggered and corresponding safety protection measures are taken.

[0014] A further improvement of the technical solution of the present invention is that the data acquisition module specifically includes:

[0015] Based on the peak-shaving requirements of thermal power units, plan the layout of data collection points for each subsystem of the thermal power units and determine the operating data to be collected, including key parameters such as temperature, pressure, flow, power, speed, and load. At the same time, establish the collection frequency and time interval, and design an operating data collection plan to ensure comprehensive coverage of the operating status of each subsystem.

[0016] According to the planned data collection plan, distributed data collection points are used to collect operating data from various subsystems of the thermal power unit. Each data collection point obtains the actual value of the corresponding key parameters through the configured sensors and transmits the data to the data acquisition module in the form of electrical signals to ensure the real-time and accuracy of the data;

[0017] Preprocess the collected raw operating data, including data cleaning and normalization steps;

[0018] The pre-processed operation data is integrated to obtain the thermal power operation sequence table, and a data warehouse is built to store the relevant data of the thermal power operation sequence table in the data warehouse.

[0019] A further improvement of the technical solution of the present invention is that the load forecasting module specifically includes:

[0020] Extract historical operating data of thermal power units from their operating records, including power, load, and corresponding timestamp information, and sort and align the data in time series to form a unified format.

[0021] Perform feature analysis on historical operating data to extract trend features related to load forecasting, including load change rate, load fluctuation amplitude, load change duration, and load change frequency. Combined with historical operating data and thermal power unit operating requirements, baseline values ​​for each trend feature are preset. Then, each trend feature and its corresponding baseline value are integrated to obtain a load trend feature sequence table.

[0022] The load trend feature sequence table is divided into a training set, a validation set, and a test set. A time series analysis-based infrastructure is selected to build a load forecasting model that inputs trend features and outputs the load change trend of thermal power units.

[0023] By using the constructed load forecasting model, the relevant data of the load trend characteristic sequence table of the current time period is input, the load trend value of the computer group is calculated, and the load change trend of the thermal power unit in the future period of the same time period is predicted.

[0024] A further improvement of the technical solution of the present invention is that the calculation process of the unit load trend value is:

[0025] Extracting trend feature-related data of the thermal power unit in the current time period from the load trend feature sequence table, including the load change rate, load fluctuation amplitude, load change duration, and load change frequency at each time point, and determining the reference values ​​corresponding to each trend feature, namely, the load change rate reference value, the load fluctuation amplitude reference value, the load change duration reference value, and the load change frequency reference value;

[0026] For each time point, the ratio of each trend feature to its baseline value is calculated to reflect the difference between the actual trend feature at that time point and the baseline value, and the relative change rate of each trend feature is obtained;

[0027] For each time point, calculate the square root of the sum of the squares of the load change rate and the relative change rate of the load fluctuation amplitude as the numerator;

[0028] At each time point, the sum of the relative change rate of the load change duration and the relative change rate of the load change frequency is substituted into the exponential function to obtain the value of the denominator. The use of the exponential function emphasizes the impact of the load change duration and the load change frequency, and the denominator as a whole reflects the comprehensive impact of the duration and frequency on the unit load change;

[0029] Divide the calculated numerator by the denominator to obtain the exponential contribution value at each time point, and then add up the exponential contribution values ​​of all time points. The sum result is the unit load trend value.

[0030] A further improvement of the technical solution of the present invention is that the working condition monitoring module specifically includes:

[0031] Extract operating data collected from various subsystems of thermal power units, including key parameters such as temperature, pressure, flow, power, speed, and load. Align data from different subsystems using timestamps. Eliminate dimensional differences using Z-score normalization or Min-Max normalization, and unify heterogeneous data into a time series dataset to ensure data consistency and comparability.

[0032] Extract the operating characteristics reflecting the operating conditions of the thermal power unit from the time series data set, and then determine the sub-features of each operating characteristic. Use the clustering algorithm to classify the historical operating data, analyze the historical operating data and the standard feature patterns under normal conditions, establish a typical operating mode library, and determine the benchmark values ​​corresponding to the sub-features of each operating characteristic. Among them, for the operating characteristics of temperature, its sub-features include main steam temperature, reheat steam temperature, condenser temperature and bearing temperature; for the operating characteristics of pressure, its sub-features include main steam pressure, condenser pressure and fuel pressure; for the operating characteristics of flow, its sub-features include fuel flow, feed water flow and air flow; for the operating characteristics of power, its sub-features include turbine output power, generator output power and reactive power; for the operating characteristics of speed, its sub-features include turbine speed and generator speed; for the operating characteristics of load, its sub-features include unit active load, load change rate and load fluctuation amplitude;

[0033] Through the similarity matching algorithm, the real-time working condition characteristics are compared with the standard characteristic patterns in the typical working condition pattern library, the working condition similarity is calculated, and the degree of deviation between the current working condition and the typical working condition pattern is quantified;

[0034] Combining historical operating data and the operating requirements of thermal power units, multi-level matching thresholds of operating condition similarity are set, namely normal matching thresholds and abnormal matching thresholds. Then, the current operating condition status is analyzed to distinguish normal from abnormal conditions, thereby determining whether the current operating condition of the system is abnormal.

[0035] A further improvement of the technical solution of the present invention is that the calculation process of the working condition similarity is:

[0036] Extract operating characteristics reflecting the operating conditions of thermal power units from time series data sets, including temperature, pressure, flow, power, speed, and load, as well as sub-features of each operating characteristic;

[0037] For each sub-feature of the working condition characteristic, the absolute difference between its actual value and the reference value is calculated, and then the ratio of the absolute difference to the reference value is calculated to determine the relative deviation degree of each sub-feature;

[0038] For each sub-feature of the operating condition characteristic, calculate the integral of the rate of change of its actual value in the time interval [0, T], and combine it with the exponential function to calculate the penalty function for the deviation of the integral of the rate of change;

[0039] For each sub-feature of the working condition feature, the relative deviation degree and the integral deviation penalty function of the change rate are combined to analyze its contribution to the working condition similarity. The static deviation degree and dynamic change degree of the sub-feature are comprehensively considered to obtain the similarity contribution value of each sub-feature.

