An adaptive precise regulation method for boiler combustion in a thermal power plant centralized control operation

Abstract

The application discloses a kind of self-adaptive precision regulation and control methods of boiler combustion in thermal power plant centralized control operation, real-time data acquisition and signal processing, real-time acquisition combustion process parameter;Data preprocessing, for key parameters, eliminate outliers, for missing data, data repair;Combustion state evaluation, evaluate current combustion state;Multi-objective optimization calculation, with the highest combustion efficiency, the lowest nitrogen dioxide emission and the smallest boiler slagging risk as optimization goal, establish multi-objective optimization function;Parameter adjustment and dynamic feedforward compensation, the optimal operating parameter set value is issued to adjust loop, when load instruction changes, execute feedforward compensation.The application realizes data acquisition by constructing hierarchical data acquisition network and differentiating sampling strategy;By multi-objective optimization algorithm, the contradiction between economy, environmental protection and safety is solved, and the optimal combination of operating parameters is output.

Description

Technical Field

[0001] This invention relates to the field of adaptive control technology for boiler combustion in power plants, and more particularly to an adaptive and precise control method for boiler combustion in centralized control operation of thermal power plants. Background Technology

[0002] Traditional boiler combustion control methods mainly employ fixed-parameter PID control or simple fuzzy logic control. These methods often struggle to achieve real-time and precise adjustment of combustion parameters when faced with changes in coal type or load fluctuations, leading to problems such as decreased combustion efficiency and increased pollutant emissions.

[0003] A boiler combustion optimization method is disclosed in invention patent application number 202510291960.6. This method utilizes a fuel quantity optimization model built based on load change rate, combined with accurate load forecasting results, to adjust the fuel delivery volume in real time, ensuring precise matching with actual load demand. It also considers factors such as boiler thermal efficiency, maintaining high thermal efficiency while ensuring load supply, reducing ineffective fuel consumption, and thus lowering power generation costs. An air quantity optimization model based on flue gas oxygen content ensures the optimal ratio of air to fuel. Appropriate air quantity allows for complete fuel combustion, reducing the generation of incomplete combustion products (such as carbon monoxide) and avoiding heat loss due to excessive air. (For example, if the flue gas carries away too much heat), it improves the energy utilization rate of fuel and further enhances combustion efficiency; by dynamically adjusting the coal-water ratio, it can stabilize the steam generation process and ensure that parameters such as steam pressure and temperature meet the requirements; by comprehensively considering multiple factors such as energy input, output and heat storage energy changes through the load prediction model, it can accurately predict the load change trend at a preset time in the future; based on this prediction, the fuel quantity, air quantity and coal-water ratio can be optimized in a timely manner, enabling the boiler combustion system to respond quickly to load changes and reduce system instability caused by load fluctuations; it effectively shortens the time interval from load change to combustion parameter adjustment, reduces the adverse effects of lag on unit operation, and improves the flexibility and adjustability of the unit;

[0004] However, the aforementioned boiler combustion optimization method only adjusts the parameters through a preset rule table, lacking real-time identification of the combustion state and adaptive parameter matching. This method does not fully consider the coordinated control of combustion efficiency and emissions of multiple pollutants. These methods often perform poorly when faced with complex and ever-changing actual operating conditions. Summary of the Invention

[0005] The main objective of this invention is to provide an adaptive and precise control method for boiler combustion in the centralized control operation of thermal power plants, which can solve the problems existing in the prior art.

[0006] Another objective of this invention is to provide an adaptive and precise control device for boiler combustion in the centralized control operation of a thermal power plant.

[0007] To achieve the above objectives, a first aspect of the present invention proposes an adaptive and precise control method for boiler combustion in centralized control operation of a thermal power plant, comprising: S1, key and non-key parameters of the boiler combustion process are acquired in real time through a hierarchical data acquisition network, and the key and non-key parameters are isolated, filtered and outlier removed based on a differentiated sampling strategy. S2, based on the processed combustion process parameters, uses the entropy weight method to evaluate the current combustion state from three dimensions: economy, environmental protection and safety, and divides the combustion state into four levels according to the comprehensive evaluation value to trigger optimization calculation; S3. Establish a multi-objective optimization function that includes maximizing combustion efficiency, minimizing nitrogen oxide emissions, and minimizing slagging risk. Solve the multi-objective optimization function using a multi-objective optimization algorithm and output the optimal combination of operating parameters. S4, the combined operating parameters are sent to the regulating loop of the boiler DCS system, and dynamic feedforward compensation is performed when the load command changes. The dynamic feedforward compensation includes adjusting the air supply and coal feed in advance according to the load change trend to maintain the optimal air-coal ratio.

[0008] In one embodiment of the present invention, S1 further includes: S11, the sampling period for key parameters is 2-5 seconds, and the sampling period for non-key parameters is 30-60 seconds. Outliers are removed using the 3σ criterion, specifically, when the deviation of a data point from the mean exceeds 3 times the standard deviation, it is considered an outlier and removed. The key parameters include flue gas oxygen content, furnace negative pressure, main steam flow rate and pressure, and coal feeder speed. The non-key parameters include coal elemental analysis, fly ash carbon content, and flue gas temperature. S12 performs moving average filtering on key parameters, with a sliding window length of 5-10 sampling points, to eliminate random measurement noise and smooth the data curve.

[0009] In one embodiment of the present invention, S2 further includes: S21, when calculating the weights of each indicator using the entropy weight method, the information entropy formula is used. Calculate information entropy value ,in For the first The first sample Individual indicator values, The total number of samples; S22, based on the comprehensive evaluation value Classifying combustion status levels, among which For the first Entropy weighting of each indicator For the first Standardized scores for each indicator This represents the total number of indicators.

[0010] In one embodiment of the present invention, S3 further includes: S31, the multi-objective optimization function is solved using the NSGA-II algorithm, with the combustion efficiency derivative as the solution. Nitrogen oxide emission concentration Slag Risk Index To optimize the objective, a Pareto optimal solution set is generated; S32, Select the optimal combination of operating parameters from the Pareto optimal solution set. The selection of the optimal solution is based on the weighted summation method. ,in The preset weighting coefficients are used for economic efficiency, environmental protection, and safety objectives.

[0011] In one embodiment of the present invention, S4 further includes: S41, in the dynamic feedforward compensation, the pre-adjustment ratio of the air supply volume when the load increases is the current load change rate. The coal feed rate is adjusted by a factor of 1:1, which is equal to the current load change rate. times, of which and These are compensation coefficients trained based on historical data; S42, the pre-adjustment ratio of the supply air volume when the load decreases is the current load change rate. times, and This is to prevent excessive air coefficient from causing decreased efficiency and increased nitrogen oxide generation.

[0012] In one embodiment of the present invention, it further includes: S5. The constraints of the multi-objective optimization function are dynamically adjusted based on the real-time monitored coal element analysis data. Specifically, when the changes in the coal element analysis data exceed the preset threshold, the entropy weight method weights are recalculated and the parameter boundaries of the multi-objective optimization function are updated.

[0013] To achieve the above objectives, a second aspect of the present invention provides an adaptive and precise control device for boiler combustion in a thermal power plant's centralized control operation, comprising: The data acquisition and preprocessing module is used to acquire key and non-key parameters of the boiler combustion process in real time through a hierarchical data acquisition network, and to perform signal isolation, filtering and outlier removal based on a differentiated sampling strategy. The multi-dimensional combustion state assessment module is used to assess the current combustion state from three dimensions: economy, environmental protection and safety, based on the processed combustion process parameters and using the entropy weight method. The combustion state is divided into four levels according to the comprehensive evaluation value to trigger optimization calculations. The multi-objective optimization parameter generation module is used to establish a multi-objective optimization function that includes maximizing combustion efficiency, minimizing nitrogen oxide emissions, and minimizing slagging risk. The multi-objective optimization function is solved by a multi-objective optimization algorithm, and the optimal combination of operating parameters is output. The dynamic feedforward compensation execution module is used to send the combination of operating parameters to the regulating loop of the boiler DCS system, and to perform dynamic feedforward compensation when the load command changes. The dynamic feedforward compensation includes pre-adjusting the air supply and coal feed according to the load change trend to maintain the optimal air-coal ratio.

