A boiler air distribution uniform flow resistance reduction adaptive control method

Abstract

This invention relates to the field of cyber-physical systems and discloses an adaptive control method for boiler air distribution with uniform flow and reduced resistance. The method includes calculating airflow deviation values ​​based on the deviation between the flow capacity corresponding to the damper opening and the actual airflow, and associating these deviation values ​​with corresponding duct units; statistically analyzing the direction and magnitude of airflow deviation values ​​for each unit within adjacent periods, and calculating the cumulative impact of abnormal flow regions on the downstream duct airflow distribution; driving the duct mapping model to perform airflow redistribution calculations based on the influence transmission relationship, forming multiple digital operating scenarios corresponding to adjustment schemes; selecting candidate scenarios from the digital scenarios where the airflow uniformity coefficient is improved and the total duct pressure drop decreases, and loading the damper adjustment amounts of the candidate scenarios into the actual system in stages; using a scoring system to correct the resistance parameters in the duct mapping model, and selecting the scheme corresponding to the scenario with the highest score as the current air distribution adjustment strategy. This invention has the advantage of improving boiler combustion stability.

Description

Technical Field

[0001] This invention relates to the field of cyber-physical systems, specifically to an adaptive control method for boiler air distribution with flow equalization and resistance reduction. Background Technology

[0002] During boiler combustion system operation, primary air, secondary air, and staged air need to be delivered to different areas of the furnace in predetermined proportions to ensure complete fuel combustion and reduce flue gas flow resistance. Existing boiler air distribution control methods typically rely on closed-loop regulation based on feedback information from airflow measurement points, air pressure measurement points, and damper opening, combined with empirical models or static resistance models to determine the air distribution parameters for each duct. However, after long-term boiler operation, some ducts gradually accumulate ash, coke, wear and deformation, or experience localized blockages, causing changes in the actual flow cross-section of the ducts and altering the true flow resistance of each duct. However, the on-site measurement system can only directly acquire a limited amount of airflow, air pressure, and damper opening data, making it difficult to reflect real-time changes in the internal structural state of the ducts.

[0003] Especially when boiler load fluctuates frequently, fuel calorific value changes significantly, or the unit participates in deep peak shaving operations, the impact of changes in the resistance of different air ducts exhibits obvious nonlinear and time-varying characteristics. Some areas may experience excessive airflow, while others may suffer from insufficient air supply due to localized blockages. Since existing control systems primarily rely on real-time monitoring feedback for adjustment, by the time monitoring data changes, an imbalance in the airflow field has often already formed, further leading to problems such as localized combustion intensity deviations in the furnace, distortion of the flue gas flow field, and increased system pressure drop.

[0004] Therefore, it is essential to design an adaptive control method for boiler air distribution to reduce resistance and improve boiler combustion stability. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a boiler air distribution flow equalization and resistance reduction adaptive control method, which has the advantage of improving boiler combustion stability and solves the problems mentioned in the background technology.

[0006] To achieve the aforementioned goal of improving boiler combustion stability, this invention provides the following technical solution: a boiler air distribution flow equalization and resistance reduction adaptive control method, comprising the following steps: Obtain boiler branch air volume, wind box static pressure, damper opening, and furnace combustion characteristic parameters. Calculate the air volume deviation value based on the degree of deviation between the flow capacity corresponding to the damper opening and the actual air volume. Associate the deviation value with the corresponding air duct unit and construct an air duct mapping model corresponding to the physical air duct in the digital space. The direction and magnitude of the air volume deviation of each unit in adjacent periods are statistically analyzed. Units that continuously deviate in the same direction and have direct airflow connection in the duct topology are grouped into abnormal flow regions. The cumulative impact of abnormal flow regions on the air volume distribution of downstream ducts is calculated, and the influence transmission relationship between abnormal flow regions is established. Based on the influence transmission relationship, the airflow redistribution calculation is performed by the duct mapping model. The airflow change results under different damper adjustment amounts are simulated, and the corresponding airflow uniformity coefficient, total duct pressure drop and oxygen dispersion are recorded to form multiple digital operation scenarios corresponding to the adjustment scheme. Candidate scenarios with improved airflow uniformity coefficient and decreased total duct pressure drop are selected from digital scenarios. The damper adjustment of the candidate scenarios is loaded into the actual system in stages. Real-time data is collected and the deviation between the actual results and the corresponding digital scenarios is calculated. Valid scenarios with deviation values ​​below the threshold are retained. The comprehensive control score is calculated based on the flow equalization gain, drag reduction benefit, and fuel stability margin of each effective scenario. The score is used to correct the drag parameters in the duct mapping model, and the scheme corresponding to the scenario with the highest score is selected as the current air distribution adjustment strategy. In the next control cycle, based on the corrected duct mapping model and the updated real-time data, the step of statistically analyzing the air volume deviation of each unit in adjacent cycles is returned to form adaptive control.

[0007] Compared with the prior art, the present invention provides an adaptive control method for boiler air distribution flow equalization and resistance reduction, which has the following beneficial effects: This invention constructs a digital twin duct mapping model that operates synchronously with the actual boiler air supply system, unifying the correlation between airflow deviation, duct topology, and combustion operating parameters. This achieves digital mapping and dynamic tracking of the air distribution system's operating status. By identifying abnormal flow regions with continuous unidirectional deviations and airflow connectivity, and establishing the influence transmission relationship between these abnormal flow regions, it can accurately locate key duct areas causing airflow imbalance and increased local resistance, avoiding the blind adjustments caused by traditional air distribution regulation relying on experience. Furthermore, by using the influence transmission relationship to drive the digital model to perform airflow redistribution calculations, it simulates the airflow under different damper adjustment schemes in advance. The changes in flow distribution, pressure distribution, and furnace oxygen distribution are analyzed to create multiple verifiable digital operating scenarios. This allows for the prediction of flow equalization and drag reduction effects before actual adjustments, reducing the number of on-site trial adjustments. By loading candidate scenarios into the actual system in stages and verifying them using actual operating data, a two-way verification between the digital model and the physical system is achieved, making the selected effective scenarios closer to real operating conditions. A control scoring mechanism is established by comprehensively considering flow equalization gain, drag reduction benefit, and combustion stability margin. The drag and flow distribution parameters in the duct mapping model are continuously corrected using on-site verification results, giving the digital twin model self-learning and adaptive optimization capabilities. This not only improves the uniformity of airflow distribution in each branch and reduces the total pressure drop and fan energy consumption of the duct system, but also reduces the dispersion of furnace oxygen distribution, improves combustion stability and boiler operating safety, and achieves coordinated optimization control of flow equalization, drag reduction, and combustion stability in the boiler air distribution system. Attached Figure Description

