Intelligent monitoring and control method for air supply of low-nitrogen combustion technology of biomass boiler

Abstract

The present application belongs to the technical field of combustion control, and particularly relates to a kind of intelligent monitoring regulation and control method of biomass boiler low-nitrogen combustion technology, comprising: collecting key operation data of furnace to construct time series data set;Current combustion quality evaluation value is calculated using time weighted sliding monitoring window;Based on historical data clustering, historical combustion state cluster is constructed, and current combustion quality benchmark value is obtained by matching;The combustion quality evaluation value is compared with the combustion quality benchmark value, the oxygen concentration time weighted value of target reference sample point is used to determine the air volume adjustment direction, and the air volume adjustment step factor is calculated;The proportional parameter of dynamic correction proportional integral differential controller is realized adaptive adjustment of air supply amount.The present application can continuously adjust the air supply amount with the change of combustion state, effectively overcome the regulation lag, improve the combustion stability, reduce the emission of nitrogen oxides and improve the boiler operation efficiency.

Description

Technical Field

[0001] This invention relates to the field of combustion control technology. More specifically, this invention relates to an intelligent monitoring and control method for air supply in low-NOx combustion technology for biomass boilers. Background Technology

[0002] Against the backdrop of advancing dual-carbon goals and stringent emission standards for industrial boilers, biomass boilers play a vital role in industrial heating and power generation. Their low-NOx combustion process is crucial for reducing pollutant emissions and improving energy efficiency. Biomass fuels have inherent characteristics such as complex composition, large moisture fluctuations, and high volatile matter content, causing their combustion state to fluctuate continuously with changes in fuel supply, boiler load, and furnace air distribution. This places extremely high demands on the precise control of the combustion process.

[0003] Currently, for air supply control in biomass boilers, existing technologies mostly employ fixed air volume ratios or hysteresis feedback regulation based on a single parameter. These methods typically rely on traditional proportional-integral-derivative (PID) controllers, adjusting damper openings or fan frequencies based on preset fixed thresholds or simple deviation feedback, attempting to maintain a balance in the air supply within the furnace, thereby ensuring the normal and stable operation of the boiler.

[0004] However, due to the highly dynamic and time-varying nature of biomass boiler combustion, the aforementioned traditional fixed air volume ratio or single-parameter feedback regulation methods are unable to respond promptly and accurately to the complex and ever-changing combustion states within the furnace. When combustion conditions fluctuate drastically, this regulation method is prone to causing air supply regulation lag, leading to insufficient air supply resulting in incomplete combustion, or excessive air supply creating localized high-temperature, oxygen-rich areas. Ultimately, this results in decreased boiler operational stability and a significant increase in nitrogen oxide emissions, failing to meet current stringent environmental protection and high-efficiency operation requirements. Summary of the Invention

[0005] To address the technical problems mentioned above, such as the inability of existing air supply control to respond promptly to the dynamic fluctuations in the combustion state of biomass boilers, leading to delayed air supply regulation, incomplete combustion or localized oxygen enrichment, resulting in unstable operation and increased nitrogen oxide emissions, this invention provides an intelligent monitoring and control method for air supply in low-NOx combustion technology for biomass boilers, comprising:

[0006] Real-time data collection of furnace temperature, furnace pressure, oxygen concentration, and nitrogen oxide concentration from biomass boilers is used to construct a time-series dataset of boiler combustion status.

[0007] Set up a sliding monitoring window, and calculate the time-weighted values ​​of temperature data fluctuation, pressure data fluctuation, and nitrogen oxide concentration based on the time weight of each data point in the sliding monitoring window, so as to obtain the combustion quality evaluation value in the current sliding monitoring window.

[0008] The current feature vector is constructed based on the time-weighted average of furnace temperature and pressure data within the sliding monitoring window; the current feature vector is matched with historical combustion state clusters pre-divided by density clustering algorithm to obtain a reference cluster corresponding to the current combustion stage, and the combustion quality benchmark value is determined based on the distribution concentration of historical data within the reference cluster.

[0009] The combustion quality evaluation value in the current sliding monitoring window is compared with the combustion quality benchmark value. In response to the combustion quality evaluation value being lower than the combustion quality benchmark value, the airflow adjustment direction is determined based on the oxygen concentration level of high-quality historical samples in the reference cluster, and the airflow adjustment step size factor is determined based on the degree of difference between the combustion quality evaluation value and the combustion quality benchmark value.

[0010] The proportional parameters of the proportional-integral-derivative control system are dynamically corrected by using the air volume adjustment step size factor. Based on the corrected proportional parameters and the air volume adjustment direction, the air intake of the biomass boiler is dynamically adjusted.

[0011] Preferably, the time weights are distributed with exponential decay based on the time distance between the data points and the latest current time.

[0012] The fluctuations in the temperature and pressure data are calculated by weighting and summing the average absolute deviations of the corresponding data according to time weights.

[0013] Preferably, the combustion quality evaluation value satisfies the expression:

[0014] ;

[0015] In the formula, This indicates the combustion quality evaluation value within the current sliding monitoring window. Indicates the index of the current sliding monitoring window; Represents the maximum value function; and These represent the fluctuation levels of temperature data and pressure data within the current sliding monitoring window, respectively. Indicates the number of times the current sliding monitoring window is filled. The time weights corresponding to each data point; This indicates the number of data points within the current sliding monitoring window; Indicates the number of times the current sliding monitoring window is filled. The normalized nitrogen oxide concentration data corresponding to each data point; This represents the weighting factor.