[0014] The adaptive and precise control method and device for boiler combustion in the centralized control operation of thermal power plants according to embodiments of the present invention acquires data by constructing a hierarchical data acquisition network and a differentiated sampling strategy; and outputs the comprehensive optimal combination of operating parameters by collaboratively solving the contradictory objectives between economy, environmental protection and safety through a multi-objective optimization algorithm.

[0015] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.

[0016] The beneficial effects of this invention are as follows: This invention achieves comprehensive perception and high-quality data acquisition of multi-dimensional parameters of boiler combustion by constructing a hierarchical data acquisition network and a differentiated sampling strategy, laying a solid foundation for precise control and enabling real-time diagnosis and early warning of combustion status. Through multi-objective optimization algorithms, it collaboratively solves the conflicting objectives of economy, environmental protection, and safety, outputting the optimal combination of operating parameters. By combining a dynamic feedforward compensation mechanism, it proactively pre-adjusts the air-fuel ratio during load changes, significantly improving the system's response speed and anti-interference capability. This effectively overcomes the shortcomings of traditional control methods, such as fixed parameters, slow response, and difficulty in simultaneously addressing multiple objectives. Attached Figure Description

[0017] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 A flowchart illustrating an adaptive and precise control method for boiler combustion in a centralized control system of a thermal power plant, provided by an embodiment of the present invention; Figure 2 A flowchart illustrating another adaptive and precise control method for boiler combustion in centralized control operation of a thermal power plant, provided by an embodiment of the present invention; Figure 3 This is a structural diagram of an adaptive precision control device for boiler combustion in a thermal power plant's centralized control operation, provided in an embodiment of the present invention. Detailed Implementation

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0020] The following description, with reference to the accompanying drawings, describes an adaptive and precise control method and apparatus for boiler combustion in centralized control operation of a thermal power plant, according to an embodiment of the present invention.

[0021] Example 1 This embodiment provides an adaptive and precise control method for boiler combustion in the centralized control operation of a thermal power plant. For example... Figure 1 As shown, it includes: S1. Key and non-key parameters of the boiler combustion process are acquired in real time through a hierarchical data acquisition network, and the key and non-key parameters are isolated, filtered and outlier removed based on a differentiated sampling strategy.

[0022] Specifically, in some implementations, "acquiring key and non-key parameters of the boiler combustion process in real time through a hierarchical data acquisition network, and performing signal isolation, filtering, and outlier removal based on a differentiated sampling strategy" is the fundamental step in achieving precise control in this invention. This step, through the three-layer architecture of the boiler DCS (Distributed Control System)—the field equipment layer, the control layer, and the monitoring layer—collaboratively completes the data acquisition and preprocessing tasks, ensuring that the data relied upon for subsequent combustion status assessment and optimization calculations possesses high accuracy, high reliability, and real-time performance.

[0023] The field equipment layer is equipped with various high-precision sensors and transmitters for real-time monitoring of key parameters during boiler combustion, including flue gas oxygen content, furnace negative pressure, main steam flow and pressure, and coal feeder speed. These parameters directly affect combustion efficiency and system stability; therefore, their sampling frequency is set to 2-5 seconds to meet the need for rapid response. Non-critical parameters, such as coal elemental analysis, fly ash carbon content, and flue gas temperature, change relatively slowly, so their sampling period is set to 30-60 seconds. This reduces the system's data processing burden and improves overall operating efficiency while ensuring data integrity.

[0024] Furthermore, to improve data quality, this step employs a differentiated signal processing strategy. For key parameters, the system performs signal isolation processing before data upload to eliminate the impact of on-site electromagnetic interference on measurement accuracy; subsequently, a moving average filtering algorithm is used to smooth the data and reduce fluctuations caused by random noise. Simultaneously, outlier removal is performed based on the 3σ criterion; that is, when a data point deviates from the mean of the current time period by more than three times the standard deviation, it is identified as an outlier and removed. The mathematical expression for this is:

[0025] in, This is the current sampled value. The mean within the sliding window. The standard deviation is denoted as . This method effectively filters out abnormal data caused by sensor malfunctions or transient disturbances, improving data reliability.

[0026] This step is widely applicable to real-time monitoring and control of boiler combustion in centralized control systems of thermal power plants. Through hierarchical data acquisition and differentiated processing, the system can continuously acquire high-quality data under complex operating conditions (such as sudden load changes and coal quality fluctuations), providing reliable input for combustion status assessment and multi-objective optimization.

[0027] By constructing a hierarchical data acquisition network and employing differentiated sampling strategies, comprehensive perception and high-quality data acquisition of multi-dimensional parameters in the boiler combustion process were achieved, laying a solid data foundation for subsequent combustion status assessment and optimized control. Simultaneously, signal isolation and filtering effectively suppressed on-site interference, ensuring data stability and continuity, thereby improving the response speed and control accuracy of the entire control system.

[0028] Furthermore, S1 includes: S11, the sampling period for key parameters is 2-5 seconds, and the sampling period for non-key parameters is 30-60 seconds. Outliers are removed using the 3σ criterion, specifically, when the deviation of a data point from the mean exceeds 3 times the standard deviation, it is considered an outlier and removed. Key parameters include flue gas oxygen content, furnace negative pressure, main steam flow rate and pressure, and coal feeder speed. Non-key parameters include coal elemental analysis, fly ash carbon content, and flue gas temperature.

[0029] Specifically, key parameters include flue gas oxygen content, furnace negative pressure, main steam flow rate and pressure, and coal feeder speed. Non-key parameters include coal elemental analysis, fly ash carbon content, and flue gas temperature. In the data preprocessing stage, the 3σ criterion is used to remove outliers from key parameters and to repair missing data, ensuring high reliability and continuity of the data relied upon for subsequent combustion state assessment and optimization calculations. The technical principle of this step is based on the normal distribution characteristic in statistics. Ideally, most data points should be symmetrically distributed around the mean, and 99.7% of data points should fall within the range of the mean ± 3 standard deviations. Therefore, when a data point deviates from the mean of the data within the current time window by more than 3 standard deviations, the data point is identified as an outlier and removed. The mathematical expression for this is:

[0030] in, Indicates the first The value of each sampling point This represents the average of the data within the current time window. The standard deviation is used. In actual operation, the system adopts a sliding window mechanism to dynamically calculate the mean and standard deviation of the current dataset, thereby adapting to the slow drift and non-steady-state characteristics of parameters during boiler combustion. The sampling period for critical parameters is set to 2-5 seconds to ensure a rapid response to the combustion process, while the sampling period for non-critical parameters is 30-60 seconds to balance data acquisition frequency with reasonable allocation of system resources.

[0031] For data repair, the system employs time-series-based interpolation methods, such as linear interpolation or predictive interpolation based on historical data, to fill in missing data caused by sensor failures or communication interruptions. This method ensures data continuity while avoiding misjudgments in control strategies due to data loss.

[0032] This step plays a crucial role in ensuring data quality throughout the entire control method. By removing outliers and repairing missing data, it effectively suppresses measurement errors caused by on-site electromagnetic interference and sensor drift, providing a high-precision, low-noise data foundation for combustion status assessment. Furthermore, this preprocessing mechanism improves the input quality of the multi-objective optimization model, thereby enhancing the robustness and accuracy of the optimization results, which is of great significance for achieving adaptive and precise control of boiler combustion.

[0033] S12 performs moving average filtering on key parameters, with a sliding window length of 5-10 sampling points, to eliminate random measurement noise and smooth the data curve.

[0034] Specifically, applying moving average filtering to key parameters is one of the core operations in the data preprocessing stage of this invention. Its technical implementation is based on the principle of time series signal processing, aiming to eliminate random measurement noise and improve the smoothness of data curves, thereby providing high-quality data input for subsequent combustion state assessment and multi-objective optimization calculation.