[0008] Figure 1 This is a schematic diagram of the method of the present invention; Figure 2 This is a schematic diagram illustrating the calculation of airflow deviation values ​​according to the present invention. Detailed Implementation

[0009] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0010] Example 1: Please refer to Figure 1 An adaptive control method for boiler air distribution flow equalization and resistance reduction in an embodiment of the present invention includes the following steps: S1: Obtain boiler branch air volume, wind box static pressure, damper opening, and furnace combustion characteristic parameters. Calculate the air volume deviation value based on the degree of deviation between the flow capacity corresponding to the damper opening and the actual air volume. Associate the deviation value with the corresponding air duct unit and construct an air duct mapping model corresponding to the physical air duct in the digital space.

[0011] The process of calculating the air volume deviation value in S1 based on the degree of deviation between the flow capacity corresponding to the damper opening and the actual air volume is as follows: Air volume measuring devices, pressure measuring devices, and oxygen detection devices are installed in each air supply branch, air box, and furnace area of ​​the boiler to synchronously collect real-time air volume, air box static pressure, damper opening, furnace oxygen content, flame center position, and furnace outlet flue gas temperature for each branch. Air volume measuring devices, such as Pitot tube flow meters, thermal mass flow meters, or ultrasonic flow meters, are installed at the primary air branch, secondary air branch, and air supply branch pipes of each burner. Pressure transmitters are installed at the air box inlet and key nodes of each air duct to measure the air box static pressure. An opening feedback encoder is configured at the damper actuator to obtain the actual damper opening in real time. An oxygen analyzer is installed in the furnace outlet flue gas duct, and an industrial camera or infrared flame detection device is installed at the furnace observation port to obtain flame images. The control system synchronously collects data from each measuring point according to a unified time base and transmits it to the data processing server via the DCS system or industrial Ethernet. The coordinates of the flame center position are extracted using image recognition methods, and the flue gas temperature measurement data at the furnace outlet are recorded to form a multi-dimensional operation dataset that includes air volume, pressure, damper status and combustion status. Based on the structural dimensions, opening position, and design flow characteristic curves of each damper, the flow capacity under corresponding opening conditions is calculated, and the air volume data for each branch is generated. The dimensions of each damper blade, the cross-sectional area of ​​the duct, and the rated flow parameters are obtained in advance from the boiler design drawings, and a correlation curve between damper opening and flow coefficient is established through cold-state flow field tests or historical operating data calibration. For any damper opening value collected at any given time, the corresponding flow characteristic curve is used to determine the current flow coefficient, and the theoretical flow capacity is calculated in conjunction with the current air box static pressure and duct cross-sectional area. The theoretical air volume under corresponding conditions is calculated based on Bernoulli's equation and duct resistance characteristics, yielding the target air volume data that each branch should achieve under the current damper opening state. During the calculation process, the influence of air temperature changes on air density can be considered simultaneously, and density corrections can be applied to the theoretical air volume to improve the accuracy of the theoretical flow capacity calculation. The actual air volume collected in real time is compared with the corresponding air volume, and the difference and deviation ratio between the two are calculated to obtain the air volume deviation value for each branch. The actual air volume and theoretical air volume are matched one-to-one according to the branch number, and the air volume difference between the actual air volume and the theoretical air volume is calculated for each duct unit. At the same time, the deviation ratio is calculated using the ratio between the air volume difference and the theoretical air volume to reflect the degree of air volume deviation. When the actual air volume is lower than the theoretical air volume, the corresponding air volume deviation value is marked as a negative deviation state; when the actual air volume is higher than the theoretical air volume, the corresponding air volume deviation value is marked as a positive deviation state. Furthermore, in order to eliminate the influence of short-term fluctuations, the air volume deviation values ​​obtained in multiple consecutive sampling periods can be processed by moving average to form a stable air volume deviation sequence. This air volume deviation sequence can reflect the degree of deviation between the damper's execution effect and the actual airflow distribution, providing a basis for identifying local flow anomalies. By combining the static pressure change trend of the bellows and the combustion characteristic parameters of the furnace, the air volume deviation value is corrected according to the operating conditions, forming a standardized air volume deviation result that reflects the degree of deviation in the current air distribution state. The change trend of the bellows static pressure within a preset time window is statistically analyzed, and the static pressure change rate and fluctuation amplitude are calculated to identify overall air volume fluctuations caused by factors such as fan load changes, duct blockage, or switching of operating conditions. Combustion characteristic parameters such as furnace oxygen content, flame center position, and furnace outlet flue gas temperature are extracted to analyze whether phenomena such as flame deflection, local oxygen deficiency combustion, or excess air exist in the current combustion state. For the portion of the overall air volume that changes synchronously due to changes in fan operating conditions, a pressure compensation coefficient is used for correction; for the portion of the local air volume demand that changes due to changes in combustion state, a combustion correction coefficient is used for compensation, forming a standardized air volume deviation result. This standardized air volume deviation result retains information on air volume distribution imbalance while reducing the interference of operating condition fluctuations on the analysis results.