[0016] Preferably, obtaining the reference cluster corresponding to the current combustion stage includes:

[0017] A density-based clustering algorithm is used to adaptively cluster the feature vectors of historical combustion stages, forming multiple historical combustion state clusters corresponding to different combustion stages. The Euclidean distance between the current feature vector and the cluster center feature vector of each historical combustion state cluster is calculated, and the historical combustion state cluster with the smallest Euclidean distance is taken as the reference cluster corresponding to the current sliding monitoring window. The cluster center feature vector is the mean vector of all historical sliding monitoring window feature vectors within the corresponding historical combustion state cluster.

[0018] Preferably, the distribution concentration satisfies the expression:

[0019] ;

[0020] In the formula, This indicates the reference cluster of the current sliding monitoring window. The distribution concentration of combustion quality evaluation values ​​corresponding to each historical sliding monitoring window; This indicates the total number of historical sliding monitoring windows contained in the reference cluster, and ; and This indicates the reference cluster of the current sliding monitoring window. The and the first Combustion quality evaluation values ​​corresponding to each historical sliding monitoring window; This represents the normalization function.

[0021] Preferably, the combustion quality benchmark value satisfies the expression:

[0022] ;

[0023] In the formula, This represents the combustion quality baseline value of the reference cluster in the current sliding monitoring window; This indicates the reference cluster of the current sliding monitoring window. The distribution concentration of combustion quality evaluation values ​​corresponding to each historical sliding monitoring window; This indicates the total number of historical sliding monitoring windows contained in the reference cluster; This indicates the reference cluster of the current sliding monitoring window. Combustion quality evaluation values ​​corresponding to each historical sliding monitoring window.

[0024] Preferably, determining the airflow adjustment direction based on the oxygen concentration level of high-quality historical samples within the reference cluster includes:

[0025] Obtain historical sliding monitoring windows whose combustion quality evaluation values ​​within the reference cluster of the current sliding monitoring window are higher than the current sliding monitoring window's combustion quality evaluation value, and use them as target reference sample points; calculate the time-weighted value of oxygen concentration in the current sliding monitoring window and the time-weighted value of oxygen concentration in each target reference sample point; count the percentage of target reference sample points whose time-weighted value of oxygen concentration is higher than the current sliding monitoring window, and based on the comparison result of the percentage and a preset constant, determine whether the air volume adjustment direction is to increase the air intake or decrease the air intake.

[0026] Preferably, the air volume adjustment step size factor satisfies the expression:

[0027] ;

[0028] In the formula, Indicates the first The airflow adjustment step size factor corresponding to each sliding monitoring window; Represents the natural constant; This represents the combustion quality baseline value of the reference cluster in the current sliding monitoring window; This indicates the combustion quality evaluation value within the current sliding monitoring window. This indicates the index of the current sliding monitoring window.

[0029] Preferably, the corrected proportional parameter satisfies the expression:

[0030] ;

[0031] In the formula, This indicates the corrected PID proportional parameter; This represents the basic proportional parameter of the PID controller; Indicates the first The airflow adjustment step size factor corresponding to each sliding monitoring window.

[0032] Preferably, the dynamic adjustment of the air intake of the biomass boiler based on the corrected proportional parameters and the air volume adjustment direction includes:

[0033] The corrected proportional parameters are input into the proportional-integral-derivative control system, and the controller continuously adjusts the fan frequency or damper opening based on the deviation between the target air supply state and the current actual air supply state. After completing the current air supply adjustment, the system continues to monitor the operating data in the subsequent sliding monitoring window in real time, and iteratively executes combustion state analysis, reference cluster matching, and air supply optimization adjustment steps to form a closed-loop dynamic control mechanism.

[0034] The beneficial effects of this invention are as follows: By real-time acquisition and time-weighted analysis of furnace temperature, furnace pressure, oxygen concentration, and nitrogen oxide concentration, this invention can accurately capture the dynamic fluctuations in combustion characteristics inside the boiler, improving the real-time perception capability of complex time-varying combustion states. Based on this, by adaptively matching historical combustion state clusters to establish the current combustion quality benchmark, it avoids misjudgments of operating conditions caused by using fixed air volume ratios or single fixed thresholds, providing a reliable dynamic reference standard for air supply regulation. Furthermore, based on the degree of difference between the current combustion state and the benchmark, the air volume adjustment direction and step size factor are determined, and the proportional parameters of the proportional-integral-derivative control system are dynamically corrected accordingly. This enables the air supply system to make timely and accurate adaptive responses to combustion fluctuations, effectively overcoming the problem of air supply regulation lag caused by traditional single feedback methods. Through the above regulation mechanism, this invention avoids localized oxygen deficiency and incomplete combustion or localized high-temperature oxygen enrichment caused by untimely regulation, ensuring boiler operational stability while reducing nitrogen oxide emissions. Attached Figure Description

[0035] Figure 1 This is a flowchart of an intelligent monitoring and control method for air supply in a low-NOx combustion technology for biomass boilers, as described in this invention.