[0035] In some implementations, the moving average filter works by setting a length of [value missing] on the time series. A sliding window is used to calculate the arithmetic mean of the data points within the window, replacing the original data value at the current time. Specifically, let's say at a certain time... The original data sequence is The output after moving average filtering It can be represented as:

[0036] in, Regarding the sliding window length, this invention recommends a range of 5 to 10 sampling points. This range effectively balances the conflict between noise suppression and response speed in practical engineering. The selection of the sliding window length needs to be combined with the sampling frequency and the dynamic characteristics of the system. For example, for key parameters with a sampling period of 2-5 seconds (such as flue gas oxygen content, main steam pressure, etc.), 5-10 sampling points correspond to a time window of 10-50 seconds, which is sufficient to smooth short-term fluctuations while avoiding excessive lag that could affect control accuracy.

[0037] Furthermore, the implementation of moving average filtering is typically embedded in the control layer of the boiler DCS system, and is completed collaboratively by the data acquisition module and the signal processing module. In practical applications, this filtering algorithm can adopt a circular buffer structure to update the data points within the window in real time, reducing the consumption of computing resources. In addition, to enhance the filtering effect, it is possible to combine it with improved methods such as weighted moving average (WMA) or exponentially weighted moving average (EWMA), but this invention prioritizes the use of simple moving average to ensure the stability and real-time performance of the algorithm.

[0038] This step has significant technical value in the centralized control operation of thermal power plants. Through filtering, random noise caused by sensor drift, electromagnetic interference, or instantaneous changes in operating conditions can be effectively suppressed, improving the signal-to-noise ratio and data continuity of key parameters. For example, in the flue gas oxygen content signal, moving average filtering can control measurement fluctuations within ±0.2%, thereby improving the accuracy of combustion air distribution control. Simultaneously, this processing method provides a stable and reliable data foundation for subsequent state assessment and multi-objective optimization calculations based on the entropy weight method, which is a key prerequisite for achieving adaptive and precise control of boiler combustion.

[0039] S2, based on the processed combustion process parameters, uses the entropy weight method to evaluate the current combustion state from three dimensions: economy, environmental protection and safety, and divides the combustion state into four levels according to the comprehensive evaluation value to trigger optimization calculation.

[0040] Specifically, this step comprehensively evaluates the current combustion state from three dimensions—economic efficiency, environmental friendliness, and safety—based on the processed combustion process parameters. The entropy weight method is used to objectively determine the weights of each evaluation indicator, thereby generating a representative comprehensive evaluation value. This evaluation value is used to classify the combustion state into four levels: Excellent (85-100 points), Good (70-85 points), Average (60-70 points), and Poor (below 60 points), serving as the basis for determining whether to trigger optimization calculations.

[0041] This step first involves constructing a multi-dimensional evaluation index system. Economic indicators include boiler efficiency (…). ) and smoke exhaust loss ( ), used to measure fuel efficiency and energy loss; environmental indicators cover nitrogen dioxide (NO2) ) and carbon monoxide ( Emission concentration reflects the cleanliness of the combustion process; safety indicators include furnace temperature distribution (…). ) and superheated steam temperature deviation ( This data is used to assess combustion stability and equipment operational risks. All indicators are collected through the DCS system and preprocessed to ensure data quality.

[0042] Entropy weighting is an objective weighting method based on information entropy. Its core idea is to reflect the importance of each indicator in the overall evaluation by calculating its information entropy. In practice, the indicators are first standardized to eliminate dimensional differences; then, the information entropy of each indicator is calculated. And based on this, its weight is obtained. ,in The total number of evaluation indicators. The comprehensive evaluation value is calculated using the entropy weight method. It can be represented as:

[0043] in For the first The standardized value of each indicator, This corresponds to the entropy weight. This formula is used to quantify the overall performance of the current combustion state.

[0044] The scoring threshold for combustion status assessment is set at four levels. The system is also configured to automatically trigger multi-objective optimization calculation in step four if the score is below "Good" for multiple consecutive periods (typically 3-5 sampling periods) or if a single assessment result is below "Poor". This mechanism ensures that the system can respond promptly and adjust parameters when combustion status continuously deteriorates or suddenly worsens.

[0045] This step is applicable to real-time status monitoring and control decisions for boiler combustion in centralized control systems of thermal power plants. Through the integration of the DCS system and PLC control loop, the system can complete a status assessment every 2-5 seconds of sampling and display it visually at the monitoring level, providing operators with auxiliary judgment criteria.

[0046] By combining multi-dimensional indicators with the entropy weight method, an objective and quantitative assessment of the combustion state is achieved, avoiding biases caused by subjective weighting. Simultaneously, the state level classification mechanism provides clear control triggering conditions for the system, improving the stability and controllability of the combustion process and providing a scientific basis for subsequent multi-objective optimization calculations, thereby achieving a dynamic balance between economy, environmental protection, and safety.

[0047] Furthermore, S2 includes: S21, when calculating the weights of each indicator using the entropy weight method, the information entropy formula is used. Calculate information entropy value ,in For the first The first sample Individual indicator values, The total number of samples.

[0048] Specifically, this step involves constructing a multi-dimensional combustion state assessment system to quantitatively analyze the boiler combustion process from three key aspects: economy, environmental protection, and safety. It also employs the entropy weight method to objectively assign weights to various indicators, thereby achieving a scientific classification and dynamic response of the combustion state.

[0049] Entropy weighting is an objective weighting method based on information entropy theory. Its core principle is to calculate the information entropy value of each indicator. This reflects the degree of variation of the value in the sample data, and thus determines its weight in the overall evaluation. Specifically, the information entropy value... The calculation formula is:

[0050] in, Indicates the first The first sample Individual indicator values, The total number of samples is represented by this formula. First, each indicator is normalized by dividing its value by the sum of all samples for that indicator, thus obtaining the relative weight. Then, by calculating information entropy This is used to measure the dispersion of the indicator. The smaller the information entropy, the greater the difference between different samples, and the stronger its ability to distinguish combustion states. Therefore, it should be given a higher weight in the comprehensive evaluation.

[0051] This invention selects several representative combustion performance indicators, including economic indicators (such as boiler efficiency and flue gas heat loss), environmental indicators (such as NO2 emission concentration and CO emission concentration), and safety indicators (such as furnace temperature distribution uniformity and superheated steam temperature deviation). These indicators are collected in real time through a DCS system, with the sampling frequency differentiated according to their sensitivity to combustion conditions. The sampling period for key parameters is 2-5 seconds, and for non-key parameters, it is 30-60 seconds. By calculating the entropy values ​​of these indicators, their objective weights can be obtained. The calculation formula is:

[0052] in, This represents the total number of evaluation indicators. This weighting method avoids the bias of subjective human weighting and enhances the scientific rigor and objectivity of the evaluation results.

[0053] This step is widely used in centralized control systems of thermal power plants, especially in scenarios involving real-time monitoring and early warning of boiler combustion status. By classifying the evaluation results into four levels (Excellent, Good, Average, and Poor), the system can determine whether to initiate multi-objective optimization calculations based on the comprehensive score of the current combustion status. For example, when the comprehensive score is below "Good" for several consecutive periods or below "Poor" in a single instance, the system will automatically trigger optimization strategies to improve combustion efficiency, reduce pollutant emissions, and control the risk of slagging.

[0054] This step objectively assigns weights to multi-source heterogeneous combustion parameters using the entropy weight method, effectively improving the accuracy and robustness of combustion state assessment. Compared to traditional assessment methods based on empirical weights, this invention can dynamically adapt to different coal types, load changes, and operating conditions, achieving real-time diagnosis and early warning of combustion state degradation. This provides reliable data support and decision-making basis for subsequent multi-objective optimization and control. This method improves boiler operating economy while also considering environmental protection and safety, demonstrating significant engineering practical value.

[0055] S22, based on the comprehensive evaluation value Classifying combustion status levels, among which For the first Entropy weighting of each indicator For the first Standardized scores for each indicator This represents the total number of indicators.