[0012] The process of constructing a wind tunnel mapping model in digital space corresponding to the physical wind tunnel in S1 is as follows: A digital duct topology network consistent with the physical ductwork is established based on the boiler air supply system structure, defining each branch damper, air box, distributor, and burner inlet as a corresponding ductwork unit node. The design drawings, duct layout diagrams, and equipment coding information of the boiler air supply system are retrieved to digitally model the ductwork structures in the primary air system, secondary air system, and auxiliary air supply system. According to the actual airflow path, key locations such as fan outlets, main air boxes, branch ducts, dampers, air volume distributors, and burner inlets are divided into independent ductwork units, each assigned a unique identifier. Connection edges between nodes are established based on the actual pipe connections, with airflow direction as the directional attribute of the edges, and duct length, cross-sectional area, and design flow rate as the basic attribute parameters of nodes and edges. For areas with multi-level branching or confluence structures, a tree-like topology is used to represent the network, ensuring the digital network fully reflects the airflow transmission relationships in the actual boiler ductwork. This method constructs a digital duct topology network consistent with the physical ductwork structure. The standardized airflow deviation results are mapped to the corresponding duct unit nodes according to the acquisition location, establishing a correspondence between airflow deviation values ​​and duct structure locations. The standardized airflow deviation results are matched with the measurement point number and equipment code to determine the duct unit location corresponding to each airflow measurement point. When the measuring device is directly installed in a duct unit, the standardized airflow deviation result is directly assigned to the corresponding node. When the measuring device is located between multiple duct units, a weighted allocation is performed based on the spatial distance between the measuring point and adjacent nodes and the airflow influence range. The airflow deviation value corresponding to each node is written into the digital duct topology network as a node operating status attribute, and the corresponding timestamp information is recorded synchronously. For duct units where measuring devices are not yet deployed, interpolation estimation can be performed based on the airflow deviation values ​​of adjacent nodes and the principle of airflow continuity, thereby forming a deviation status distribution map covering the entire duct system. By establishing the mapping relationship between airflow deviation values ​​and duct structure locations, the analysis can directly locate the location and propagation path of airflow anomalies. The connection relationships, design resistance parameters, and flow direction relationships between each duct unit are extracted to form a topological correlation matrix describing the airflow transmission path. Connection edge information between nodes is read from the digital duct topology network, and the starting node, ending node, and airflow direction corresponding to each connection edge are statistically analyzed. Further, based on duct design data, the design resistance coefficient, local resistance coefficient, number of bends, expansion / contraction port structural parameters, and damper flow resistance parameters of each duct section are extracted to form a corresponding resistance characteristic database. Using duct units as matrix nodes, an adjacency matrix is ​​constructed using the existence of direct airflow transmission relationships between nodes as matrix elements, and the flow direction relationship and resistance parameters are loaded into the matrix elements. When there is a direct connection between two duct units, the corresponding matrix element is assigned a non-zero correlation weight; when there is no direct airflow transmission relationship, the corresponding matrix element is assigned a value of zero. The correlation weights are corrected according to the resistance parameters so that the matrix can simultaneously characterize the airflow transmission direction and the ease of transmission, ultimately forming a topological correlation matrix describing the airflow propagation characteristics of the entire boiler duct. Based on the airflow deviation, static pressure status, and topological correlation matrix of each duct unit, a duct mapping model synchronously operating with the actual boiler duct is constructed in digital space. Synchronous operation, according to a preset refresh cycle, continuously updates the real-time collected airflow, static pressure, and damper opening data of each node in the physical duct to the corresponding nodes in the duct mapping model, and recalculates the dynamic airflow distribution and pressure drop status of each node based on mass conservation and pressure balance relationships. The digital duct topology network serves as the model skeleton, the airflow deviation as the node operating status parameter, and the topological correlation matrix as the airflow transmission constraint. Real-time collected data on box static pressure, branch static pressure, and damper opening are synchronously loaded into the corresponding nodes and connecting edges, forming a complete set of dynamic operating parameters. The system continuously updates the status of each node according to a preset refresh cycle, calculates the airflow exchange between nodes through mass conservation and pressure balance relationships, and corrects the flow distribution results corresponding to each connecting edge in real time. For areas where the airflow deviation continuously increases or the static pressure fluctuates significantly, the model automatically records its state change trajectory and updates the influence range of the associated nodes. By continuously receiving on-site operation data and updating model parameters synchronously, the air volume distribution, pressure distribution, and flow state in the digital space are kept consistent with the actual boiler air duct, thereby constructing an air duct mapping model that operates synchronously with the physical air duct.

[0013] S2: Statistically analyze the direction and magnitude of the airflow deviation of each unit within adjacent periods. Units that continuously deviate in the same direction and have direct airflow connectivity in the duct topology are grouped into abnormal flow regions. Calculate the cumulative impact of abnormal flow regions on the airflow distribution of downstream ducts and establish the influence transmission relationship between abnormal flow regions.

[0014] In S2, the process of combining units that continuously deviate in the same direction and have direct airflow connectivity in the duct topology into an abnormal flow region is as follows: The system invokes a duct mapping model to continuously acquire the airflow deviation value sequence of each duct unit within multiple adjacent sampling periods. It also invokes the duct mapping model to continuously read the standardized airflow deviation results corresponding to each duct unit at a preset sampling period. The sampling period can be set to 5 seconds, 10 seconds, or 30 seconds according to the boiler operating characteristics. Each sampling records the duct unit number, sampling time, airflow deviation value, and corresponding static pressure state. For each duct unit in continuous operation, airflow deviation data from multiple adjacent sampling periods are stored in chronological order to form a corresponding airflow deviation value sequence. For example, for a secondary air branch duct unit, the deviation value changes within the most recent N sampling periods can be continuously recorded to construct a time series database. To avoid measurement noise affecting subsequent analysis, moving average filtering, median filtering, or exponential weighted average methods can be used to smooth the original deviation values, resulting in a more stable airflow deviation value sequence. The direction and magnitude of increase / decrease between adjacent periods are calculated based on the airflow deviation value sequence to form corresponding deviation evolution characteristics. The airflow deviation value sequence of each duct unit is compared period by period to calculate the change in deviation value between the current sampling period and the previous sampling period. When the change is greater than zero, it is determined to be a direction of increasing deviation; when the change is less than zero, it is determined to be a direction of decreasing deviation; when the change is close to zero and lower than a preset fluctuation threshold, it is determined to be a stable state. Simultaneously, the proportion of the change in airflow deviation value between adjacent periods to the deviation value of the previous period is calculated as an indicator of the change magnitude. The direction changes, average change magnitude, maximum change magnitude, and duration of change are statistically analyzed over multiple consecutive sampling periods in chronological order to form a corresponding deviation evolution feature vector. This deviation evolution feature can reflect the development trend of the deviation state of a certain duct unit, providing a dynamic basis for identifying local flow anomalies, rather than relying solely on the degree of deviation at a single moment. A set of duct units with consistent deviation direction and amplitude exceeding a preset change threshold is selected from those that maintain the same deviation direction over multiple consecutive sampling periods. The preset change threshold is dynamically set based on the standard deviation of the airflow deviation values ​​of each duct unit during historical normal operation, ranging from 1.5 to 2.5 times the standard deviation. All duct units are screened and analyzed based on their deviation evolution characteristics. A continuous judgment window is set, such as 5 or 10 consecutive sampling periods, within which the consistency of the deviation direction of each duct unit is statistically analyzed. When a duct unit continuously maintains an increasing or decreasing deviation state throughout the judgment window, it is marked as a unit with continuous direction. The average change amplitude within the corresponding window is calculated and compared with the preset change threshold. When the average change amplitude exceeds the change threshold, it indicates that the deviation state of the duct unit is not a random fluctuation but a continuous development process. Duct units that simultaneously meet the conditions of continuous direction and amplitude exceeding the threshold are included in the set of units with continuous deviation. The deviation direction, duration, degree of deviation, and location within the duct are further recorded for each unit with continuous deviation. Based on the topological correlation matrix, the airflow connectivity between continuously deviating units is retrieved. Units with direct airflow transmission paths and consistent deviation directions are grouped to form an abnormal flow region. In the duct topology, from the outlet of the first duct unit to the inlet of the second duct unit, there are no intermediate nodes with independent airflow regulation functions or abrupt cross-sectional structures. Nodes with independent airflow regulation functions include dampers and regulating baffles; abrupt cross-sectional structures include distributor inlet narrowing and bellows branch interfaces. If two units are connected through a straight pipe section without regulation function or the internal space of a bellows with a constant cross-section, it is still considered a direct airflow transmission path. For units with direct connectivity, their deviation directions are compared to see if they are consistent, for example, both showing a continuous increase in airflow or both showing a continuous decrease in airflow. When multiple adjacent duct units simultaneously meet the conditions of direct connectivity and consistent deviation direction, they are combined and merged according to the airflow direction to form an abnormal flow region. For each formed abnormal flow region, the number of duct units included, the total deviation of the region, the average deviation degree of the region, and the pressure change characteristics within the region are statistically analyzed, and the upstream and downstream boundaries of the region are determined. The corresponding abnormal flow area identifier is generated in the air duct mapping model to achieve automatic location of local airflow imbalance areas.