[0040] The similarity contribution values ​​of the sub-features of all working condition features are summed and averaged to calculate the working condition similarity, which is the overall working condition similarity, reflecting the degree of deviation between the current working condition and the typical working condition pattern.

[0041] A further improvement of the technical solution of the present invention is that: the distributed control module includes a collaborative control unit, a bat algorithm optimization unit and a control execution unit;

[0042] The collaborative control unit is used to distribute the control functions to the controllers corresponding to each subsystem. Each subcontroller independently completes the preliminary control of its subsystem and performs local collaborative control.

[0043] The bat algorithm optimization unit is used to use the bat algorithm to perform global optimization on the load changes and operating condition monitoring results of the thermal power units under the distributed control framework to find the optimal control parameters and scheduling scheme;

[0044] The control execution unit is used to generate control instructions based on the optimal control parameters and scheduling plan, perform real-time control of the various subsystems of the thermal power unit, adjust operating parameters, and achieve independent and coordinated control of the various subsystems through distributed control, thereby improving the flexibility and reliability of the system.

[0045] A further improvement of the technical solution of the present invention is that the bat algorithm optimization unit specifically includes:

[0046] Within a distributed control framework, the load changes and operating condition monitoring results of thermal power units are transformed into an optimization problem. The objective function and constraints are defined. The objective function is to maximize peak-shaving efficiency, and the constraints are load response rate, upper and lower limits of equipment output, and operating condition stability. Then, a bat population is initialized, with each individual representing a set of control parameters. An initial solution is generated through random sampling. Algorithm parameters, including the pulse frequency range, loudness decay rate, and search space boundary, are set to ensure that the population diversity covers the feasible solution region. The pulse frequency range is [0, 2], the loudness decay rate is 0.9, and the search space boundary is ±10% of the equipment rating.

[0047] Based on the biological characteristics of the bat algorithm, the search behavior of individuals in the solution space is simulated. Each bat adjusts the pulse frequency and loudness (initial value 1, decaying with iteration) through a dynamic update formula. A new solution is generated by combining the current global optimal solution. The speed update formula is used to update the bat's search speed in the solution space to explore the solution space. At the same time, a dynamic inertia weight is introduced. The weight is increased in the early stage of the search to enhance global exploration ability, and reduced in the later stage to accelerate local convergence. This balances global search and local development capabilities to ensure that all individuals always meet physical constraints.

[0048] The objective function value of each individual bat is calculated. The global optimal solution and the individual historical optimal solution are updated based on the objective function value. A greedy strategy is used to select the better solution, and the bat hunting behavior is simulated. If the new solution is better than the current solution and the loudness meets the conditions, the new solution is accepted and the loudness is reduced. Otherwise, the original solution is retained and the pulse frequency is adjusted. The population is updated iteratively to gradually approach the global optimal solution. At the same time, the convergence trend of each iteration is recorded to evaluate the efficiency of the algorithm. If the objective function value fluctuation is <0.1% for 5 consecutive iterations, the termination condition is triggered.

[0049] The feasibility of the final converged global optimal solution is verified to check whether the constraints (load response rate ≥ 2% / min, equipment output within the rated range) are met. If the verification is passed, the optimal control parameters and scheduling plan are output. If not, a local search is started or the population is reinitialized, and the optimization results are synchronized to the collaborative control unit to drive each sub-controller to adjust the operating parameters, thereby performing dynamic optimization of the load of the thermal power unit and adaptive adjustment of the operating conditions.

[0050] A further improvement of the technical solution of the present invention is that the control execution unit specifically includes:

[0051] The control execution unit receives the optimal control parameters and scheduling plan output by the bat algorithm optimization unit, parses them into executable control instructions for each subsystem, and decomposes the global optimization target into corresponding control parameters such as boiler fuel quantity, turbine valve opening, and generator power set value. It dynamically modifies the parameters based on the current operating conditions and generates an instruction set containing timing constraints and priority tags to ensure compatibility of the instructions with the distributed control node interface protocol.

[0052] Through distributed control, control instructions are distributed to each subsystem controller in real time, triggering independent control processes. Each subsystem performs closed-loop control based on local sensor data and performs autonomous adjustments to ensure that operating parameters quickly respond to instruction requirements. At the same time, the execution status is uploaded to the control execution unit.

[0053] The control execution unit continuously monitors the operating status of each subsystem of the thermal power unit, collects real-time operating data, and compares it with the optimized control parameters and scheduling plan. If deviations or abnormalities are found, the control instructions are adjusted and the operating parameters are corrected to ensure that the subsystem operates in an optimized state.

[0054] A further improvement of the technical solution of the present invention is that the security monitoring and early warning module specifically includes:

[0055] The safety monitoring and early warning module continuously monitors the key parameters of thermal power units during deep peak regulation, and conducts a comprehensive assessment of the operating results of the thermal power units to analyze whether the key parameters are within the normal operating range;

[0056] By comparing the current key parameters with the preset safety thresholds, it is determined whether the current operating status meets the safety requirements. Once any key parameter is found to exceed the preset safety threshold, the alarm mechanism is immediately triggered and an alarm message is issued. The alarm message is issued through various means such as sound and light alarms, SMS notifications, and system prompts to notify operators and maintenance personnel to perform corresponding maintenance measures. At the same time, relevant information of the alarm event is recorded;

[0057] Continuously monitor the status of thermal power units, track key parameters that issue alarms, and feed back the implementation of maintenance measures to the monitoring system until key parameters return to normal range.