[0036] Figure 2 This is a schematic diagram of historical combustion state clustering in this invention;

[0037] Figure 3 This is a comparison chart showing the improvement in combustion quality in this invention. Detailed Implementation

[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] For example, Figure 1 This is a flowchart of an intelligent monitoring and control method for air supply in a low-NOx combustion technology for biomass boilers according to the present invention. The specific implementation of the present invention will be described in detail below with reference to the flowchart.

[0040] This invention discloses a method for intelligent monitoring and control of air supply in low-NOx combustion technology of biomass boilers, comprising steps S1-S5:

[0041] S1: Collect furnace temperature data, furnace pressure data, oxygen concentration data, and NOx concentration data to construct a time-series dataset of boiler combustion status.

[0042] To achieve intelligent monitoring and air supply control of the low-NOx combustion process in biomass boilers, a furnace temperature sensor is installed inside the boiler furnace, a furnace pressure sensor is installed at the top of the furnace, and a nitrogen oxide (NOx) concentration sensor and an oxygen concentration sensor are installed in the boiler's tail flue. The temperature data collected by the furnace temperature sensor reflects the combustion intensity, the oxygen concentration data collected by the oxygen concentration sensor characterizes the air supply status during combustion, the NOx concentration data collected by the NOx concentration sensor reflects the low-NOx combustion effect, and the furnace pressure data collected by the furnace pressure sensor monitors combustion stability and airflow disturbance within the furnace.

[0043] By using furnace temperature sensors, NOx concentration sensors, oxygen concentration sensors, and furnace pressure sensors, key operating data such as furnace temperature, NOx concentration, oxygen concentration, and furnace pressure are collected in real time. The collected key operating data are then constructed into a boiler combustion state time-series dataset in chronological order, providing a data foundation for subsequent combustion state analysis and intelligent air supply control.

[0044] S2: Calculate the degree of fluctuation in temperature data and pressure data based on a time-weighted sliding window and obtain the combustion quality evaluation value.

[0045] It should be noted that in the low-NOx combustion process of biomass boilers, air supply control is a crucial factor affecting combustion stability, combustion efficiency, and NOx emission levels. Due to the complex composition, large moisture fluctuations, and high volatile content of biomass fuel, its combustion state fluctuates continuously with changes in fuel supply, boiler load, and furnace air distribution. Insufficient air supply can easily lead to incomplete combustion, localized oxygen deficiency, and decreased thermal efficiency; while excessive air supply can easily create localized high-temperature, oxygen-rich areas, resulting in increased NOx formation. Furthermore, the biomass boiler combustion process is highly dynamic and time-varying. Traditional fixed air volume ratios or delayed adjustment methods are insufficient to respond promptly to changes in furnace combustion state, easily causing incomplete combustion and delayed air supply regulation. Therefore, this invention monitors parameters such as furnace temperature, oxygen content, NOx concentration, and air supply status in real time, and dynamically adjusts the air supply volume according to changes in combustion state, thereby achieving the goals of reducing NOx emissions, improving combustion efficiency, and enhancing boiler operational stability.

[0046] Specifically, the collected boiler combustion state time series dataset is normalized and filtered to eliminate dimensional differences and noise effects between different types of data, thereby improving the consistency of subsequent analysis. In this embodiment, the maximum-minimum normalization method is used to normalize key operating data such as furnace temperature, NOx concentration, oxygen concentration, and furnace pressure in the boiler combustion state time series dataset, and the moving average filtering algorithm is used to filter the normalized key operating data. In other embodiments, implementers can select the normalization method and filtering algorithm according to the actual implementation situation.

[0047] A sliding monitoring window is set to continuously analyze the combustion state inside the furnace. In this embodiment, the length of the sliding monitoring window is set to 5 seconds, and the sliding step size is set to half the window length, i.e., 2.5 seconds. Key operating data within the current sliding monitoring window are acquired to analyze the current combustion state inside the furnace: Since data closer to the latest time within the sliding monitoring window better reflects the current combustion state, while data farther from the latest time reflects more historical operating states, different time weights are assigned to the data at each time point within the sliding monitoring window to improve the real-time performance of the current combustion state analysis. Data closer to the latest time has a higher weight, as expressed below:

[0048] ;

[0049] In the formula, Indicates the number of times the current sliding monitoring window is filled. The time weights corresponding to each data point; Represents the natural constant; This represents the normalization function. In this embodiment, the maximum and minimum value normalization function is used. In other embodiments, the implementer can choose the normalization function according to the actual implementation situation. Indicates the number of times the current sliding monitoring window is filled. Index of data points This indicates the index of the latest data point within the current sliding monitoring window. Indicates the first Data points and the latest data point The time distance between them; This indicates the number of data points currently within the sliding monitoring window. As the [number of data points increases / decreases]... The closer the time corresponding to the i-th data point is to the current time, the better the i-th data point... The smaller the time distance between the first data point and the latest data point, the better. The greater the time weight of each data point, the more real-time and accurate the analysis of the current combustion state in the furnace is improved.