[0015] The process of establishing the influence transmission relationship between abnormal flow regions in S2 is as follows: Extract information about each abnormal flow region and its contained duct units, and determine the corresponding airflow outlet boundary for each abnormal flow region. Retrieve the abnormal flow region information from the duct mapping model, reading the duct unit number, average airflow deviation, pressure state, and corresponding topological location for each abnormal flow region. Then, determine the main flow direction of the airflow within the region based on the duct topology, and search for the connection points between the abnormal flow region and external normal duct units along the main flow direction. When the last abnormal duct unit in the region connects with a downstream non-abnormal duct unit, this connection point is defined as the airflow outlet boundary; if the abnormal region has multiple downstream branching directions, determine the corresponding outlet boundaries for each. Record the airflow, static pressure, and connecting duct number information corresponding to the outlet boundaries to form an abnormal flow region boundary dataset. This clarifies the airflow output location between the abnormal flow region and the downstream duct system. The transmission path from each abnormal flow region to the downstream duct unit is traced along the airflow direction in the duct mapping model, recording the airflow attenuation ratio and static pressure change along the path. Starting from the airflow outlet boundary, the downstream duct unit is searched level by level along the airflow direction according to the topological correlation matrix in the duct mapping model. During the search, the duct nodes, branch nodes, and confluence nodes passed by the airflow are recorded to form the corresponding airflow transmission path. For each level of transmission node, real-time airflow and static pressure data are read, and the airflow change ratio and static pressure change relative to the previous level node are calculated. The airflow attenuation ratio can be determined by the ratio of the difference between the current node's airflow and the upstream node's airflow to the upstream node's airflow; the static pressure change can be calculated by the static pressure difference between adjacent nodes. For cases with multiple downstream branches, the airflow attenuation characteristics and pressure change characteristics corresponding to each branch path are recorded. Finally, a complete airflow transmission path database from the outlet boundary of the abnormal flow region to each downstream duct unit is formed to reflect the propagation process of abnormal conditions in the duct system. Based on the airflow deviation, path length, and pressure loss coefficient in each transmission path, the cumulative impact of abnormal flow regions on downstream airflow distribution is calculated. For each airflow transmission path, the airflow deviation value corresponding to the starting position of the path is extracted as the initial influence source intensity. Subsequently, the duct length, number of bends, number of diversions, and number of local resistance components along the path are statistically analyzed, and the corresponding pressure loss coefficient is obtained based on design data. The degree of influence attenuation during the propagation of abnormal states is calculated step by step along the transmission path. When the path length or pressure loss increases, the influence contribution of the abnormal state on downstream nodes is reduced accordingly. The influence contributions received by each node in the path are accumulated to obtain the comprehensive influence degree of the abnormal flow region on the corresponding downstream duct unit. For the case where the same duct unit is simultaneously affected by multiple abnormal flow regions, the influence contributions are superimposed to form the final influence result. The cumulative impact of the abnormal flow region on the downstream air volume distribution is obtained through the above calculations. This is used to quantify the propagation range and intensity of the abnormal state. The cumulative impact is calculated as follows: the air volume deviation at the outlet of the abnormal flow region is taken as the initial impact value. It is then multiplied by the attenuation coefficient of the corresponding pipe segment along the transmission path segment by segment, and the impact contribution of all downstream nodes on the path is summed. The attenuation coefficient is determined by normalization based on the product of the pressure loss coefficient of the corresponding pipe segment and the path length. This study statistically analyzes the influence intensity and transmission direction between abnormal flow regions via airflow paths, constructing a directed influence transmission network. Based on the cumulative influence, it analyzes the degree of influence each abnormal flow region has on the locations of other abnormal flow regions. When an abnormal flow region propagates into the duct range of another abnormal flow region, an influence transmission relationship is determined. Directed connections are established with the source region as the starting node and the affected region as the target node, with the cumulative influence as the weight of the corresponding connection. Characteristic parameters such as the main transmission path, transmission distance, and pressure attenuation during the influence propagation process are recorded. For cases where multiple abnormal flow regions act on the same area, corresponding directed connections are established for each region. Using abnormal flow regions as network nodes and influence intensity and transmission direction as connection edge attributes, a directed influence transmission network is constructed. This network not only reflects the propagation relationship of abnormal airflow in the boiler duct system but also identifies key influence source regions and main propagation channels.

[0016] S3: Based on the influence transmission relationship, the duct mapping model is driven to perform air volume redistribution calculation, simulate the air volume change results under different damper adjustment amounts, record the corresponding air volume uniformity coefficient, total duct pressure drop and oxygen dispersion, and form multiple digital operation scenarios corresponding to the adjustment scheme.