[0058] Due to the adoption of the above technical solution, the present invention has the following technical advancements compared to the prior art:

[0059] 1. The present invention provides a deep peak-shaving control system for thermal power units. By introducing a distributed control architecture and bat algorithm optimization, the subsystems of the thermal power unit are divided into multiple local control units to achieve independent and coordinated control, significantly improving the flexibility and response speed of the thermal power unit during deep peak-shaving. Distributed control ensures that each subsystem can quickly adjust its operating parameters according to actual load requirements, and the global optimization capability of the bat algorithm further enhances the system's adaptability and decision-making efficiency in the face of complex working conditions, thereby effectively solving the problem of response lag under the traditional centralized control mode.

[0060] 2. The present invention provides a deep peak-shaving control system for thermal power units. It uses big data analysis and load forecasting technology, combined with the real-time and historical operating data of the thermal power units, to predict load change trends and formulate optimal control parameters and scheduling plans accordingly. This not only improves peak-shaving efficiency, but also reduces operating costs by reducing unnecessary energy waste and equipment loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0062] Figure 1 Schematic diagram of the system function modules of the present invention;

[0063] Figure 2 Schematic diagram of the workflow of the bat algorithm optimization unit of the present invention. DETAILED DESCRIPTION

[0064] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0065] Example 1, as Figure 1 、 Figure 2 As shown, the present invention provides a deep peak-shaving control system for a thermal power unit, including a deep peak-shaving control center, which is communicatively connected to a data acquisition module, a load forecasting module, a working condition monitoring module, a distributed control module, and a safety monitoring and early warning module, wherein electrical signals are connected between the modules;

[0066] The deep peak shaving control center, serving as the command center of the entire control system, is responsible for communication and coordination between modules, integrating information from each module and issuing instructions to achieve centralized management and unified control of the deep peak shaving process of thermal power units, ensuring the coordinated operation of each module.

[0067] The data acquisition module is used to collect and pre-process the operating data of each subsystem of the thermal power unit in real time, including key parameters such as temperature, pressure, flow, power, speed and load. Through the distributed arrangement of data collection points, the comprehensiveness and real-time nature of the data are ensured. According to the peak-shaving needs of the thermal power unit, the layout of the data collection points of each subsystem of the thermal power unit is planned, and the operating data to be collected is determined, including key parameters such as temperature, pressure, flow, power, speed and load. At the same time, the collection frequency and time interval are formulated, and the operation data collection plan is designed to ensure that the operating status of each subsystem can be fully covered. Among them, the temperature includes the main steam temperature, reheat steam temperature, exhaust temperature, bearing temperature, etc., the pressure includes the main steam pressure, reheat steam pressure, feed water pressure, lubricating oil pressure, etc., the flow includes the main steam flow, feed water flow, fuel flow, cooling water flow, etc., the power includes the generator output power, turbine shaft power, etc., and the speed Speed ​​includes turbine speed, feed water pump speed, etc., and load includes grid demand load, actual load of the unit, etc. According to the planned collection plan, distributed data collection points are used to collect the operating data of each subsystem of the thermal power unit. Each data collection point obtains the actual value of the corresponding key parameter through the configured sensor, and transmits the data to the data acquisition module in the form of electrical signals to ensure the real-time and accuracy of the data. The collected raw operating data is preprocessed, including the steps of data cleaning and normalization. Among them, through data cleaning, noise and outliers are removed, data is filtered, data curves are smoothed, and data is normalized to unify the data to the same dimension and range, thereby improving data quality and consistency. The preprocessed operating data is integrated to obtain a thermal power operation sequence table, and a data warehouse is built to store relevant data of the thermal power operation sequence table in the data warehouse;

[0068] The load forecasting module is used to analyze the load change trend of thermal power units using historical operation data. The historical operation data of the thermal power units, including power, load and corresponding timestamp information, are extracted from the operation records of the thermal power units. The data are sorted and aligned according to the time series to form a unified format. The historical operation data are feature analyzed to extract trend features related to load forecasting, namely load change rate, load fluctuation amplitude, load change duration and load change frequency. In combination with historical operation data and the operation requirements of the thermal power units, the benchmark values ​​of each trend feature are preset, and then each trend feature and the corresponding benchmark value are integrated to obtain a load trend feature sequence table. Among them, the load change rate is the change in load per unit time, in MW / min (megawatts / minute), reflecting the speed of load change, which is used to analyze the response ability of the thermal power unit to load fluctuations. The load fluctuation amplitude is the difference between the maximum and minimum load values ​​in a specific time period, usually in MW, quantifying the severity of load changes, which is used to analyze the fluctuation characteristics of the load. The load change duration is the time when the load changes from a certain state. The duration from one state to another, usually measured in minutes, is used to evaluate the persistence of load changes and is used for unit start-up and shutdown decisions and fuel supply planning. The load change frequency is the number of load changes per unit time, usually measured in times / hour, reflecting the frequency of load changes and used to evaluate fatigue damage and life of unit equipment. The load trend feature sequence table is divided into a training set, a validation set, and a test set. A basic architecture based on time series analysis is selected to build a load forecasting model. The trend features are input and the load change trend of the thermal power unit is output. The model is trained using the training set, and the model parameters are adjusted to optimize the prediction accuracy. The performance of the model is evaluated using the validation set to ensure that the model can accurately capture the trends and patterns of load changes. The constructed model is evaluated using the test set to improve the prediction accuracy and generalization ability of the model, thereby obtaining the final load forecasting model. The constructed load forecasting model is used to input the relevant data of the load trend feature sequence table of the current time period, calculate the load trend value of the unit, and predict the load change trend of the thermal power unit in the future period of the same time period.