[0050] It should be noted that during the combustion process of a biomass boiler, fluctuations in furnace temperature reflect the fuel combustion intensity and flame stability, while changes in furnace pressure reflect airflow disturbance and combustion uniformity. When furnace temperature and pressure remain relatively stable within a certain time range, and NOx concentration remains at a low level overall, it usually indicates that the air distribution inside the furnace is reasonable, the combustion state is stable, and the low-NOx combustion effect is good. Therefore, this invention uses temperature stability, furnace pressure stability, and overall NOx level as important criteria for evaluating the combustion quality inside the furnace.

[0051] Specifically, the fluctuation level of temperature or furnace pressure data within the current sliding monitoring window is obtained using time weighting to reflect the stability of combustion within the furnace within the current sliding monitoring window. This fluctuation level satisfies the expression:

[0052] ;

[0053] In the formula, This indicates the degree of fluctuation of the target parameter within the current sliding monitoring window, where the target parameter is normalized temperature data or normalized furnace pressure data. Indicates the number of times the current sliding monitoring window is filled. The time weights corresponding to each data point; This indicates the number of data points within the current sliding monitoring window; Indicates the number of times the current sliding monitoring window is filled. The target parameter values ​​corresponding to each data point; This represents the mean value of the target parameters within the current sliding monitoring window; This represents the weighted average absolute deviation of the target parameters within the current sliding monitoring window. The larger this value, the greater the fluctuation in combustion within the furnace and the worse the combustion stability.

[0054] Based on the fluctuations in temperature and pressure data, as well as the NOx concentration level, the combustion quality evaluation value within the current sliding monitoring window is obtained:

[0055] ;

[0056] In the formula, This indicates the combustion quality evaluation value within the current sliding monitoring window. Indicates the index of the current sliding monitoring window; Represents the maximum value function; and These represent the fluctuation levels of temperature data and pressure data within the current sliding monitoring window, respectively. Indicates the number of times the current sliding monitoring window is filled. The time weights corresponding to each data point; This indicates the number of data points within the current sliding monitoring window; Indicates the number of times the current sliding monitoring window is filled. The normalized NOx concentration data corresponding to each data point. This represents the maximum value between the fluctuation of temperature data and the fluctuation of pressure data within the current sliding monitoring window. It is used to characterize the current degree of combustion fluctuation in the furnace. The smaller the value, the more stable the current combustion state in the furnace. This represents the time-weighted value of the NOx concentration within the current sliding monitoring window. The smaller the value, the lower the current nitrogen oxide concentration and the better the combustion quality. This represents a weighting factor used to balance the weighting between in-furnace combustion stability and nitrogen oxide emission concentration. Its value ranges from (0,1), and in this embodiment, it is set to 0.5. If greater emphasis is placed on in-furnace combustion stability in actual production, this factor can be appropriately increased. The value of can be adjusted accordingly; conversely, if more emphasis is placed on the effectiveness of nitrogen oxide emission control, the value can be appropriately reduced. The value can be set according to actual production needs.

[0057] This invention utilizes the aforementioned time-weighted sliding window-based in-furnace combustion state analysis method, combined with furnace temperature fluctuations, furnace pressure fluctuations, and NOx concentration levels, to comprehensively evaluate the current boiler combustion state. This method can more realistically and accurately reflect the current in-furnace combustion stability and low-NOx combustion effect, effectively enhancing the ability to perceive changes in combustion state. It provides a more reliable decision-making basis for subsequent intelligent air supply control, which is beneficial for improving boiler combustion stability, reducing NOx emissions, and improving the overall operating efficiency of the boiler.

[0058] S3: Construct the current feature vector to match the historical reference cluster to determine the combustion quality benchmark value.

[0059] It should be noted that, due to the significant differences in furnace temperature and pressure of biomass boilers under different operating loads and combustion stages, evaluating the current combustion quality solely based on a uniform and fixed reference standard can easily lead to misjudgments of the combustion state due to variations in operating conditions. To improve the accuracy of combustion state analysis, this invention utilizes the time-weighted average of temperature data and furnace pressure data within the current sliding monitoring window as a characteristic representation of the current combustion stage in the furnace. Specifically, the calculation method is as follows: the time weight corresponding to each data point within the current sliding monitoring window is used as the weight; this time weight is then used to perform a weighted summation of the normalized temperature data within the current sliding monitoring window to obtain the time-weighted average of the temperature within the current sliding monitoring window; similarly, this time weight is used to perform a weighted summation of the normalized furnace pressure data within the current sliding monitoring window to obtain the time-weighted average of the furnace pressure within the current sliding monitoring window.

[0060] Furthermore, feature vectors for historical combustion stages are constructed based on the time-weighted average of temperature and the time-weighted average of furnace pressure corresponding to historical sliding monitoring windows. Using the Euclidean distance between the feature vectors corresponding to different historical sliding monitoring windows as the clustering distance criterion, the HDBSCAN clustering algorithm is used to cluster the historical sliding monitoring windows, forming multiple historical combustion state clusters corresponding to different combustion stages. It should be noted that since the HDBSCAN clustering algorithm does not require pre-setting parameters such as cluster radius and can achieve adaptive clustering based on the distribution of historical data, it is more suitable for combustion state classification under complex and fluctuating operating conditions of biomass boilers. Therefore, this embodiment uses this algorithm for processing.