[0017] The process of simulating the air volume change results under different damper adjustment amounts in S3 is as follows: The directed influence transmission network is invoked, and the influence intensity of each abnormal flow region is input into the duct mapping model as an airflow adjustment constraint. The directed influence transmission network for abnormal flow regions is then invoked to extract the influence intensity, influence range, and influence transmission direction information corresponding to each abnormal flow region. Abnormal flow regions with high influence intensity are identified as priority adjustment regions; for duct units affected by multiple abnormal regions, their comprehensive influence intensity is calculated. The influence intensity of each abnormal flow region is converted into corresponding airflow adjustment constraint parameters, such as mapping the influence intensity to the maximum allowable airflow change range, the maximum damper opening change range, or the target deviation correction amount. Based on the current boiler load, total air supply demand, and combustion stability constraints, the above adjustment constraints are boundary-limited to avoid local air shortages or excessive air supply during airflow adjustment. The adjustment constraint parameters corresponding to each region are loaded into the duct mapping model, so that subsequent simulation calculations consider not only local airflow deviations but also the propagation impact of abnormal states throughout the entire duct network. The dampers corresponding to each abnormal flow area are selected as adjustment targets, and multiple combinations of different damper adjustment amounts are generated according to a preset adjustment step size. Based on the priority adjustment area information, the damper devices directly associated with each abnormal flow area are determined as adjustment targets. For example, for a secondary air branch corresponding to a certain abnormal flow area, its inlet regulating damper can be selected as the control object; for multiple abnormal areas sharing the same damper, this damper is used as the joint adjustment object. The damper adjustment step size is set, for example, using an opening change of 1%, 2%, or 5% as the basic adjustment unit, and the adjustable range is determined according to the current damper opening. On this basis, multiple adjustment targets are combined and arranged to generate multiple sets of different damper adjustment amount combination schemes. For example, if two dampers each have an adjustment range of 5%, candidate adjustment schemes corresponding to different opening combinations can be formed. Combination schemes that may lead to excessive total air supply or violation of operational safety constraints are eliminated, forming a set of damper adjustment schemes for simulation analysis. For each combination of damper adjustment amounts, the local resistance coefficient and flow capacity parameters of the corresponding duct unit are updated. For each combination of damper adjustment amounts, the flow coefficient of the corresponding damper is recalculated based on the change in damper opening. Since changes in damper opening alter the degree of local throttling, the local resistance coefficient of the duct unit containing the damper is updated synchronously. For dampers with increased opening, the local resistance coefficient is decreased accordingly; for dampers with decreased opening, the local resistance coefficient is increased. Based on the corrected flow coefficient, the flow capacity parameters of the duct unit are recalculated, including maximum allowable airflow, flow rate per unit pressure difference, and equivalent flow area. For duct areas affected by multiple dampers, the resistance parameters and flow capacity parameters of the relevant duct units are synchronously corrected according to the airflow transmission relationship. Through this parameter update process, the digital model can accurately reflect the changes in the flow characteristics of the duct structure after damper adjustment. Based on the updated duct mapping model, airflow redistribution calculations are performed to obtain the branch airflow distribution, pressure distribution, and airflow transmission status under each damper adjustment scheme. After parameter updates, the duct mapping model is invoked to perform airflow redistribution calculations using the boiler's current total air supply, fan operating conditions, and combustion load as boundary conditions. During the calculation, the inflow and outflow of each node are balanced according to the mass conservation principle, and the pressure distribution between each duct unit is solved based on the pressure balance relationship. For duct areas with branching and converging structures, the airflow is redistributed according to the updated resistance parameters, and the calculation is iterative until the flow and pressure at each node reach convergence. After the calculation is completed, the airflow distribution results for each branch, the static pressure distribution results for each key node, and the airflow transmission path status in the duct network are output. The deviation correction degree, downstream airflow balance, and pressure recovery status of each abnormal flow area are recorded, and these results are used as the digital operation results of the corresponding damper adjustment scheme.

[0018] The process of creating multiple digital operation scenarios corresponding to the adjustment scheme in S3 is as follows: Obtain the branch airflow distribution and pressure distribution results corresponding to each damper adjustment scheme; call the simulation results output after the airflow redistribution calculation is completed, and extract the branch airflow distribution data and pressure distribution data corresponding to each damper adjustment scheme one by one. For each digital simulation scheme, record the airflow results corresponding to the primary air branch, secondary air branch, and each burner inlet, and simultaneously record the static pressure data at the air box inlet, key nodes of the air duct, and burner inlet. To ensure comparability between different schemes, all schemes are calculated using the same boiler load, total air supply volume, and combustion boundary conditions. Establish a corresponding data storage structure according to the scheme number, and uniformly associate the airflow of each branch, node pressure, and damper adjustment amount to form the corresponding scheme operation dataset. Through the above processing, each damper adjustment scheme has complete airflow distribution and pressure distribution information; The airflow uniformity coefficient is calculated based on the deviation between the airflow of each branch and the average airflow, and the total pressure drop of the duct is obtained by statistically analyzing the pressure difference between the duct inlet and outlet. For each set of operational datasets, the airflow data of all target branches are statistically analyzed, and the average airflow of each branch is calculated. Subsequently, the deviation and proportion between the airflow of each branch and the average airflow are calculated, and the dispersion of the deviation values ​​of all branches is statistically analyzed. When the airflow of each branch is closer to the average airflow, it indicates a more balanced airflow distribution, and the corresponding airflow uniformity coefficient is higher; when the difference between branches is large, the corresponding airflow uniformity coefficient decreases. At the same time, the static pressure data corresponding to the inlet and outlet positions of the duct system are extracted, and the total duct pressure drop is obtained by subtracting the outlet static pressure from the inlet static pressure. For cases with multiple parallel duct branches, the pressure drop of each branch can be statistically analyzed separately, and a weighted average can be calculated as the overall total duct pressure drop. The airflow uniformity coefficient and total duct pressure drop evaluation results corresponding to each set of damper adjustment schemes are formed to measure the flow uniformity and drag reduction capabilities. Based on the current combustion characteristic parameters of the furnace and the coupled influence of the adjusted airflow distribution on the oxygen distribution in the furnace, the predicted oxygen value for each measuring point is calculated through the airflow-combustion coupling relationship in the duct mapping model. The standard deviation and coefficient of variation of the oxygen distribution in the furnace cross-section are calculated to obtain the oxygen dispersion. After completing the airflow distribution simulation, the oxygen supply results obtained from each burner region are used to call the furnace combustion prediction module to calculate the oxygen distribution state under the corresponding operating conditions. Specifically, an oxygen prediction model trained using historical operating data can be used, or a combustion numerical simulation model can be used to predict the oxygen level at different locations in the furnace. The predicted oxygen values ​​are obtained from multiple oxygen monitoring locations arranged along the furnace cross-section, and the average and standard deviation of the oxygen data at all measuring points are calculated. The standard deviation characterizes the degree of fluctuation in oxygen distribution, and the coefficient of variation is calculated by the ratio of the standard deviation to the average value, characterizing the consistency of oxygen distribution between different operating conditions. When the oxygen distribution is more uniform, its standard deviation and coefficient of variation are smaller; when local oxygen enrichment or deficiency is obvious, its standard deviation and coefficient of variation increase. The standard deviation and coefficient of variation are combined to form the oxygen dispersion index, which is used to reflect the degree of influence of the air volume adjustment scheme on the uniformity of combustion in the furnace. The airflow uniformity coefficient, total duct pressure drop, oxygen dispersion, and corresponding damper adjustment are correlated to form multiple digital operation scenarios. The results of the airflow uniformity coefficient, total duct pressure drop, and oxygen dispersion are uniformly summarized and mapped to the damper adjustment amounts of the corresponding schemes. For each damper adjustment scheme, its damper opening change information, branch airflow distribution status, pressure distribution status, airflow uniformity coefficient, total duct pressure drop, and oxygen dispersion are encapsulated and stored as complete operational features within the same scenario. A digital operation scenario library is established based on scenario numbers, recording the degree of improvement in abnormal flow areas, the effect of airflow deviation correction, and pressure changes at key nodes for each scenario. Newly generated digital operation scenarios are simultaneously written into the digital twin platform database, forming traceable scenario archives and creating multiple digital operation scenarios corresponding to different damper adjustment schemes.