[0056] Specifically, in the combustion state assessment step, the system uses an entropy weighting method to objectively assign weights to each indicator based on multi-dimensional combustion performance indicators, and then evaluates the results using a comprehensive evaluation value. The current combustion state is quantitatively assessed and classified. Among other things, Indicates the first Entropy weighting of each indicator The standardized score for this indicator, To assess the total number of indicators involved, its technical implementation principle is based on the information entropy theory. By calculating the information entropy value of each indicator, it reflects the uncertainty of its role in combustion state assessment, thereby determining its weight and avoiding subjective bias caused by human weighting.

[0057] In practice, several key indicators are first selected from three dimensions: economy, environmental protection, and safety. These indicators include boiler efficiency, flue gas loss, NO2 emission concentration, CO emission concentration, furnace temperature distribution, and superheated steam temperature deviation. The raw data for each indicator needs to be standardized, typically using range standardization to map the raw data to the [0,1] interval to eliminate dimensional differences. Subsequently, the entropy weight method is used to calculate the weight of each indicator. The calculation process includes: constructing a standardized scoring matrix and calculating the information entropy of each indicator. Determine the coefficient of difference Finally, the weights are obtained by normalization. .

[0058] Overall evaluation value The calculation results are used to classify the combustion status into four levels: Excellent (85–100 points), Good (70–85 points), Average (60–70 points), and Poor (below 60 points). This assessment mechanism has significant application value in the centralized control system of thermal power plants, especially in operating environments with frequent load fluctuations and large changes in coal quality. It can identify the deterioration trend of combustion status in real time, providing a scientific basis for subsequent multi-objective optimization and control. When the assessment level is below "Good" for several consecutive cycles or the single assessment result is "Poor," the system will automatically trigger the optimization calculation process, thereby realizing closed-loop control and dynamic adjustment of the combustion process.

[0059] S3. Establish a multi-objective optimization function that includes maximizing combustion efficiency, minimizing nitrogen oxide emissions, and minimizing slagging risk. Solve the multi-objective optimization function using a multi-objective optimization algorithm and output the optimal combination of operating parameters.

[0060] Specifically, this invention achieves synergistic optimization of the boiler combustion process among economy, environmental protection, and safety by establishing a multi-objective optimization function. The core of this step lies in constructing a mathematical model that includes three optimization objectives: maximizing combustion efficiency, minimizing nitrogen oxide (NOx) emissions, and minimizing slagging risk. A multi-objective optimization algorithm is then used to solve this model, outputting the optimal combination of operating parameters, thereby providing a scientific and precise basis for the control of the boiler combustion system.

[0061] The form of the multi-objective optimization function is:

[0062] in, This represents the derivative of combustion efficiency, used to measure the changing trend of fuel energy utilization. It indicates the concentration of nitrogen dioxide emissions, reflecting the environmental performance of the combustion process; This represents the slagging risk index, used to assess the likelihood of slagging on the heated surfaces inside the furnace. This is a controllable parameter vector, typically including key operating parameters such as the air-to-coal ratio, secondary air ratio, and burner tilt angle. This function constructs a multi-dimensional optimization space by quantifying multiple conflicting objectives, enabling the system to achieve optimal combustion efficiency while meeting emission and safety constraints.

[0063] combustion efficiency derivative Modeling is typically based on the real-time rate of change of boiler thermal efficiency and fuel consumption rate, with units of % / min or kJ / (kg·min); NOx emission concentration Measured in mg / Nm³, emissions must meet national or industry standards (such as GB 13223-2011 "Emission Standard of Air Pollutants for Thermal Power Plants"); slagging risk index It is calculated by comprehensively considering parameters such as furnace temperature distribution, ash melting point, and coal ash composition. Its value range is usually [0,1]. The smaller the value, the lower the risk of slagging.

[0064] This multi-objective optimization function is suitable for boiler combustion control in thermal power plants under complex operating conditions such as load fluctuations, coal quality changes, and seasonal operation. By collecting real-time data such as flue gas oxygen content, main steam pressure, and coal quality analysis, the system can dynamically adjust the weights of the optimization objectives to adapt to the needs of different operating stages. For example, during high-load operation, the system may focus more on improving combustion efficiency; while during periods with strict environmental requirements, it will prioritize controlling NOx emissions.

[0065] This step uses multi-objective optimization algorithms (such as NSGA-II, MOEA / D, etc.) to solve the above function, which can find the Pareto optimal solution set among multiple conflicting objectives, thereby outputting a set of operating parameters that achieves the best balance between combustion efficiency, emission control, and slagging risk. This method effectively overcomes the limitations of traditional single-objective optimization in taking multiple operational objectives into account, improves the overall performance and operational stability of the boiler combustion system, and has significant economic and environmental benefits.

[0066] Furthermore, S3 includes: S31, the multi-objective optimization function is solved using the NSGA-II algorithm, with the combustion efficiency derivative as the solution. Nitrogen oxide emission concentration Slag Risk Index To optimize the objective, a Pareto optimal solution set is generated.

[0067] Specifically, this invention employs NSGA-II (Non-dominated sorting genetic algorithm II) to solve a multi-objective optimization function, thereby achieving synergistic optimization of the boiler combustion process among economy, environmental protection, and safety. The optimization function is defined as follows:

[0068] in, This represents the derivative of combustion efficiency, used to measure the trend of change in combustion efficiency. It represents the concentration of nitrogen oxides (NOx) emissions, which is a key indicator of pollutant emissions during combustion. This represents the slagging risk index, reflecting the likelihood of slagging on the boiler's heating surfaces. It is a vector of controllable operating parameters, including air-coal ratio, secondary air ratio, burner tilt angle, etc. These parameters directly affect combustion efficiency, pollutant generation and boiler operation safety.

[0069] The NSGA-II algorithm performs collaborative optimization of the three objective functions mentioned above through non-dominated sorting and crowding calculation. First, the algorithm performs Pareto non-dominated sorting on the individuals in the population, dividing the solution set into multiple levels, with the first level representing the non-dominated optimal solution. Then, through a crowding comparison mechanism, diversity is maintained for solutions at the same level, preventing excessive concentration of the solution set in the objective space. NSGA-II employs simulated binary crossover (SBX) and polynomial mutation (PM) operations to ensure good convergence and distribution in the search space. In this invention, the population size is set to 100, the maximum number of iterations is 200, and the crossover probability is... Probability of mutation This is to ensure that a high-quality Pareto front is obtained within a reasonable computation time.

[0070] In practical applications, this step operates within the optimization control module of the power plant's centralized control system, interacting with the DCS system in real time. When the combustion status assessment module detects that the combustion status is below "good" for multiple consecutive cycles or a single assessment is "poor," the system automatically invokes the NSGA-II algorithm for optimization calculation. The optimization results are output in the form of a Pareto optimal solution set, which is then used by the subsequent parameter adjustment and feedforward compensation modules to select the optimal solution or for human-machine interactive decision-making.

[0071] Through a multi-objective optimization algorithm, the nonlinear, multi-constraint, and multi-conflict relationships between combustion efficiency, pollutant emissions, and slagging risk are effectively resolved, achieving adaptive and precise control under complex operating conditions. Compared to traditional single-objective optimization or rule-table control methods, NSGA-II can optimize other objectives without sacrificing any one objective, thereby improving the overall performance and stability of boiler operation.

[0072] S32, Select the optimal combination of operating parameters from the Pareto optimal solution set. The selection of the optimal solution is based on the weighted summation method. ,in The preset weighting coefficients are used for economic efficiency, environmental protection, and safety objectives.

[0073] Specifically, in the multi-objective optimization calculation step, this invention employs a weighted summation method to screen the Pareto optimal solution set for comprehensive optimal solutions. Its core formula is: ,in This represents the derivative of combustion efficiency. Indicates the concentration of nitrogen dioxide emissions. This indicates the boiler slagging risk index. These are the weighting coefficients for the preset economic, environmental, and safety objectives. This method, by introducing a weighting mechanism, transforms the originally conflicting multi-objective problem into a single-objective optimization problem, thereby achieving a comprehensive optimal selection of operating parameters.