[0069] The calculation process of the unit load trend value is:

[0070] Extract the trend feature related data of the thermal power unit in the current time period from the load trend feature sequence table, including the load change rate, load fluctuation amplitude, load change duration and load change frequency at each time point, and determine the reference value corresponding to each trend feature, namely the load change rate reference value, load fluctuation amplitude reference value, load change duration reference value and load change frequency reference value. For each time point, calculate the ratio of each trend feature to its reference value to reflect the high or low of the actual trend feature at that time point relative to the reference value, and obtain the relative change rate of each trend feature. For each time point, calculate the relative change rate average of the load change rate and load fluctuation amplitude. The square root of the sum of squares is used to combine the relative changes in the load change rate and the load fluctuation amplitude to form an indicator reflecting the joint effect of the two. This indicator is used as the numerator. For each time point, the result of adding the relative change rate of the load change duration and the relative change rate of the load change frequency is substituted into the exponential function to obtain the value of the denominator. The use of the exponential function emphasizes the influence of the load change duration and the load change frequency. The denominator as a whole reflects the comprehensive influence of the duration and frequency on the unit load change. The calculated numerator is divided by the denominator to obtain the exponential contribution value of each time point. The exponential contribution values ​​of all time points are then added together, and the sum is the unit load trend value.

[0071] The calculation expression of the unit load trend value is as follows:

[0072]

[0073] Where L is the load trend value of the unit, which reflects the comprehensive trend of the load change of the thermal power unit. The larger the value, the more drastic the load change. n is the number of sample points calculated in the time period. OCR i The load change rate at the i-th time point is expressed in MW / min, which indicates the load change per unit time and reflects the speed of load change. base It is the load change rate benchmark value, which is preset based on historical operating data and thermal power unit operating requirements and is used to measure the relative level of the actual load change rate. i The load fluctuation amplitude at the i-th time point is expressed in MW, which represents the difference between the maximum and minimum load values ​​in a specific time period and quantifies the severity of the load change. OVA base is the load fluctuation amplitude reference value, OCD i The duration of the load change at the i-th time point, in minutes, represents the duration of the load change from one state to another. OCD base OCF is the load change duration reference value. i The load change frequency at the i-th time point is times / hour, which means the number of load changes per unit time and reflects the frequency of load changes. OCFbase The index is the load change frequency benchmark value. When all actual load characteristics are far below the benchmark value, the index is close to 0, indicating that the load change is gentle. When the actual load characteristics are close to or exceed the benchmark value, the index gradually increases, indicating that the load change is drastic. When the unit load trend value is low, it indicates that the load change of the thermal power unit is relatively stable and the unit is in a relatively stable operating state. At this time, the unit's operating risk is low, and the intensity of peak regulation can be appropriately reduced to improve operating efficiency. When the unit load trend value is high, the load change is drastic, and the unit needs to pay close attention and may need to take more active peak regulation measures.

[0074] The operating condition monitoring module is used to monitor the operating conditions of thermal power units, identify abnormalities in the current operating conditions, and help the system adopt adjustment strategies based on the operating conditions, thereby improving the accuracy and efficiency of the peak regulation process. It extracts operating data collected from various subsystems of the thermal power unit, including key parameters such as temperature, pressure, flow, power, speed, and load, and synchronizes the data of different subsystems through timestamps. It uses Z-score standardization or Min-Max normalization to eliminate dimensional differences, unifies heterogeneous data into time series data sets, ensures data time series consistency and comparability, extracts operating condition features reflecting the operating conditions of the thermal power unit from the time series data sets, and then determines the sub-features of each operating condition feature. It also uses clustering algorithms to classify historical operating data and analyze historical operating conditions. The standard characteristic patterns under normal operating conditions and row data are established, a typical operating pattern library is established, and the benchmark values ​​corresponding to the sub-features of each operating condition characteristic are determined. Among them, for the operating characteristics of temperature, its sub-features include main steam temperature, reheat steam temperature, condenser temperature and bearing temperature. The main steam temperature reflects the efficiency of boiler combustion and steam generation, as well as the operating conditions of the turbine. The reheat steam temperature affects the efficiency and operating safety of the turbine. The condenser temperature affects the back pressure and cycle efficiency of the turbine. The bearing temperature reflects the mechanical operating status of the unit. Too high may indicate lubrication or alignment problems. For the operating characteristics of pressure, its sub-features include main steam pressure, condenser pressure and fuel pressure. The main steam pressure affects the steam flow and the output power of the turbine. The condenser pressure The pressure and condenser temperature jointly affect the thermal efficiency of the steam turbine. The fuel pressure affects the working state and combustion efficiency of the fuel supply system. For the flow operating characteristics, its sub-characteristics include fuel flow, feed water flow and air flow. The fuel flow is directly related to the combustion intensity and steam generation of the boiler. The feed water flow affects the water circulation and steam generation of the boiler. The air flow affects the combustion efficiency and pollutant emissions. For the power operating characteristics, its sub-characteristics include turbine output power, generator output power and reactive power. The turbine output power reflects the power generation capacity and operating efficiency of the unit. The generator output power and turbine power jointly reflect the overall power generation performance of the unit. The reactive power reflects the unit's ability to support the grid voltage. For the speed operating characteristics, its sub-characteristics include turbine output power, generator output power and reactive power. The characteristics include turbine speed and generator speed. Turbine speed affects the output power and mechanical stability of the turbine. Generator speed is usually synchronized with turbine speed and affects the power generation frequency. For the load operating characteristics, its sub-characteristics include unit active load, load change rate and load fluctuation amplitude. Unit active load reflects the grid active power demand borne by the unit, load change rate reflects the speed and flexibility of unit load adjustment, and load fluctuation amplitude reflects the stability of unit load. Through the similarity matching algorithm, the real-time operating condition characteristics are compared with the standard feature patterns in the typical operating condition pattern library, the operating condition similarity is calculated, and the degree of deviation between the current operating condition and the typical operating condition pattern is quantified. In combination with the sliding window technology, the trend of similarity change in the time dimension is analyzed.Identify short-term fluctuations and long-term deviations, generate a heat map of operating condition deviations, and intuitively display abnormal tendencies in each feature dimension. Combined with historical operating data and the operating requirements of thermal power units, set multi-level matching thresholds for operating condition similarity, namely normal matching thresholds and abnormal matching thresholds. Then, analyze the current operating condition to distinguish between normal and abnormal conditions, and thus determine whether the current system operating condition is abnormal.