[0019] S4: Select candidate scenarios from the digital scenarios where the air volume uniformity coefficient is improved and the total pressure drop of the air duct is reduced. Load the damper adjustment of the candidate scenarios into the actual system in stages, collect real-time data and calculate the deviation between the actual results and the corresponding digital scenarios, and retain the valid scenarios where the deviation value is lower than the threshold.

[0020] The process of loading the candidate scenario damper adjustment amount into the actual system in stages in S4 is as follows: Multiple digital operating scenarios were screened, and scenarios with a higher airflow uniformity coefficient and lower duct total pressure drop than the current operating state were selected as candidate scenarios. A digital operating scenario library was acquired, and the airflow uniformity coefficient and duct total pressure drop under the current actual boiler operating conditions were simultaneously read as benchmark evaluation values. The airflow uniformity coefficient, duct total pressure drop, and oxygen dispersion index corresponding to each digital operating scenario were extracted and compared with the current operating state. When the airflow uniformity coefficient of a digital operating scenario is higher than the current operating state, it indicates that the scenario can improve the uniformity of branch airflow distribution; when its duct total pressure drop is lower than the current operating state, it indicates that the scenario has the potential to reduce air supply resistance and fan energy consumption. Digital operating scenarios that simultaneously meet both of the above conditions are marked as candidate scenarios. Furthermore, an oxygen dispersion constraint can be added; scenarios with oxygen dispersion exceeding a preset upper limit are eliminated to avoid situations where flow uniformity and resistance reduction are achieved but combustion uniformity deteriorates. This forms a set of candidate scenarios that meet the requirements of flow uniformity optimization and resistance reduction optimization, providing alternative solutions for on-site loading verification. Candidate scenarios are ranked from highest to lowest based on their comprehensive control scores to form an adjustment priority sequence. For each candidate scenario, the corresponding target damper opening parameter is extracted and compared with the damper opening under the current actual operating state to calculate the opening change of each damper. The sum of the absolute values, maximum change, and average change of all damper opening changes in each candidate scenario are statistically analyzed to characterize the adjustment magnitude of the scenario to the current operating state. Scenarios with smaller adjustment magnitudes but significant optimization effects are prioritized to reduce the risk of operational disturbances during on-site adjustments. Scenarios are ranked from smallest to largest based on their opening changes, and a comprehensive evaluation is conducted based on the importance of abnormal flow areas, the intensity of influence in the influence transmission network, and the flow equalization gain of the corresponding scenario to generate a scenario loading priority. If multiple candidate scenarios have similar optimization effects, scenarios with smaller damper adjustment magnitudes are prioritized to form an adjustment priority sequence for the candidate scenarios. Based on the adjustment priority sequence, the target damper opening change is divided into multiple progressive loading stages, and a corresponding stable observation period is set for each stage. The damper opening in the actual boiler system is adjusted step by step according to the progressive loading stages. According to the adjustment priority sequence, the candidate scenario with the highest priority is selected as the current verification object. The target damper opening change between this candidate scenario and the current operating state is extracted and divided into multiple progressive loading stages according to a preset ratio. For example, when the target adjustment is 10%, it can be divided into four progressive stages: 25%, 50%, 75%, and 100%, corresponding to opening changes of 2.5%, 5%, 7.5%, and 10% respectively. For dampers with larger adjustment ranges, the number of loading stages can be further refined to reduce system disturbance. Simultaneously, a stable observation period is set for each loading stage, such as 3 minutes, 5 minutes, or 10 minutes, to ensure sufficient stabilization time for the boiler air supply system and combustion system after each adjustment. During the observation period, air volume, static pressure, oxygen content, and furnace combustion status are continuously monitored to ensure that the system has reached a new dynamic equilibrium state before entering the next loading stage. The control system invokes the damper adjustment command corresponding to the target candidate scenario and executes the damper adjustment sequentially according to the progressive loading stages. At each loading stage, a new opening setpoint is sent to the corresponding damper actuator, and real-time data on damper feedback opening, branch airflow, wind box static pressure, and furnace oxygen content are collected to confirm that the adjustment action has been accurately executed. During the stable observation period, the boiler operating status is continuously monitored. When airflow fluctuations, pressure anomalies, or oxygen deviations exceeding the preset safety threshold are detected, the subsequent loading process is paused and the system returns to the previous stable state. If the current operating status is stable and all parameters meet the safety operation requirements, the system automatically enters the next loading stage, gradually approaching the target damper opening corresponding to the candidate scenario. When all progressive loading stages are completed, the actual boiler system reaches the air distribution state corresponding to the candidate scenario, generating the corresponding on-site operating results.