[0074] This step first constructs a multi-objective optimization function F(x) = [f_1(x), f_2(x), f_3(x)] based on the combustion state indices obtained from the evaluation, where... This is a vector of controllable parameters, including the air-to-coal ratio, secondary air ratio, and burner tilt angle. A Pareto-optimal solution set is obtained using a non-dominated sorting genetic algorithm (NSGA-II) or other multi-objective optimization algorithms. This solution set represents a compromise solution where further optimization of economy, environmental protection, and safety is impossible under the current operating conditions. Subsequently, the system calculates the optimal solution based on preset weight coefficients. For each solution, a weighted sum is performed to calculate its comprehensive objective function value. Then, the parameter combination corresponding to the minimum value is selected as the final comprehensive optimal operation parameter.

[0075] Weighting coefficient The settings need to be adjusted according to the actual operating strategy of the thermal power plant, and usually meet the following requirements. For example, during the operation phase where environmental protection is emphasized, [measures can be set up]. , , However, in stages where economic efficiency is emphasized, it can be... Increased to above 0.5. Furthermore, The gradient rate of change of boiler efficiency is usually used as the optimization objective. Based on the real-time monitoring value of NO2 concentration in flue gas, The results are calculated using models such as furnace temperature distribution and ash melting point prediction, and are used to quantify the risk of slagging.

[0076] This step is widely used in centralized control systems of thermal power plants, especially under operating conditions with frequent load fluctuations and significant changes in coal quality. By collecting key parameters such as flue gas oxygen content, main steam pressure, and coal feeder speed in real time, combined with coal quality analysis data, the system can dynamically adjust the weights of optimization objectives to adapt to different operating strategies. For example, when the grid's peak-shaving demand is high, the system can prioritize load response speed and combustion efficiency, while under stricter environmental regulations, it can strengthen the control of NO2 emissions.

[0077] S4, the combined operating parameters are sent to the regulating loop of the boiler DCS system, and dynamic feedforward compensation is performed when the load command changes. The dynamic feedforward compensation includes adjusting the air supply and coal feed in advance according to the load change trend to maintain the optimal air-coal ratio.

[0078] Specifically, this invention achieves adaptive and precise control of the combustion process by sending the optimal combination of operating parameters obtained in step four to the regulating loop of the boiler DCS system and executing feedforward compensation control when the load command changes dynamically. This step is the execution terminal of the entire control method, and its technical implementation relies on the DCS system's rapid response capability to control parameters and the scientific design of the feedforward compensation strategy.

[0079] In some implementations, the combination of operating parameters includes key control variables such as the air-to-coal ratio, secondary air ratio, and burner tilt angle. These parameters are obtained by solving a multi-objective optimization function, aiming to achieve an optimal balance between maximizing combustion efficiency, minimizing nitrogen oxide emissions, and minimizing slagging risk. Specifically, when the load command undergoes a step change or continuous fluctuation, the system uses a load prediction model to predict the load change trend at a future moment (e.g., 5-10 seconds later) and adjusts the air supply and coal feed rates in advance in the control loop accordingly. For example, during the load increase phase, the system can proportionally increase the air supply while fine-tuning the coal feed rate based on coal quality characteristics to maintain the optimal air-to-coal ratio. To ensure complete fuel combustion; during periods of reduced load, the air supply volume should be reduced in advance to prevent excessive air coefficient. If the volume is too large, it will reduce the heat loss from flue gas and the amount of nitrogen oxides generated.

[0080] The frequency of parameter adjustment and feedforward compensation is usually related to the load change rate. Relatedly, when the load change rate exceeds a set threshold (e.g.) When the load changes, the system will automatically activate the feedforward compensation mechanism. The compensation amount is calculated based on the dynamic relationship model between historical load changes and combustion parameter responses, typically using linear interpolation or a neural network-based prediction model for real-time estimation. In this step, the DCS system receives the optimized parameters via the OPC interface or hard-wiring and writes them into the setpoint register of the PID controller, achieving closed-loop control of actuators such as dampers and coal feeders.

[0081] In practical applications, this step is typically deployed at the control layer of the DCS system in thermal power plants, directly interacting with the actuators at the field equipment level (such as variable frequency blowers and coal feeders). Its operating environment must meet real-time requirements, with a control cycle generally of 1-3 seconds to ensure rapid response and stable adjustment of combustion parameters. The technical effect of this step is to significantly improve the dynamic response capability and adjustment accuracy of the boiler combustion system, effectively mitigating combustion imbalance caused by load changes, thereby achieving synergistic optimization of combustion efficiency and environmental performance while ensuring unit output.

[0082] Furthermore, S4 includes: S41, in dynamic feedforward compensation, the pre-adjustment ratio of the supply air volume when the load increases is the current load change rate. The coal feed rate is adjusted by a factor of 1:1, which is equal to the current load change rate. times, of which and These are compensation coefficients trained based on historical data.

[0083] Specifically, the technology is based on the dynamic relationship between load change rate and combustion parameters. By pre-adjusting the air supply volume and fine-tuning the coal feed rate, the air-coal ratio is optimized in real time, thereby improving combustion efficiency and reducing pollutant emissions.

[0084] In some implementations, when load commands show an upward trend, the system adjusts the load change rate accordingly. proportional Pre-adjusted air supply volume, i.e. ,in The compensation coefficient, trained based on historical operating data, is used to quantify the sensitivity of the air supply volume to load changes. Pre-adjustment of the air supply volume can effectively avoid the increase of incomplete combustion products such as carbon monoxide due to incomplete fuel combustion, while preventing furnace negative pressure fluctuations caused by air supply lag, thus improving combustion stability.

[0085] Furthermore, the fine-tuning ratio of the coal feed rate is 1 / 3 of the current load change rate. times, that is ,in Similarly, the parameters, obtained through historical data modeling and training with machine learning algorithms (such as support vector regression and random forest), are used to characterize the strength of the fuel supply's response to load changes. This fine-tuning mechanism can make appropriate adjustments to the fuel supply in the early stages of load increases, avoiding the risk of slagging and excessive air coefficient caused by excessive coal supply, thereby achieving a balance between economy and safety.

[0086] This compensation mechanism is typically integrated into the regulating loop of the boiler's DCS (Distributed Control System) and works in conjunction with the load forecasting model. When the load forecasting model detects that the load will increase at a certain future time, the system anticipates this increase based on the current load change rate. and compensation coefficient , Feedforward adjustments to airflow and coal flow are made to shorten control response time and improve system dynamic performance. Compensation coefficient. and The value range is usually between 0.5 and 1.5. The specific value needs to be optimized and adjusted online or offline based on the boiler type, coal characteristics and historical operating data.

[0087] By introducing a dynamic feedforward compensation mechanism, the response speed and adjustment accuracy of the boiler combustion system to load changes are significantly improved, effectively alleviating the combustion instability problem caused by the lag effect in traditional PID control. Simultaneously, through active adjustment of the air-coal ratio, synergistic optimization of combustion efficiency and pollutant emissions is achieved, providing reliable technical support for centralized control operation of thermal power plants under complex operating conditions.

[0088] S42, the pre-adjustment ratio of the supply air volume when the load decreases is the current load change rate. times, and This is to prevent excessive air coefficient from causing decreased efficiency and increased nitrogen oxide generation.

[0089] Specifically, when the load decreases, the pre-adjustment ratio of the air supply volume is set to the current load change rate. times, and This is one of the key control strategies in the parameter adjustment and dynamic feedforward compensation steps of this invention. The core technical principle of this step lies in responding to load change trends in advance through a dynamic feedforward mechanism, thereby avoiding excessive air coefficient caused by lag in air supply volume adjustment. Excessive heat can lead to decreased combustion efficiency and increased formation of nitrogen oxides (NOx).

[0090] The system first obtains the current load change rate through the control layer of the boiler DCS system. This rate of change reflects the magnitude of the decrease in unit load per unit time. Subsequently, the system adjusts according to the preset adjustment coefficient. Calculate the pre-adjustment increment of the air supply volume This increment is then used in the feedforward control of the air supply regulation loop. Because... The value is less than the adjustment coefficient used when the load increases. This strategy adopts a more conservative air supply adjustment strategy during the load reduction phase to prevent incomplete combustion and increased heat loss caused by excess air.