[0075] The calculation process of working condition similarity is:

[0076] The operating characteristics reflecting the operating conditions of the thermal power unit are extracted from the time series data set, including temperature, pressure, flow, power, speed and load, as well as the sub-features of each operating characteristic. For each sub-feature of the operating characteristic, the absolute difference between its actual value and the reference value is calculated, and then the ratio of the absolute difference to the reference value is calculated to determine the relative deviation degree of each sub-feature. By dividing by the reference value, the influence of different feature dimensions is eliminated, making the deviation degree of different features comparable. For each sub-feature of the operating characteristic, the change rate integral of its actual value in the time interval [0, T] is calculated, and combined with the exponential function, the change rate integral deviation penalty function is calculated. The change rate integral reflects the sub-feature in the time interval. The greater the rate of change, the greater the integral value, indicating that the sub-feature has a greater fluctuation within the time interval. The value of the exponential function decreases as the integral value increases, thereby giving a greater deviation penalty to features with large changes. For each sub-feature of the working condition feature, the relative deviation degree and the rate of change integral deviation penalty function are combined to analyze its contribution to the working condition similarity. The static deviation degree and dynamic change degree of the sub-feature are comprehensively considered to obtain the similarity contribution value of each sub-feature. The similarity contribution values ​​of the sub-features of all working condition features are summed and averaged to calculate the working condition similarity, which is the overall working condition similarity, reflecting the degree of deviation between the current working condition and the typical working condition mode.

[0077] The calculation expression of working condition similarity is as follows:

[0078]

[0079] Where S is the similarity of the working condition, which is used to quantify the degree of deviation between the current working condition and the typical working condition mode. The smaller the value, the smaller the deviation and the closer the working condition is to normal. N is the total number of sub-features of the working condition feature, and F is the number of sub-features of the working condition feature. j is the actual value of the sub-feature of the j-th working condition feature, F base，j is the sub-feature reference value of the j-th operating condition feature, that is, the standard value under normal operating conditions, which is used to measure the deviation degree of the actual value. j(t) is a function that changes with time, which represents the actual value of the sub-feature of the j-th operating condition feature at time t. T is the upper bound of the time interval, that is, the end time of the considered time range. The value range of S is between 0 and 1. When S is close to 0, it means that the current operating condition is highly consistent with the typical operating condition mode, and the deviation is extremely small. When S is close to 1, it means that there is a large deviation between the current operating condition and the typical operating condition mode, and the operating condition is abnormal. The normal operating condition is that under normal operating conditions, the actual value of the sub-feature of each operating condition feature is close to the benchmark value, the relative deviation is small, and the rate of change is also small. Therefore, the S value is close to 0. The abnormal operating condition is when one or more sub-features of the operating condition feature are abnormal, the relative deviation and the rate of change increase, the S value will increase significantly, exceed the set normal matching threshold, and enter the abnormal state;

[0080] The distributed control module is used to divide the various subsystems of the thermal power unit into multiple local control units. Each control unit is independently adjusted and coordinated with other units. It also uses the bat algorithm to analyze the load changes and operating condition monitoring results of the thermal power unit to find the optimal control parameters and scheduling plan, and then generate and execute control instructions;

[0081] The safety monitoring and early warning module is used to evaluate the operating performance of thermal power units and continuously monitor various key parameters during the deep peak regulation process of thermal power units. Once it is found that the key parameters exceed the safety threshold, the alarm mechanism will be triggered and corresponding safety protection measures will be taken.

[0082] Example 2, as Figure 1 、 Figure 2 As shown, based on Example 1, the present invention provides a technical solution: preferably, the distributed control module includes a collaborative control unit, a bat algorithm optimization unit and a control execution unit;

[0083] Among them, the collaborative control unit is used to distribute the control functions to the controllers corresponding to each subsystem. Each subcontroller independently completes the preliminary control of its subsystem and performs local collaborative control.

[0084] The collaborative control unit specifically includes: based on a hierarchical distributed architecture, decomposing the overall control task of the thermal power unit into multiple subtasks, including boiler combustion control, turbine load regulation, generator power balancing, and fuel supply management, and assigning the subtasks to the controllers corresponding to each subsystem. Each subcontroller is logically isolated through an independent computing unit to ensure that the basic operation of its subsystem can be maintained in the event of communication interruption or failure. The task allocation stage clearly defines the scope of authority of each subcontroller (for example, the boiler controller only adjusts the fuel amount and air volume), and establishes a redundant backup mechanism to ensure the reliable execution of control tasks. Each subcontroller performs preliminary control of its subsystem according to a preset control strategy (PID control). The subcontroller adjusts the operating parameters of the subsystem based on real-time collected operating data and the preset control strategy to meet the basic operating requirements of the subsystem. The goal of preliminary control is to ensure that each subsystem operates within the normal range. After the preliminary control is completed, the subcontrollers exchange operating status information through a standardized communication interface (CAN bus) and adjust their own operating parameters according to the needs of adjacent subsystems or related subsystems to achieve local collaborative operation.