[0061] To avoid the impact of historical abnormal operating conditions or noise data on clustering results, this embodiment constrains historical combustion state clusters by setting a minimum cluster size (min_cluster_size). This automatically classifies abnormal sliding monitoring windows with a small number of samples as noise points, thereby improving the stability and reliability of historical combustion state clusters. Specifically, to adapt to historical sample data of different scales, this embodiment uses a proportional adaptive method to determine the minimum cluster size parameter of the HDBSCAN clustering algorithm, the expression of which is as follows:

[0062] ;

[0063] In the formula, This represents the minimum number of samples in a cluster. This represents a user-defined scaling parameter with a value range of (0,1). In this embodiment, it will... The value is set to 0.1 because, according to empirical theory, a complete cluster can be formed at this scale, while clusters smaller than this scale may be noise or abnormal combustion stages; N represents the total number of historical sliding monitoring windows. This represents the floor function. It should be noted that, to ensure the effectiveness of clustering, a minimum number of samples is required for each cluster. That is, when When the calculation result is less than 2, take it directly. This invention, by setting a minimum number of cluster samples, can automatically divide small and discrete abnormal combustion windows into noise points, thereby avoiding interference from abnormal operating conditions with historical combustion state reference results.

[0064] By clustering the feature vectors of historical combustion stages within historical sliding monitoring windows, multiple historical combustion state clusters corresponding to different combustion stages can be obtained. Further, the current feature vector is constructed by combining the time-weighted average of temperature and the time-weighted average of furnace pressure corresponding to the current sliding monitoring window, and then matched with the historical combustion state clusters to obtain the reference cluster corresponding to the current combustion stage. Specifically, the matching method is as follows: the Euclidean distance between the current feature vector and the cluster center feature vector of each historical combustion state cluster is calculated, where the cluster center feature vector is the mean vector of all historical sliding monitoring window feature vectors within that historical combustion state cluster; the historical combustion state cluster with the smallest Euclidean distance is selected as the reference cluster corresponding to the current sliding monitoring window, thereby achieving the matching between the current combustion stage and historically similar combustion stages.

[0065] For example, Figure 2 This is a schematic diagram of historical combustion state clustering, such as... Figure 2 As shown in the figure, the horizontal axis represents the time-weighted mean of temperature, and the vertical axis represents the time-weighted mean of furnace pressure. Each independent data cluster represents a cluster of historical combustion states corresponding to multiple different combustion stages, formed after adaptive classification of the feature vectors of historical combustion stages using the HDBSCAN clustering algorithm. Figure 2 The clusters of historical combustion states exhibit a highly compact distribution characteristic. Based on this, the reference cluster core feature vector determined has extremely high representativeness of the operating conditions, providing an accurate and stable dynamic reference boundary for obtaining the combustion quality benchmark value under the current combustion stage.

[0066] It should be noted that the more concentrated the combustion quality evaluation values ​​of each historical sliding monitoring window within the reference cluster are, the more representative the evaluation value is of the combustion quality level of most historical sliding monitoring windows under the current combustion stage. Therefore, in order to establish a combustion quality reference baseline corresponding to the current combustion stage, this invention analyzes the degree of concentration of the distribution of combustion quality evaluation values ​​corresponding to each historical sliding monitoring window within the reference cluster.

[0067] Specifically, the reference cluster’s The distribution concentration of combustion quality evaluation values ​​corresponding to each historical sliding monitoring window satisfies the expression:

[0068] ;

[0069] In the formula, This indicates the reference cluster of the current sliding monitoring window. The distribution concentration of combustion quality evaluation values ​​corresponding to each historical sliding monitoring window; This represents the total number of historical sliding monitoring windows contained in the reference cluster, due to the minimum cluster sample size parameter of the HDBSCAN clustering algorithm. Greater than or equal to 2, therefore ; and This indicates the reference cluster of the current sliding monitoring window. The and the first Combustion quality evaluation values ​​corresponding to each historical sliding monitoring window; The function represents the normalization function. In this embodiment, the maximum and minimum value normalization function is used. In other embodiments, the implementer can choose the normalization function according to the actual implementation situation. This indicates the reference cluster of the current sliding monitoring window. The average absolute value of the difference between the combustion quality evaluation value corresponding to the current historical sliding monitoring window and the combustion quality evaluation values ​​corresponding to the remaining historical sliding monitoring windows in the reference cluster. The smaller this value, the more concentrated the majority of combustion quality evaluation values ​​in the reference cluster of the current sliding monitoring window are on the same historical sliding monitoring window. The combustion quality evaluation value corresponding to the first historical sliding monitoring window is near the first one, i.e., the first... The higher the combustion quality evaluation value corresponding to each historical sliding monitoring window, the more representative it is of the mainstream combustion quality level under the current combustion stage.