[0021] The process of retaining valid scenarios with deviation values ​​below the threshold in S4 is as follows: After each loading stage is completed, real-time data on branch air volume, wind box static pressure, furnace oxygen content, and flue gas temperature distribution are collected for the corresponding time period. After the progressive loading stage is completed, the current damper opening remains unchanged, and a stable observation period begins. During the observation period, the boiler DCS system continuously collects operating parameters such as air volume, wind box static pressure, furnace oxygen content, and furnace outlet flue gas temperature for each air supply branch. For air volume data, real-time flow rates for primary air, secondary air, and burnout air branches are obtained; for pressure data, static pressure values ​​at the main wind box and key air duct nodes are collected; for oxygen content data, oxygen content at the furnace outlet and oxygen monitoring values ​​at different cross-sectional locations in the furnace are collected; for flue gas temperature data, the flue gas temperature at the furnace outlet and the multi-point flue gas temperature distribution in the flue are obtained. All parameters are continuously recorded according to a preset sampling period, and abnormal data such as start-up / shutdown fluctuations and instantaneous disturbances are removed to form an actual operating dataset for the corresponding loading stage, accurately reflecting the boiler's operating performance under the current air distribution conditions. Extract the predicted airflow distribution, predicted pressure distribution, and predicted oxygen distribution results for the corresponding stages in the digital operation scenario. Based on the current actual damper adjustment scheme and loading stage information, retrieve the corresponding digital operation scenario from the digital operation scenario library. Extract the predicted branch airflow distribution, duct pressure distribution, and furnace oxygen distribution results for this digital operation scenario under the same operating conditions. The airflow distribution results include the predicted airflow values ​​for each branch; the pressure distribution results include the predicted static pressure values ​​for each key node; and the oxygen distribution results include the predicted oxygen values ​​for each measuring point in the furnace and the corresponding oxygen dispersion index. To ensure the accuracy of the comparison results, the predicted flue gas temperature and combustion state results corresponding to the digital scenario can also be extracted simultaneously, and data matching is performed according to the measuring point locations consistent with the field data, thereby forming a prediction dataset that corresponds one-to-one with the actual operation data. The deviations in airflow, pressure, and oxygen content between the actual operating results and the predicted results of the digital operating scenario are calculated separately, and a comprehensive scenario deviation value is formed according to preset weights. The actual operating data and predicted data are matched according to the measurement point number, and the difference and relative error between the actual and predicted airflow for each branch are calculated. The average value of the airflow error for all branches is then calculated to form the airflow deviation value. The difference between the actual and predicted static pressure at each key pressure measurement point is calculated to form the pressure deviation value. Simultaneously, the difference between the actual and predicted oxygen content at each oxygen measurement point is calculated to form the oxygen deviation value. For cases where flue gas temperature prediction data exists, the flue gas temperature deviation value can be further calculated as an auxiliary verification indicator. Each deviation value is normalized by dividing it by the corresponding rated value or upper limit of the design range to obtain dimensionless airflow deviation factors, pressure deviation factors, and oxygen deviation factors. These deviation factors are then weighted and summed according to preset weights to form a comprehensive scenario deviation value. The preset weights are set based on the degree of influence of each parameter on the air distribution control effect, with the airflow deviation factor weight being 0.4–0.5, the pressure deviation factor weight being 0.3–0.4, and the oxygen deviation factor weight being 0.2–0.3, and the sum of the three being 1. This comprehensive scenario deviation value is used to characterize the consistency between the digital scenario and the actual operating results; the smaller the deviation value, the closer the prediction results of the digital twin model are to the actual operating conditions. Candidate scenarios with a comprehensive scenario deviation value lower than the scenario verification threshold are identified as valid scenarios, while other scenarios are removed from the digital operation scenario library. The comprehensive scenario deviation value is compared and analyzed with the preset scenario verification threshold, which can be set based on historical verification results, model prediction accuracy requirements, and boiler operation stability requirements. When the comprehensive scenario deviation value corresponding to a candidate scenario is lower than the scenario verification threshold, it indicates that the digital prediction result of that scenario can accurately reflect the actual operating state. Therefore, this scenario is marked as a valid scenario, and its corresponding damper adjustment scheme, air volume distribution results, and evaluation indicators are retained. For candidate scenarios with a comprehensive scenario deviation value higher than the scenario verification threshold, it is considered that their prediction results deviate significantly from the actual operating state and cannot be used as a reliable control basis; therefore, they are removed from the digital operation scenario library. Simultaneously, the error source information corresponding to this scenario is recorded, such as inaccurate local wind resistance parameters, large oxygen prediction errors, or pressure model deviations, for subsequent digital twin model correction. This forms a set of valid scenarios after on-site verification and screening, providing reliable candidate objects for the next step of comprehensive control score calculation and adaptive control strategy generation.

[0022] S5: Calculate the comprehensive control score based on the flow equalization gain, drag reduction benefit and fuel stability margin of each effective scenario. Use the score to correct the wind resistance parameters in the duct mapping model, and select the scenario with the highest score as the current air distribution adjustment strategy. In the next control cycle, based on the corrected duct mapping model and the updated real-time data, return to the step of executing the step of statistically analyzing the air volume deviation of each unit in adjacent cycles to form adaptive control.