[0085] The bat algorithm optimization unit is used to use the bat algorithm to globally optimize the load changes and operating condition monitoring results of thermal power units under the distributed control framework, find the optimal control parameters and scheduling scheme, and transform the load changes and operating condition monitoring results of thermal power units into optimization problems under the distributed control framework. The objective function and constraints are defined. The objective function is to maximize the peak-shaving efficiency, and the constraints are the load response rate, the upper and lower limits of the equipment output, and the operating condition stability. Then, the bat population is initialized, and each individual represents a set of control parameters. The initial solution is generated by random sampling, and the algorithm parameters including the pulse frequency range, the loudness attenuation rate, and the search space boundary are set to ensure the population Diversity covers the feasible solution area, with a pulse frequency range of [0, 2], a loudness decay rate of 0.9, and a search space boundary of ±10% of the device rating. Based on the biological characteristics of the bat algorithm, the search behavior of individuals in the solution space is simulated. Each bat adjusts the pulse frequency and loudness (initial value 1, decaying with iteration) through a dynamic update formula. A new solution is generated based on the current global optimal solution. The speed update formula is used to update the bat's search speed in the solution space to explore the solution space. At the same time, a dynamic inertia weight is introduced. The weight is increased in the early stage of the search to enhance global exploration capabilities, and the weight is reduced in the later stage to accelerate local convergence. This balances global search and local development capabilities to ensure that all individuals always meet physical constraints.

[0086] The calculation expression of the dynamic update formula is as follows:

[0087] f k =f min +(f max -f min )·rand;

[0088] Where, f k is the current pulse frequency of the k-th bat, f min is the minimum value of the bat pulse frequency, usually 0 or close to 0, f max is the maximum value of the bat pulse frequency, usually 2, rand is a random number uniformly distributed in the interval [0,1], f k The value range is [f min , f max ], as the iterations proceed, the bats will gradually tend to use higher frequencies for more refined searches;

[0089] The calculation expression of the speed update formula is as follows:

[0090]

[0091] Where, is the speed of the k-th bat in the u+1 generation, ω is the dynamic inertia weight, which is used to balance the global search and local development capabilities. Its value range is usually between [0,1] and will gradually decrease as the iteration proceeds. is the speed of the k-th bat in the u-th generation, x best is the position of the current global optimal solution, is the position of the k-th bat in the u-th generation. In the early stage of the bat algorithm, the dynamic inertia weight ω is large, which enables the bat to maintain a large speed, enhance the global search ability, and explore a wider solution space. As the iteration proceeds, ω gradually decreases, and the bat's speed will also decrease accordingly, thereby accelerating local convergence and improving search accuracy. During the search process, the bat's speed will be dynamically adjusted according to the distance between the current position and the global optimal solution, the pulse frequency, and the dynamic inertia weight to balance the global search and local development capabilities;

[0092] Calculate the objective function value of each bat individual, update the global optimal solution and individual historical optimal solution based on the objective function value, adopt a greedy strategy to select the better solution, and simulate the bat hunting behavior. If the new solution is better than the current solution and the loudness meets the conditions, accept the new solution and reduce the loudness. Otherwise, retain the original solution and adjust the pulse frequency. Update the population through iteration, gradually approach the global optimal solution, and record the convergence trend of each iteration to evaluate the efficiency of the algorithm. If the objective function value fluctuates by <0.1% for 5 consecutive iterations, the termination condition is triggered, and the feasibility of the final converged global optimal solution is verified to check whether the constraints (load response rate ≥2% / min, equipment output within the rated range) are met. If the verification passes, the optimal control parameters and scheduling plan are output. If not, start local search or reinitialize the population, synchronize the optimization results to the collaborative control unit, drive each sub-controller to adjust the operating parameters, and perform dynamic optimization of the load of the thermal power unit and adaptive adjustment of the operating conditions;

[0093] The calculation expression of the objective function value of each bat individual is as follows:

[0094]

[0095] Where Fit is the objective function value of each bat individual, indicating the degree of optimization of peak-shaving efficiency. The larger the value, the higher the peak-shaving efficiency. K is the number of load demand points, and P is the number of load demand points. p is the actual output of the pth load demand point, in MW, P base，p is the benchmark output of the pth load demand point, in MW, indicating the output of the point under normal operating conditions, M is the number of equipment, R o is the load response rate of the oth device, in MW / min, R base，o is the benchmark load response rate of the oth device, in MW / min, indicating the load response rate of the device under normal operating conditions. λ is the weight coefficient, which is used to balance the contribution of load demand and device output constraints. Its value range is usually between 0 and 1. When the objective function value is close to 0, it means that the peak-shaving efficiency is extremely low, and the constraints of load demand and device output are not met. When the objective function value gradually increases, it means that the peak-shaving efficiency is increasing, and the constraints of load demand and device output are met.

[0096] The control execution unit is used to generate control instructions based on the optimal control parameters and scheduling plan, perform real-time control of each subsystem of the thermal power unit, adjust operating parameters, and achieve independent and coordinated control of each subsystem through distributed control, thereby improving the flexibility and reliability of the system. The control execution unit receives the optimal control parameters and scheduling plan output by the bat algorithm optimization unit, parses them into executable control instructions for each subsystem, and decomposes the global optimization target into corresponding control parameters such as boiler fuel quantity, turbine valve opening, and generator power set value. It dynamically modifies the parameters based on the current operating conditions and generates an instruction set containing timing constraints and priority tags to ensure that the instructions are compatible with the distributed control node interface protocol. The control instructions are distributed to the controllers of each subsystem in real time through distributed control, triggering independent control processes. Each subsystem performs closed-loop control based on local sensor data and performs autonomous adjustment to ensure that operating parameters quickly respond to instruction requirements. The execution status is also uploaded to the control execution unit. The control execution unit continuously monitors the operating status of each subsystem of the thermal power unit, collects real-time operating data, and compares it with the optimized control parameters and scheduling plan. If deviations or abnormalities are found, the control instructions are adjusted and the operating parameters are corrected to ensure that the subsystem operates in an optimized state.