[0091] Adjustment coefficient The settings need to be calibrated by combining the boiler's combustion characteristics, aerodynamic model, and historical operating data, usually in... The selected value should be within a certain range, and the specific value needs to be determined through simulation or actual operation testing. Simultaneously, the air supply volume adjustment must meet the dynamic response requirements of the boiler combustion system to air volume, and is generally controlled within a certain range. Adjustments are completed within seconds to ensure a balance between system stability and response speed.

[0092] This procedure applies to thermal power plants operating under conditions of rapid load decline, such as sudden changes in grid dispatch instructions or unit peak shaving. Under these conditions, failure to reduce the air supply in a timely manner will result in an excess air coefficient. Excessive air volume not only reduces boiler thermal efficiency but also promotes NOx formation due to the high-temperature, oxygen-rich environment, increasing environmental pressure. Through feedforward control in this step, the system can pre-adjust the air supply volume at the initial stage of load reduction, thereby maintaining the economy and environmental friendliness of the combustion process.

[0093] This step effectively improves the dynamic response and control precision of the boiler combustion system, reducing the risk of combustion instability caused by delayed air supply adjustments. Through reasonable settings... The system can precisely control the air supply during load reduction, thereby maintaining the optimal air-to-coal ratio, reducing the introduction of ineffective air, improving combustion efficiency and suppressing NOx generation, providing a strong guarantee for the efficient and clean operation of thermal power plants.

[0094] Also includes: S5. The constraints of the multi-objective optimization function are dynamically adjusted based on the real-time monitored coal element analysis data. Specifically, when the changes in the coal element analysis data exceed the preset threshold, the entropy weight method weights are recalculated and the parameter boundaries of the multi-objective optimization function are updated.

[0095] Specifically, in some implementations, the "multi-objective optimization calculation" of this invention achieves adaptive and precise control of the boiler combustion process by dynamically adjusting the constraints of the multi-objective optimization function. Specifically, when real-time monitored coal elemental analysis data (such as volatile matter, fixed carbon, ash, and moisture) undergo significant changes, and the magnitude of these changes exceeds preset thresholds (e.g., volatile matter changes exceeding ±5%, ash changes exceeding ±3%, and moisture changes exceeding ±2%), the system will automatically trigger a recalculation of the entropy weight method weights and update the parameter boundaries of each optimization objective in the multi-objective optimization function accordingly, ensuring that the optimization results are highly matched with the current combustion conditions.

[0096] This step, based on coal quality data collected by the boiler's DCS system and combined with combustion status assessment results, uses the entropy weight method to dynamically assign weights to economic, environmental, and safety indicators. The entropy weight method reflects the importance of each indicator in the system by calculating its information entropy; the weight calculation formula is as follows:

[0097] in, For the first The weight of each indicator, This represents the information entropy of the indicator. Using this method, the system can objectively reflect the weighting of the impact of coal quality changes on combustion efficiency, pollutant emissions, and slagging risk, thereby dynamically adjusting the priority of each objective in the optimization function.

[0098] The multi-objective optimization function is defined as:

[0099] in, This represents the derivative of combustion efficiency. Indicates the concentration of nitrogen dioxide emissions. Indicates the risk index of slagging. This is a controllable parameter vector, including the air-to-coal ratio, secondary air ratio, burner tilt angle, etc. When the coal quality parameters change beyond the set threshold, the system will recalculate the entropy weight method weights and adjust the parameter boundaries of each optimization objective according to the new weights, such as adjusting the air volume adjustment range and the coal-to-water ratio control range, to adapt to the combustion characteristic deviation caused by changes in coal quality characteristics.

[0100] This step is applicable to thermal power plants operating in scenarios involving frequent coal type switching, large load fluctuations, or complex and variable combustion conditions. For example, when the calorific value of coal decreases or the ash content increases, the system can automatically identify the trend of coal quality deterioration and guide the combustion parameters towards a safer and more environmentally friendly direction by adjusting the constraints of the optimization function, thereby avoiding combustion instability or excessive emissions caused by coal quality mismatch.

[0101] This step effectively enhances the adaptability and robustness of the boiler combustion control system. By dynamically updating the constraints of the optimization function, the system can respond quickly to changes in coal quality, avoiding control deviations caused by fixed weights or static boundaries. This achieves a better balance between economy, environmental protection, and safety, improving combustion efficiency, reducing pollutant emissions, and minimizing the risk of boiler slagging. It also significantly enhances the intelligence level and control precision of thermal power plant operation.

[0102] The adaptive precision control method for boiler combustion in thermal power plants according to this invention dynamically adjusts the constraints of the multi-objective optimization function based on real-time coal quality element analysis data. When the coal quality change exceeds a preset threshold, the entropy weight method weights and parameter boundaries are automatically updated, further improving the adaptability of the combustion control system to coal quality fluctuations. This enables more precise combustion parameter optimization under complex operating conditions and significantly enhances the synergistic optimization level of combustion efficiency and emission control.

[0103] Example 2 This invention proposes yet another adaptive and precise control method for boiler combustion in the centralized control operation of thermal power plants, such as... Figure 2 As shown, it includes: Step 1: Real-time data acquisition and signal processing, acquiring combustion process parameters in real time; Step 2: Data preprocessing, including removing outliers from key parameters and repairing missing data; Step 3: Combustion status assessment; assess the current combustion status. Step 4: Multi-objective optimization calculation. With the optimization objectives of maximizing combustion efficiency, minimizing nitrogen dioxide emissions, and minimizing boiler slagging risk, a multi-objective optimization function is established. Step 5: Parameter adjustment and dynamic feedforward compensation. The optimal operating parameter setpoints are sent to the control loop, and feedforward compensation is performed when the load command changes.

[0104] This invention achieves data acquisition by constructing a hierarchical data acquisition network and a differentiated sampling strategy; it also outputs the optimal combination of operating parameters by collaboratively solving the conflicting objectives of economy, environmental protection and safety through a multi-objective optimization algorithm.

[0105] The adaptive precision control method for boiler combustion in the centralized control operation of this thermal power plant utilizes the collaborative work of the field equipment layer, control layer, and monitoring layer of the boiler DCS system. It collects and preprocesses combustion process parameters in real time, such as key data like flue gas oxygen content and air volume, as well as non-critical data like coal quality analysis. Signal isolation, filtering, and outlier removal based on the 3σ criterion are performed to obtain high-quality data. Next, the processed data is used to evaluate the combustion state from three dimensions: economy, environmental protection, and safety. The entropy weight method is used to determine the weights of indicators to classify the state levels. Then, a multi-objective optimization function is established with the goals of maximizing combustion efficiency, minimizing nitrogen oxide emissions, and minimizing slagging risk. The optimal combination of operating parameters is obtained through algorithmic solutions. Finally, these parameter setpoints are sent to the DCS control loop, and dynamic feedforward compensation is executed when the load changes, thereby achieving precise, stable, and efficient optimized control of boiler combustion.

[0106] Furthermore, the present invention will be further described in detail below with reference to embodiments. These embodiments are only used to explain the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0107] The boiler, operating under centralized control in a thermal power plant, is one of the core pieces of equipment in a modern thermal power plant. It is a large, complex, and technologically advanced pressure vessel whose core task is to efficiently and safely convert water into superheated steam with specific pressure and temperature to drive the turbine generator unit to generate electricity. Under centralized control, the boiler is no longer an isolated device but an organic whole that works closely with the turbine, generator, and other main and auxiliary equipment through a distributed control system, enabling centralized monitoring, operation, and optimized control. The boiler system mainly consists of two parts: the boiler body and auxiliary equipment. The boiler body comprises two major systems: the "boiler" and the "furnace." The "boiler" system, or steam-water system, is responsible for heating and flowing the working fluid. It mainly includes the steam drum, downcomer, water-cooled walls, superheater, reheater, and economizer. The process involves feedwater being preheated by the economizer before entering the steam drum, then distributed through the downcomer to the water-cooled walls to absorb radiant heat from the furnace to generate steam. The steam is finally heated to its rated parameters in the superheater and reheater. The "furnace" system, or combustion system, is responsible for the efficient and clean combustion of fuel. It mainly includes the furnace, burner, air preheater, and flue. The process involves mixing pulverized coal with air, injecting it into the furnace through the burner for complete combustion, and then transferring heat to the heating surfaces. After purification treatment such as dust removal and desulfurization, the resulting high-temperature flue gas is discharged into the atmosphere by an induced draft fan. Auxiliary equipment for the boiler includes a pulverizing system (such as a coal mill), ventilation equipment (forced and induced draft fans), feedwater equipment (feed pumps), dust removal and ash removal equipment, and a desulfurization system. These provide necessary support for the stable operation of the boiler itself.