[0070] Based on the distribution concentration of combustion quality evaluation values ​​corresponding to each historical sliding monitoring window within the reference cluster of the current sliding monitoring window, the baseline combustion quality value for the current combustion stage is obtained:

[0071] ;

[0072] In the formula, This represents the combustion quality baseline value of the reference cluster in the current sliding monitoring window; This indicates the reference cluster of the current sliding monitoring window. The distribution concentration of combustion quality evaluation values ​​corresponding to each historical sliding monitoring window; This indicates the total number of historical sliding monitoring windows contained in the reference cluster; This indicates the reference cluster of the current sliding monitoring window. The combustion quality evaluation value corresponding to each historical sliding monitoring window; this invention will As a weight for the corresponding combustion quality evaluation value, the greater the distribution concentration, the more representative the corresponding combustion quality evaluation value is of the main quality distribution state under the current combustion stage. This represents the weighted average of the combustion quality evaluation values ​​of the reference cluster in the current sliding monitoring window. By weighting and fusing the combustion quality evaluation values ​​within the reference cluster, a more stable and representative combustion quality benchmark value can be obtained.

[0073] S4: Compare the combustion quality evaluation value with the combustion quality benchmark value to determine the airflow adjustment direction and airflow adjustment step size factor.

[0074] The combustion quality evaluation value corresponding to the current sliding monitoring window Combustion mass benchmark value corresponding to the reference cluster Comparative analysis: Response to This indicates that the current combustion quality has reached the mainstream quality level for the current combustion stage, at which point the boiler combustion state is relatively stable and airflow adjustment is not required; in response to If the current combustion quality is lower than the normal level corresponding to similar historical combustion stages, the decline in current combustion quality may be caused by insufficient air supply leading to incomplete combustion, or excessive air supply leading to local oxygen-rich combustion and increased NOx generation. In this case, further air supply optimization and adjustment are required.

[0075] To clarify the optimization direction of the current air intake, historical sliding monitoring windows with combustion quality evaluation values ​​higher than the current sliding monitoring window's combustion quality evaluation value are obtained and used as target reference sample points. The time weights corresponding to each data point within the current sliding monitoring window are used as weights, and the normalized oxygen concentration data within the current sliding monitoring window are weighted and summed to obtain the time-weighted oxygen concentration value for the current sliding monitoring window. Similarly, the time weights corresponding to each data point within the historical sliding monitoring windows corresponding to each target reference sample point are used as weights, and the oxygen concentration data within the corresponding historical sliding monitoring windows are weighted and summed to obtain the time-weighted oxygen concentration value for each target reference sample point.

[0076] Since the combustion quality at the target reference sample point is better than the current state, the current airflow adjustment direction is analyzed using the oxygen concentration level at the target reference sample point to obtain the airflow adjustment direction:

[0077] ;

[0078] ;

[0079] In the formula, This indicates the weighted value of oxygen concentration in the current sliding monitoring window compared to the first... Comparison results between the weighted values ​​of oxygen concentration at each target reference sample point; and They represent the first The time-weighted value of oxygen concentration for each target reference sample point and the current sliding monitoring window; This represents the time-weighted value of the oxygen concentration in the current sliding monitoring window and the first... The ternary expression for the comparison results of the time-weighted oxygen concentration values ​​of each target reference sample point, when... hour, The value is 1, otherwise The value is 0; Indicates the direction of airflow adjustment corresponding to the current sliding monitoring window; This indicates the number of target reference sample points corresponding to the current sliding monitoring window; This indicates the proportion of samples in the target reference sample point where the oxygen concentration is higher than that in the current sliding monitoring window. A ternary expression representing the comparison result between the sample proportion and a preset constant of 0.5, when... When the current combustion quality is optimal within the reference cluster, no airflow adjustment is needed; when... If the proportion of samples in the target reference sample point with an oxygen concentration higher than that in the current sliding monitoring window is relatively large, then... If the current oxygen concentration is lower than that of historical samples with better combustion quality, it indicates that the air intake should be increased appropriately to improve combustion completeness. A value of 1 indicates that the air intake needs to be increased; conversely, a value of 1 indicates that oxygen-rich combustion may be occurring, in which case the air intake needs to be reduced to decrease NOx formation and improve low-NOx combustion efficiency. A value of -1 indicates that the air intake volume needs to be reduced.

[0080] Furthermore, after obtaining the airflow adjustment direction corresponding to the current sliding monitoring window, in order to improve the airflow adjustment efficiency and avoid excessive or insufficient adjustment range, this invention obtains the airflow adjustment step size factor based on the difference between the combustion quality evaluation value of the current sliding monitoring window and the combustion quality benchmark value of the reference cluster:

[0081] ;

[0082] In the formula, Indicates the first The airflow adjustment step size factor corresponding to each sliding monitoring window. Indicates the index of the current sliding monitoring window; Represents the natural constant; This represents the combustion quality baseline value of the reference cluster in the current sliding monitoring window; This represents the combustion quality evaluation value of the current sliding monitoring window. The greater the difference between the current combustion quality evaluation value of the sliding monitoring window and the combustion quality benchmark value of the reference cluster, the higher the evaluation value. The larger the value, the more significantly the current combustion state deviates from the historical optimal combustion state. Therefore, the corresponding airflow adjustment step size factor is larger, thereby improving the targeting and response efficiency of airflow adjustment.

[0083] S5: Dynamically corrects proportional parameters and adjusts the air intake of the biomass boiler based on the air volume adjustment step size factor.