[0097] The security monitoring and early warning module specifically includes:

[0098] The safety monitoring and early warning module continuously monitors the key parameters of the thermal power unit during the deep peak regulation process, and conducts a comprehensive assessment of the operating effect of the thermal power unit, analyzes whether the key parameters are within the normal operating range, and judges whether the current operating status meets the safety requirements by comparing the current key parameters with the preset safety thresholds. Once any key parameter is found to exceed the preset safety threshold, the alarm mechanism is immediately triggered and an alarm message is issued. The alarm message is issued through various means such as sound and light alarms, SMS notifications, system prompts, etc., to notify operators and maintenance personnel to perform corresponding maintenance measures. At the same time, the relevant information of the alarm event is recorded, the status of the thermal power unit is continuously monitored, the key parameters that issued the alarm are tracked, and the implementation of the maintenance measures is fed back to the monitoring system until the key parameters return to the normal range.

[0099] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A deep peak shaving control system for a thermal power unit, comprising a deep peak shaving control center, characterized in that: The deep peak shaving control center is connected to the data acquisition module, load forecasting module, working condition monitoring module, distributed control module and safety monitoring and early warning module, wherein the electrical signal connection between each module; The data acquisition module is used to collect and pre-process the operating data of each subsystem of the thermal power unit in real time, including key parameters such as temperature, pressure, flow, power, speed and load; The load forecasting module is used to analyze the load change trend of the thermal power unit using historical operation data; The operating condition monitoring module is used to monitor the operating condition of the thermal power unit and identify abnormalities in the current operating condition; The distributed control module is used to divide the various subsystems of the thermal power unit into multiple local control units, and use the bat algorithm to analyze the load changes and operating condition monitoring results of the thermal power unit to find the optimal control parameters and scheduling plan, and then generate and execute control instructions; The safety monitoring and early warning module is used to continuously monitor various key parameters during the deep peak regulation process of the thermal power unit. If the safety threshold is exceeded, the alarm mechanism is triggered and corresponding safety protection measures are taken.

2. A deep peak regulation control system for thermal power units according to claim 1, characterized in that: The data acquisition module specifically includes: Based on the peak-shaving requirements of thermal power units, plan the data collection point layout for each subsystem of the thermal power units, determine the operating data to be collected, including key parameters such as temperature, pressure, flow, power, speed, and load, and establish the collection frequency and time interval, and design an operating data collection plan; According to the planned data collection plan, distributed data collection points are used to collect operating data from various subsystems of the thermal power unit. Each data collection point obtains the actual value of the corresponding key parameters through the configured sensors and transmits the data to the data acquisition module in the form of electrical signals; Preprocess the collected raw operating data, including data cleaning and normalization steps; The pre-processed operation data is integrated to obtain the thermal power operation sequence table, and a data warehouse is built to store the relevant data of the thermal power operation sequence table in the data warehouse.

3. A deep peak regulation control system for thermal power generation units according to claim 1, characterized in that: The load forecasting module specifically includes: Extract historical operating data of thermal power units from their operation records, including power, load, and corresponding timestamp information, and sort and align the data in time series. Perform feature analysis on historical operating data to extract trend features related to load forecasting, including load change rate, load fluctuation amplitude, load change duration, and load change frequency. Combined with historical operating data and thermal power unit operating requirements, baseline values ​​for each trend feature are preset. Then, each trend feature and its corresponding baseline value are integrated to obtain a load trend feature sequence table. The load trend feature sequence table is divided into a training set, a validation set, and a test set. A time series analysis-based infrastructure is selected to build a load forecasting model that inputs trend features and outputs the load change trend of thermal power units. By using the constructed load forecasting model, the relevant data of the load trend characteristic sequence table of the current time period is input, the load trend value of the computer group is calculated, and the load change trend of the thermal power unit in the future period of the same time period is predicted.

4. A deep peak regulation control system for thermal power generation units according to claim 3, characterized in that: The calculation process of the unit load trend value is as follows: Extracting trend feature-related data of the thermal power unit in the current time period from the load trend feature sequence table, including the load change rate, load fluctuation amplitude, load change duration, and load change frequency at each time point, and determining the reference values ​​corresponding to each trend feature, namely, the load change rate reference value, the load fluctuation amplitude reference value, the load change duration reference value, and the load change frequency reference value; For each time point, the ratio of each trend feature to its baseline value was calculated to obtain the relative change rate of each trend feature; For each time point, calculate the square root of the sum of the squares of the load change rate and the relative change rate of the load fluctuation amplitude as the numerator; For each time point, the sum of the relative change rate of the load change duration and the relative change rate of the load change frequency is substituted into the exponential function to obtain the value of the denominator. Divide the calculated numerator by the denominator to obtain the exponential contribution value at each time point, and then add up the exponential contribution values ​​of all time points. The sum result is the unit load trend value.

5. A deep peak regulation control system for thermal power generation units according to claim 1, characterized in that: The working condition monitoring module specifically includes: Extract operational data collected from various subsystems of thermal power units, including key parameters such as temperature, pressure, flow, power, speed, and load. Then, through timestamp alignment and standardization, unify data in different formats into a time series dataset. Extract the operating characteristics reflecting the operating conditions of the thermal power unit from the time series data set, and then determine the sub-features of each operating characteristic. Use the clustering algorithm to classify the historical operating data, analyze the historical operating data and the standard feature patterns under normal conditions, establish a typical operating mode library, and determine the benchmark values ​​corresponding to the sub-features of each operating characteristic. Among them, for the operating characteristics of temperature, its sub-features include main steam temperature, reheat steam temperature, condenser temperature and bearing temperature; for the operating characteristics of pressure, its sub-features include main steam pressure, condenser pressure and fuel pressure; for the operating characteristics of flow, its sub-features include fuel flow, feed water flow and air flow; for the operating characteristics of power, its sub-features include turbine output power, generator output power and reactive power; for the operating characteristics of speed, its sub-features include turbine speed and generator speed; for the operating characteristics of load, its sub-features include unit active load, load change rate and load fluctuation amplitude; Through the similarity matching algorithm, the real-time working condition characteristics are compared with the standard characteristic patterns in the typical working condition pattern library, the working condition similarity is calculated, and the degree of deviation between the current working condition and the typical working condition pattern is quantified; Combining historical operating data and the operating requirements of thermal power units, multi-level matching thresholds of operating condition similarity are set, namely normal matching thresholds and abnormal matching thresholds. Then, the current operating condition status is analyzed to distinguish normal from abnormal conditions, thereby determining whether the current operating condition of the system is abnormal.