[0108] Based on this, according to Figure 2 As shown in this embodiment, the adaptive and precise control method for boiler combustion in the centralized control operation of a thermal power plant includes the following steps: Step 1: Real-time Data Acquisition and Signal Processing. Combustion process parameters, including fuel characteristic parameters, airflow parameters, temperature parameters, and emission parameters, are acquired in real time through the boiler's DCS system. Before data upload, signal isolation and filtering are performed to suppress electromagnetic interference. The boiler DCS system comprises a field equipment layer, a control layer, and a monitoring layer. The field equipment layer monitors flue gas composition in real time and includes sensors, transmitters, electric regulating valves, and frequency converters. The control layer receives data uploaded from the field equipment layer, performs calculations according to pre-set control strategies, and issues control commands to the actuators. The monitoring layer monitors the boiler's operating status in real time. Combustion process parameter acquisition includes critical and non-critical data. Critical data includes flue gas oxygen content, furnace negative pressure, main steam flow and pressure, and coal feeder speed signals, with a sampling period of 2-5 seconds. Non-critical data includes coal elemental analysis, fly ash carbon content, and flue gas temperature, with a sampling period of 30-60 seconds. The boiler DCS system employs a three-layer architecture—field equipment layer, control layer, and monitoring layer—working collaboratively. First, various sensors and transmitters at the field equipment layer collect real-time data on key parameters such as flue gas oxygen content and furnace negative pressure, as well as non-critical parameters like coal elemental analysis. As the foundational support for the entire control system, a hierarchical data acquisition network is constructed, and a differentiated sampling strategy is implemented. This enables comprehensive perception and high-quality acquisition of multi-dimensional parameters of the boiler combustion process, providing accurate, real-time, and complete raw data sources for subsequent analysis and decision-making. Simultaneously, signal preprocessing effectively suppresses field interference, ensuring the reliability of the system's input information.

[0109] Step 2: Data preprocessing. For key parameters, outliers are removed using the 3σ criterion; for missing data, time-series-based interpolation is used for data repair. The 3σ criterion considers data points deviating from the mean by more than three times the standard deviation as outliers and removes them. Moving average filtering is used to eliminate random measurement noise, making the data curve smoother. This process plays a crucial role in safeguarding data quality, transforming raw data into clean, continuous, and reliable time-series data through outlier removal, noise filtering, and missing value repair. Its purpose is to eliminate measurement errors and transmission interference, laying a high-quality data foundation for combustion state assessment and optimization calculations.

[0110] Step 3: Combustion Status Assessment. The current combustion status is assessed from three dimensions: economic indicators, environmental indicators, and safety indicators. The entropy weight method is used to determine the weight of each indicator. Based on the assessment results, the combustion status is classified into four levels: Excellent (85-100 points), Good (70-85 points), Average (60-70 points), and Poor (below 60 points). When the assessment level is below "Good" for several consecutive periods or drops to "Poor" in a single instance, the system automatically triggers the optimization calculation in Step 4. Economic indicators include boiler efficiency and flue gas losses; environmental indicators include nitrogen dioxide and carbon monoxide emission concentrations; and safety indicators include furnace temperature distribution and superheated steam temperature deviation. This achieves quantitative diagnosis and scientific evaluation of combustion operation status. Through a multi-dimensional indicator system and objective weight allocation, complex operating characteristics are transformed into intuitive status level scores. Its core function is to accurately identify the trends of good and bad combustion processes, providing a quantitative decision-making basis for whether to initiate optimization control.

[0111] Step four involves multi-objective optimization calculations. With the optimization objectives of maximizing combustion efficiency, minimizing nitrogen dioxide emissions, and minimizing boiler slagging risk, a multi-objective optimization function is established to find the optimal balance point among multiple conflicting objectives. By establishing a multi-objective optimization model and employing advanced algorithms for solution, the inherent contradictions between economy, environmental protection, and safety are effectively resolved. The optimal combination of operating parameters is output, providing scientifically precise control commands for boiler combustion and achieving multi-objective collaborative optimization.

[0112] The multi-objective optimization function is:

[0113] in, f 1 ( x () represents the derivative of combustion efficiency. f 2 ( x ( ) represents the nitrogen dioxide emission concentration. f 3 ( x () represents the slagging risk index. x It is a vector of controllable parameters.

[0114] Step 5: Parameter Adjustment and Dynamic Feedforward Compensation. The optimal operating parameter setpoints obtained in Step 4 are sent to the boiler DCS system control loop, and feedforward compensation is executed when the load command changes. Operating parameters include the air-coal ratio, secondary air ratio, and burner tilt angle. Feedforward compensation pre-increases the air supply by a certain proportion when the load increases and fine-tunes the coal feed rate to maintain the optimal air-coal ratio. When the load decreases, it pre-decreases the air supply to prevent excessive air coefficient from causing efficiency loss and increased nitrogen dioxide generation. By accurately issuing optimized parameters and combining them with load forecasting for feedforward compensation, precise execution and dynamic adjustment of control commands are achieved. This significantly improves system response speed and anti-interference capability, ensuring that the boiler maintains a stable and efficient operating state under load fluctuation conditions.

[0115] The adaptive precision control method for boiler combustion in the centralized control operation of this thermal power plant utilizes the collaborative work of the field equipment layer, control layer, and monitoring layer of the boiler DCS system. It collects and preprocesses combustion process parameters in real time, such as key data like flue gas oxygen content and air volume, as well as non-critical data like coal quality analysis. Signal isolation, filtering, and outlier removal based on the 3σ criterion are performed to obtain high-quality data. Next, the processed data is used to evaluate the combustion state from three dimensions: economy, environmental protection, and safety. The entropy weight method is used to determine the weights of indicators to classify the state levels. Then, a multi-objective optimization function is established with the goals of maximizing combustion efficiency, minimizing nitrogen oxide emissions, and minimizing slagging risk. The optimal combination of operating parameters is obtained through algorithmic solutions. Finally, these parameter setpoints are sent to the DCS control loop, and dynamic feedforward compensation is executed when the load changes, thereby achieving precise, stable, and efficient optimized control of boiler combustion.

[0116] Example 3 To implement the methods of the above embodiments, the present invention also provides an adaptive precision control device for boiler combustion in centralized control operation of a thermal power plant, such as... Figure 3 As shown, it includes: The data acquisition and preprocessing module 100 is used to acquire key and non-key parameters of the boiler combustion process in real time through a hierarchical data acquisition network, and to perform signal isolation, filtering and outlier removal based on a differentiated sampling strategy. The multi-dimensional combustion state assessment module 200 is used to assess the current combustion state from three dimensions—economic efficiency, environmental protection, and safety—based on the processed combustion process parameters and using the entropy weight method. The combustion state is divided into four levels according to the comprehensive evaluation value to trigger optimization calculations. The multi-objective optimization parameter generation module 300 is used to establish a multi-objective optimization function that includes maximizing combustion efficiency, minimizing nitrogen oxide emissions, and minimizing slagging risk. The multi-objective optimization function is solved by a multi-objective optimization algorithm, and the optimal combination of operating parameters is output. The dynamic feedforward compensation execution module 400 is used to send the combination of operating parameters to the regulating loop of the boiler DCS system, and to perform dynamic feedforward compensation when the load command changes. The dynamic feedforward compensation includes pre-adjusting the air supply and coal feed according to the load change trend to maintain the optimal air-coal ratio.