[0084] It should be noted that after obtaining the airflow adjustment direction and airflow adjustment step size factor corresponding to the current sliding monitoring window, this invention uses the airflow adjustment step size factor as a dynamic correction factor for the proportional adjustment parameter in the PID control system, so as to adaptively adjust the PID control intensity according to the changes in the current combustion state. Since the proportional parameter in the PID controller can reflect the system's response intensity to the current deviation, when the difference between the current combustion quality evaluation value and the reference value is large, the corresponding airflow adjustment step size factor is large. At this time, by increasing the PID proportional parameter, the response speed of the PID controller to the current air supply deviation is enhanced, thereby improving the air supply adjustment efficiency. Conversely, when the current combustion state gradually stabilizes and the difference from the reference value decreases, the corresponding airflow adjustment step size factor gradually decreases. At this time, the corrected PID proportional parameter will gradually approach the PID basic proportional parameter, thereby slowing down the air supply adjustment amplitude and avoiding furnace combustion disturbance caused by frequent fluctuations in air supply.

[0085] Specifically, the corrected PID proportional parameter satisfies the expression:

[0086] ;

[0087] In the formula, This indicates the corrected PID proportional parameter; This represents the basic proportional parameter of the PID controller; Indicates the first The airflow adjustment step size factor corresponding to each sliding monitoring window. The more significantly the combustion state of the current sliding monitoring window deviates from the historical optimal combustion state, the more... The larger the air volume adjustment step size factor corresponding to each sliding monitoring window, the larger the corrected PID proportional parameter will be, thereby improving the air supply adjustment response capability.

[0088] Furthermore, the corrected PID proportional parameters are substituted into the PID control system. The PID controller continuously adjusts the fan frequency or damper opening based on the deviation between the target air supply state and the current actual air supply state, thereby achieving dynamic and stable control of the boiler air supply and improving the operational stability and control reliability of the biomass boiler during low-NOx combustion.

[0089] Meanwhile, due to the inherent dynamic lag in the biomass boiler combustion process, after completing the current air supply adjustment, the system will continue to monitor the furnace temperature, furnace pressure, oxygen concentration, NOx concentration, and combustion quality evaluation value in real time within the subsequent sliding monitoring window. It will also iteratively execute steps S2 (combustion state analysis and combustion quality evaluation), S3 (reference cluster matching and combustion quality benchmark value acquisition), and S4 (air supply optimization adjustment). Through continuous iteration, a closed-loop dynamic control mechanism based on historical combustion state references and real-time combustion state perception is formed. This closed-loop dynamic control mechanism allows the boiler air supply to be continuously and adaptively adjusted according to changes in the combustion state within the furnace, thereby improving combustion stability and NOx control in the low-NOx combustion process of the biomass boiler, and enhancing overall operating efficiency.

[0090] For example, Figure 3 This is a comparison chart showing the improvement in combustion quality in this invention. Figure 3 The changes in combustion quality score before and after air supply adjustment using the method of this invention are shown. The horizontal axis represents the operating time, and the vertical axis represents the combustion quality evaluation value. Figure 3 The curve includes three lines. One curve represents the change in combustion quality evaluation value without adjustment; this curve fluctuates significantly and has a low overall evaluation value. Another curve represents the change in combustion quality evaluation value after adjustment using the method of this invention; this curve can quickly improve and stabilize at a higher level. The third dashed line represents the historical combustion quality benchmark value. Figure 3 As can be seen, after adjustment using the method of the present invention, the combustion quality evaluation value can be improved from a low level to a level close to the historical combustion quality benchmark value in a short period of time, and can fluctuate stably around the benchmark value during subsequent operation. In contrast, the combustion quality evaluation value in the unadjusted state remains at a low level and fluctuates violently. This indicates that the method of the present invention can effectively improve the combustion quality of biomass boilers, make the combustion state quickly approach the historical high-quality combustion state, and maintain stable operation.

[0091] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.

[0092] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A method for intelligent monitoring and control of air supply in a low-NOx combustion technology for biomass boilers, characterized in that, include: Real-time data collection of furnace temperature, furnace pressure, oxygen concentration, and nitrogen oxide concentration from biomass boilers is used to construct a time-series dataset of boiler combustion status. Set up a sliding monitoring window, and calculate the time-weighted values ​​of temperature data fluctuation, pressure data fluctuation, and nitrogen oxide concentration based on the time weight of each data point in the sliding monitoring window, so as to obtain the combustion quality evaluation value in the current sliding monitoring window. The current feature vector is constructed based on the time-weighted average of furnace temperature and pressure data within the sliding monitoring window; the current feature vector is matched with historical combustion state clusters pre-divided by density clustering algorithm to obtain a reference cluster corresponding to the current combustion stage, and the combustion quality benchmark value is determined based on the distribution concentration of historical data within the reference cluster. The combustion quality evaluation value in the current sliding monitoring window is compared with the combustion quality benchmark value. In response to the combustion quality evaluation value being lower than the combustion quality benchmark value, the airflow adjustment direction is determined based on the oxygen concentration level of high-quality historical samples in the reference cluster, and the airflow adjustment step size factor is determined based on the degree of difference between the combustion quality evaluation value and the combustion quality benchmark value. The proportional parameters of the proportional-integral-derivative control system are dynamically corrected by using the air volume adjustment step size factor. Based on the corrected proportional parameters and the air volume adjustment direction, the air intake of the biomass boiler is dynamically adjusted.