6. A deep peak regulation control system for thermal power generation units according to claim 5, characterized in that: The calculation process of the working condition similarity is: Extract the operating characteristics reflecting the operating conditions of the thermal power units and the sub-characteristics of each operating characteristic from the time series data set; For each sub-feature of the working condition characteristic, the absolute difference between its actual value and the reference value is calculated, and then the ratio of the absolute difference to the reference value is calculated to determine the relative deviation degree of each sub-feature; For each sub-feature of the operating condition characteristic, calculate the integral of the rate of change of its actual value within the time interval, and combine it with the exponential function to calculate the penalty function for the deviation of the integral of the rate of change; For each sub-feature of the working condition feature, the relative deviation degree and the integral deviation penalty function of the change rate are combined to analyze its contribution to the working condition similarity and obtain the similarity contribution value of each sub-feature; The similarity contribution values ​​of the sub-features of all working condition features are summed and averaged to calculate the working condition similarity, which is the overall working condition similarity, reflecting the degree of deviation between the current working condition and the typical working condition pattern.

7. A deep peak regulation control system for thermal power generation units according to claim 1, characterized in that: The distributed control module includes a collaborative control unit, a bat algorithm optimization unit and a control execution unit; The collaborative control unit is used to distribute the control functions to the controllers corresponding to each subsystem and perform local collaborative control; The bat algorithm optimization unit is used to use the bat algorithm to perform global optimization on the load changes and operating condition monitoring results of the thermal power units to find the optimal control parameters and scheduling solutions; The control execution unit is used to generate control instructions based on the optimal control parameters and scheduling plan, perform real-time control on various subsystems of the thermal power unit, and adjust operating parameters.

8. A deep peak regulation control system for thermal power generation units according to claim 7, characterized in that: The bat algorithm optimization unit specifically includes: Within a distributed control framework, the load changes and operating condition monitoring results of thermal power units are transformed into an optimization problem, and the objective function and constraints are defined. The objective function is to maximize the peak-shaving efficiency, and the constraints are the load response rate, the upper and lower limits of equipment output, and the operating condition stability. Then, a bat population is initialized, with each individual representing a set of control parameters. An initial solution is generated through random sampling, and algorithm parameters are set, including the pulse frequency range, loudness decay rate, and search space boundary. Based on the biological characteristics of bat algorithms, the algorithm simulates the search behavior of individuals in the solution space. Each bat adjusts the pulse frequency and loudness through a dynamic update formula, combines the current global optimal solution to generate a new solution, and uses the speed update formula to update the bat's search speed in the solution space to explore the solution space. At the same time, a dynamic inertia weight is introduced to balance global search and local development capabilities. The objective function value of each individual bat is calculated. The global optimal solution and the individual historical optimal solution are updated based on the objective function value. A greedy strategy is used to select the better solution, and the bat hunting behavior is simulated. If the new solution is better than the current solution and the loudness meets the conditions, the new solution is accepted and the loudness is reduced. Otherwise, the original solution is retained and the pulse frequency is adjusted. The population is updated iteratively to gradually approach the global optimal solution. At the same time, the convergence trend of each iteration is recorded to evaluate the efficiency of the algorithm. If the objective function value fluctuation is <0.1% for 5 consecutive iterations, the termination condition is triggered. The feasibility of the final converged global optimal solution is verified to check whether the constraints are met. If the verification is passed, the optimal control parameters and scheduling plan are output. If not, a local search is started or the population is reinitialized, and the optimization results are synchronized to the collaborative control unit to drive each sub-controller to adjust the operating parameters, and perform dynamic optimization of the thermal power unit load and adaptive adjustment of the operating conditions.

9. A deep peak regulation control system for thermal power generation units according to claim 8, characterized in that: The control execution unit specifically includes: The control execution unit receives the optimal control parameters and scheduling scheme output by the bat algorithm optimization unit, parses them into control instructions that can be executed by each subsystem, decomposes the global optimization target into corresponding control parameters, and dynamically modifies them based on the current working condition characteristics to generate an instruction set containing timing constraints and priority tags; Through distributed control, control instructions are distributed to each subsystem controller in real time, triggering independent control processes. Each subsystem performs closed-loop control based on local sensor data, performs autonomous adjustments, and uploads the execution status to the control execution unit. The control execution unit continuously monitors the operating status of each subsystem of the thermal power unit, collects real-time operating data, and compares it with the optimized control parameters and scheduling plan. If deviations or abnormalities are found, the control instructions are adjusted and the operating parameters are corrected.

10. A deep peak regulation control system for thermal power generation units according to claim 1, characterized in that: The security monitoring and early warning module specifically includes: The safety monitoring and early warning module continuously monitors the key parameters of thermal power units during deep peak regulation, and conducts a comprehensive assessment of the operating results of the thermal power units to analyze whether the key parameters are within the normal operating range; By comparing the current key parameters with the preset safety thresholds, it is determined whether the current operating status meets the safety requirements. Once any key parameter is found to exceed the preset safety threshold, the alarm mechanism is immediately triggered and an alarm message is issued to notify the operator and maintenance personnel to perform the corresponding maintenance measures. At the same time, relevant information of the alarm event is recorded; Continuously monitor the status of thermal power units, track key parameters that issue alarms, and feed back the implementation of maintenance measures to the monitoring system until key parameters return to normal range.

Images