[0117] Furthermore, the data acquisition and preprocessing module is also used for: A sampling period of 2-5 seconds is set for key parameters, and a sampling period of 30-60 seconds is set for non-key parameters. Outliers are removed using the 3σ criterion, specifically, when the deviation of a data point from the mean exceeds 3 times the standard deviation, it is considered an outlier and removed. The key parameters include flue gas oxygen content, furnace negative pressure, main steam flow rate and pressure, and coal feeder speed. The non-key parameters include coal elemental analysis, fly ash carbon content, and flue gas temperature. The key parameters are processed by moving average filtering with a sliding window length of 5-10 sampling points to eliminate random measurement noise and smooth the data curve.

[0118] Furthermore, the multi-dimensional combustion state assessment module is also used for: Using the information entropy formula Calculate information entropy value ,in For the first The first sample Individual indicator values, The total number of samples; Based on the comprehensive evaluation value Classifying combustion status levels, among which For the first Entropy weighting of each indicator For the first Standardized scores for each indicator This represents the total number of indicators.

[0119] Furthermore, the multi-objective optimization parameter generation module is also used for: The multi-objective optimization function is solved using the NSGA-II algorithm, with the combustion efficiency derivative as the criterion. Nitrogen oxide emission concentration Slag Risk Index To optimize the objective, a Pareto optimal solution set is generated; The optimal combination of operating parameters is selected from the Pareto optimal solution set. The selection of the optimal solution is based on the weighted summation method. ,in The preset weighting coefficients are used for economic efficiency, environmental protection, and safety objectives.

[0120] The adaptive precision control device for boiler combustion in the centralized control operation of thermal power plants according to this invention acquires data by constructing a hierarchical data acquisition network and a differentiated sampling strategy; it outputs the optimal combination of operating parameters by collaboratively solving the conflicting objectives between economy, environmental protection and safety through a multi-objective optimization algorithm.

[0121] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0122] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0123] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. An adaptive and precise control method for boiler combustion in centralized control operation of a thermal power plant, characterized in that, include: S1, key and non-key parameters of the boiler combustion process are acquired in real time through a hierarchical data acquisition network, and the key and non-key parameters are isolated, filtered and outlier removed based on a differentiated sampling strategy. S2, based on the processed combustion process parameters, uses the entropy weight method to evaluate the current combustion state from three dimensions: economy, environmental protection and safety, and divides the combustion state into four levels according to the comprehensive evaluation value to trigger optimization calculation; S3. Establish a multi-objective optimization function that includes maximizing combustion efficiency, minimizing nitrogen oxide emissions, and minimizing slagging risk. Solve the multi-objective optimization function using a multi-objective optimization algorithm and output the optimal combination of operating parameters. S4, the combined operating parameters are sent to the regulating loop of the boiler DCS system, and dynamic feedforward compensation is performed when the load command changes. The dynamic feedforward compensation includes adjusting the air supply and coal feed in advance according to the load change trend to maintain the optimal air-coal ratio.

2. The method as described in claim 1, characterized in that, S1 further includes: S11, the sampling period for key parameters is 2-5 seconds, and the sampling period for non-key parameters is 30-60 seconds. Outliers are removed using the 3σ criterion, specifically, when the deviation of a data point from the mean exceeds 3 times the standard deviation, it is considered an outlier and removed. The key parameters include flue gas oxygen content, furnace negative pressure, main steam flow rate and pressure, and coal feeder speed. The non-key parameters include coal elemental analysis, fly ash carbon content, and flue gas temperature. S12 performs moving average filtering on key parameters, with a sliding window length of 5-10 sampling points, to eliminate random measurement noise and smooth the data curve.

3. The method as described in claim 1, characterized in that, S2 further includes: S21, when calculating the weights of each indicator using the entropy weight method, the information entropy formula is used. Calculate information entropy value ,in For the first The first sample Individual indicator values, The total number of samples; S22, based on the comprehensive evaluation value Classifying combustion status levels, among which For the first Entropy weighting of each indicator For the first Standardized scores for each indicator This represents the total number of indicators.

4. The method as described in claim 1, characterized in that, S3 further includes: S31, the multi-objective optimization function is solved using the NSGA-II algorithm, with the combustion efficiency derivative as the solution. Nitrogen oxide emission concentration Slag Risk Index To optimize the objective, a Pareto optimal solution set is generated; S32, Select the optimal combination of operating parameters from the Pareto optimal solution set. The selection of the optimal solution is based on the weighted summation method. ,in The preset weighting coefficients are used for economic efficiency, environmental protection, and safety objectives.

5. The method as described in claim 1, characterized in that, S4 further includes: S41, in the dynamic feedforward compensation, the pre-adjustment ratio of the air supply volume when the load increases is the current load change rate. The coal feed rate is adjusted by a factor of 1:1, which is equal to the current load change rate. times, of which and These are compensation coefficients trained based on historical data; S42, the pre-adjustment ratio of the supply air volume when the load decreases is the current load change rate. times, and This is to prevent excessive air coefficient from causing decreased efficiency and increased nitrogen oxide generation.

6. The method as described in claim 1, characterized in that, Also includes: S5. The constraints of the multi-objective optimization function are dynamically adjusted based on the real-time monitored coal element analysis data. Specifically, when the changes in the coal element analysis data exceed the preset threshold, the entropy weight method weights are recalculated and the parameter boundaries of the multi-objective optimization function are updated.

7. An adaptive precision control device for boiler combustion in a thermal power plant's centralized control operation, characterized in that, include: The data acquisition and preprocessing module is used to acquire key and non-key parameters of the boiler combustion process in real time through a hierarchical data acquisition network, and to perform signal isolation, filtering and outlier removal based on a differentiated sampling strategy. The multi-dimensional combustion state assessment module is used to assess the current combustion state from three dimensions: economy, environmental protection and safety, based on the processed combustion process parameters and using the entropy weight method. The combustion state is divided into four levels according to the comprehensive evaluation value to trigger optimization calculations. The multi-objective optimization parameter generation module is used to establish a multi-objective optimization function that includes maximizing combustion efficiency, minimizing nitrogen oxide emissions, and minimizing slagging risk. The multi-objective optimization function is solved by a multi-objective optimization algorithm, and the optimal combination of operating parameters is output. The dynamic feedforward compensation execution module is used to send the combination of operating parameters to the regulating loop of the boiler DCS system, and to perform dynamic feedforward compensation when the load command changes. The dynamic feedforward compensation includes pre-adjusting the air supply and coal feed according to the load change trend to maintain the optimal air-coal ratio.

8. The apparatus as claimed in claim 7, characterized in that, The data acquisition and preprocessing module is also used for: A sampling period of 2-5 seconds is set for key parameters, and a sampling period of 30-60 seconds is set for non-key parameters. Outliers are removed using the 3σ criterion, specifically, when the deviation of a data point from the mean exceeds 3 times the standard deviation, it is considered an outlier and removed. The key parameters include flue gas oxygen content, furnace negative pressure, main steam flow rate and pressure, and coal feeder speed. The non-key parameters include coal elemental analysis, fly ash carbon content, and flue gas temperature. The key parameters are processed by moving average filtering with a sliding window length of 5-10 sampling points to eliminate random measurement noise and smooth the data curve.

9. The apparatus as claimed in claim 7, characterized in that, The multi-dimensional combustion state assessment module is also used for: Using the information entropy formula Calculate information entropy value ,in For the first The first sample Individual indicator values, The total number of samples; Based on the comprehensive evaluation value Classifying combustion status levels, among which For the first Entropy weighting of each indicator For the first Standardized scores for each indicator This represents the total number of indicators.

10. The apparatus as claimed in claim 7, characterized in that, The multi-objective optimization parameter generation module is also used for: The multi-objective optimization function is solved using the NSGA-II algorithm, with the combustion efficiency derivative as the criterion. Nitrogen oxide emission concentration Slag Risk Index To optimize the objective, a Pareto optimal solution set is generated; The optimal combination of operating parameters is selected from the Pareto optimal solution set. The selection of the optimal solution is based on the weighted summation method. ,in The preset weighting coefficients are used for economic efficiency, environmental protection, and safety objectives.

Images