2. The intelligent monitoring and control method for air supply in a low-NOx combustion technology for biomass boilers according to claim 1, characterized in that, The time weights are distributed exponentially based on the time distance between the data points and the latest time. The fluctuations in the temperature and pressure data are calculated by weighting and summing the average absolute deviations of the corresponding data according to time weights.

3. The intelligent monitoring and control method for air supply in a low-NOx combustion technology for biomass boilers according to claim 1, characterized in that, The combustion quality evaluation value satisfies the expression: ； In the formula, This indicates the combustion quality evaluation value within the current sliding monitoring window. Indicates the index of the current sliding monitoring window; Represents the maximum value function; and These represent the fluctuation levels of temperature data and pressure data within the current sliding monitoring window, respectively. Indicates the number of times the current sliding monitoring window is filled. The time weights corresponding to each data point; This indicates the number of data points within the current sliding monitoring window; Indicates the number of times the current sliding monitoring window is filled. The normalized nitrogen oxide concentration data corresponding to each data point; This represents the weighting factor.

4. The intelligent monitoring and control method for air supply in a low-NOx combustion technology for biomass boilers according to claim 1, characterized in that, The acquisition of the reference cluster corresponding to the current combustion stage includes: A density-based clustering algorithm is used to adaptively cluster the feature vectors of historical combustion stages, forming multiple historical combustion state clusters corresponding to different combustion stages. The Euclidean distance between the current feature vector and the cluster center feature vector of each historical combustion state cluster is calculated, and the historical combustion state cluster with the smallest Euclidean distance is taken as the reference cluster corresponding to the current sliding monitoring window. The cluster center feature vector is the mean vector of all historical sliding monitoring window feature vectors within the corresponding historical combustion state cluster.

5. The intelligent monitoring and control method for air supply in a low-NOx combustion technology for biomass boilers according to claim 1, characterized in that, The distribution concentration satisfies the expression: ； In the formula, This indicates the reference cluster of the current sliding monitoring window. The distribution concentration of combustion quality evaluation values ​​corresponding to each historical sliding monitoring window; This indicates the total number of historical sliding monitoring windows contained in the reference cluster, and ; and This indicates the reference cluster of the current sliding monitoring window. The and the first Combustion quality evaluation values ​​corresponding to each historical sliding monitoring window; This represents the normalization function.

6. The intelligent monitoring and control method for air supply in a low-NOx combustion technology for biomass boilers according to claim 1, characterized in that, The combustion quality benchmark value satisfies the expression: ； In the formula, This represents the combustion quality baseline value of the reference cluster in the current sliding monitoring window; This indicates the reference cluster of the current sliding monitoring window. The distribution concentration of combustion quality evaluation values ​​corresponding to each historical sliding monitoring window; This indicates the total number of historical sliding monitoring windows contained in the reference cluster; This indicates the reference cluster of the current sliding monitoring window. Combustion quality evaluation values ​​corresponding to each historical sliding monitoring window.

7. The intelligent monitoring and control method for air supply in a low-NOx combustion technology for biomass boilers according to claim 1, characterized in that, The method of determining the airflow adjustment direction based on the oxygen concentration level of high-quality historical samples within the reference cluster includes: Obtain historical sliding monitoring windows whose combustion quality evaluation values ​​within the reference cluster of the current sliding monitoring window are higher than the current sliding monitoring window's combustion quality evaluation value, and use them as target reference sample points; calculate the time-weighted value of oxygen concentration in the current sliding monitoring window and the time-weighted value of oxygen concentration in each target reference sample point; count the percentage of target reference sample points whose time-weighted value of oxygen concentration is higher than the current sliding monitoring window, and based on the comparison result of the percentage and a preset constant, determine whether the air volume adjustment direction is to increase the air intake or decrease the air intake.

8. The intelligent monitoring and control method for air supply in a low-NOx combustion technology for biomass boilers according to claim 1, characterized in that, The air volume adjustment step size factor satisfies the expression: ； In the formula, Indicates the first The airflow adjustment step size factor corresponding to each sliding monitoring window; Represents the natural constant; This represents the combustion quality baseline value of the reference cluster in the current sliding monitoring window; This indicates the combustion quality evaluation value within the current sliding monitoring window. This indicates the index of the current sliding monitoring window.

9. The intelligent monitoring and control method for air supply in a low-NOx combustion technology for biomass boilers according to claim 1, characterized in that, The corrected scaling parameter satisfies the expression: ； In the formula, This indicates the corrected PID proportional parameter; This represents the basic proportional parameter of the PID controller; Indicates the first The airflow adjustment step size factor corresponding to each sliding monitoring window.

10. The intelligent monitoring and control method for air supply in a low-NOx combustion technology for biomass boilers according to claim 1, characterized in that, The dynamic adjustment of the air intake of the biomass boiler based on the corrected proportional parameters and the air volume adjustment direction includes: The corrected proportional parameters are input into the proportional-integral-derivative control system, and the controller continuously adjusts the fan frequency or damper opening based on the deviation between the target air supply state and the current actual air supply state. After completing the current air supply adjustment, the system continues to monitor the operating data in the subsequent sliding monitoring window in real time, and iteratively executes combustion state analysis, reference cluster matching, and air supply optimization adjustment steps to form a closed-loop dynamic control mechanism.

Images