[0023] The process of selecting the scenario with the highest score as the current air distribution adjustment strategy in S5 is as follows: Extract the airflow uniformity coefficient, total duct pressure drop, and oxygen dispersion index corresponding to each effective scenario; call the effective scenario set and read the corresponding operational evaluation parameters for each effective scenario from the digital twin platform database. The extracted data includes at least the airflow uniformity coefficient, total duct pressure drop, oxygen dispersion, damper adjustment amount, branch airflow distribution results, and pressure distribution results. For the same effective scenario, the furnace outlet flue gas temperature distribution data and combustion stability state data recorded during the field verification phase are also extracted simultaneously. The flow equalization gain is calculated based on the ratio of the increase in the airflow uniformity coefficient of the effective scenario relative to the current operating state to the current airflow uniformity coefficient. The drag reduction benefit is calculated based on the ratio of the difference between the current total duct pressure drop and the total duct pressure drop of the effective scenario to the current total duct pressure drop. The combustion stability margin is calculated by weighting and normalizing the inverse of the oxygen dispersion and the inverse of the flue gas temperature fluctuation amplitude, where the combustion stability margin is negatively correlated with both oxygen dispersion and flue gas temperature fluctuation amplitude. The airflow uniformity coefficient, total duct pressure drop, oxygen dispersion, and flue gas temperature fluctuation amplitude corresponding to the boiler's current operating state are obtained as baseline parameters. For each effective scenario, the increase ratio of its airflow uniformity coefficient relative to the current state is calculated, and this increase ratio is defined as the flow equalization gain index, used to characterize the degree of improvement in airflow distribution. The decrease ratio of the total duct pressure drop of the effective scenario relative to the current state is calculated, and this is defined as the drag reduction benefit index, used to characterize the reduction in duct resistance and the improvement in fan energy consumption. The distribution of oxygen content at various measuring points in the furnace and the changes in flue gas temperature at the furnace outlet were statistically analyzed under effective scenarios. When the oxygen content dispersion is small and the flue gas temperature fluctuation is low, it indicates a more stable combustion state. Based on this, a combustion stability margin index was established to reflect the stability of furnace combustion and the operational safety margin under the current air distribution scheme. Three core evaluation indicators were formed: flow equalization gain, drag reduction benefit, and combustion stability margin. After normalizing the flow equalization gain, drag reduction benefit, and fuel stability margin by dividing them by their respective maximum value ranges, the results are weighted and summed according to preset weights to obtain the comprehensive control score. Weight parameters for each evaluation indicator are set based on the boiler's operating objectives. For example, the weight of drag reduction benefit is appropriately increased under energy-saving priority conditions, the weight of fuel stability margin is increased under combustion stability priority conditions, and the weight of flow equalization gain is increased under severe airflow imbalance conditions. The flow equalization gain, drag reduction benefit, and fuel stability margin for each effective scenario are standardized to eliminate differences in the dimensions of different indicators. The comprehensive control score is then calculated using a weighted summation method according to preset weights. The comprehensive control score increases with increasing flow equalization gain, drag reduction benefit, and fuel stability margin. For scenarios where a single indicator is found to be abnormally low during the scoring process, a penalty coefficient can be added for correction to prevent a scenario where one indicator deteriorates significantly while other indicators are better from being mistakenly judged as the optimal solution. The comprehensive control score results for all effective scenarios are generated and sorted according to the score. The comprehensive control score of each effective scenario is compared with the benchmark score of the current operating state. The score difference is calculated, and the effective scenarios are sorted from largest to smallest according to the score difference. The effective scenario with the largest score difference, i.e. the highest comprehensive control score, is selected as the optimal scenario. The optimal scenario's damper adjustment values ​​are extracted. Based on the deviation between the current wind resistance parameters of each duct unit in the duct mapping model and the expected wind resistance parameters under the target operating state, an optimization problem is solved iteratively, with constraints of duct unit node mass conservation and loop pressure balance, and the objective function being the maximization of the comprehensive control score. The calculated correction values ​​are then superimposed onto the corresponding unit's wind resistance and flow distribution parameters in the duct mapping model. The scenario score results are compared with the scenario verification deviation results to analyze the sources of difference between the digital twin model's prediction results and the actual field results. When certain duct units consistently exhibit prediction errors in multiple effective scenarios, model parameter deviations are identified in that area. Combining the differences between actual operating data and digital scenario prediction data, the equivalent wind resistance change of the corresponding duct unit is calculated in reverse, and the local drag coefficient, friction coefficient, and flow distribution parameters in the duct mapping model are corrected. For diversion and convergence nodes, their airflow distribution ratio parameters can also be corrected based on field verification results. After parameter correction, model validation calculations are re-executed to ensure the updated digital duct mapping model more accurately reflects the actual duct operation. Through continuous reverse correction, the digital twin model achieves self-learning optimization, improving the accuracy of subsequent scenario predictions. The effective scenario with the highest comprehensive control score is selected, and the corresponding damper adjustment scheme is determined as the current boiler air distribution regulation strategy, serving as the initial operating scheme for the next control cycle. Based on the comprehensive control score ranking, the effective scenario with the highest score is selected as the current optimal operating scenario. The damper adjustment scheme corresponding to this optimal scenario is extracted, including the target opening degree of each branch damper, the air volume distribution ratio at each level, and the corresponding pressure control parameters. The control system writes this damper adjustment scheme into the boiler automatic control module and sets it as the current boiler air distribution regulation strategy. At the start of the next control cycle, the system directly uses this scheme as the initial operating state and continues to collect new operating data to perform air volume deviation analysis, abnormal flow identification, scenario simulation, and on-site verification processes. If the boiler load, fuel characteristics, or combustion state changes during subsequent operation, a new digital operating scenario is regenerated based on the updated operating data, and a new round of optimization decisions is executed. Through the above closed-loop optimization process, continuous flow equalization, drag reduction, and adaptive combustion control of the boiler air distribution system are achieved.

[0024] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "include," "contain," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0025] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A boiler air distribution flow equalization and resistance reduction adaptive control method, characterized in that, Includes the following steps: Obtain boiler branch air volume, wind box static pressure, damper opening, and furnace combustion characteristic parameters. Calculate the air volume deviation value based on the degree of deviation between the flow capacity corresponding to the damper opening and the actual air volume. Associate the deviation value with the corresponding air duct unit and construct an air duct mapping model corresponding to the physical air duct in the digital space. The direction and magnitude of the air volume deviation of each unit in adjacent periods are statistically analyzed. Units that continuously deviate in the same direction and have direct airflow connection in the duct topology are grouped into abnormal flow regions. The cumulative impact of abnormal flow regions on the air volume distribution of downstream ducts is calculated, and the influence transmission relationship between abnormal flow regions is established. Based on the influence transmission relationship, the airflow redistribution calculation is performed by the duct mapping model. The airflow change results under different damper adjustment amounts are simulated, and the corresponding airflow uniformity coefficient, total duct pressure drop and oxygen dispersion are recorded to form multiple digital operation scenarios corresponding to the adjustment scheme. Candidate scenarios with improved airflow uniformity coefficient and decreased total duct pressure drop are selected from digital scenarios. The damper adjustment of the candidate scenarios is loaded into the actual system in stages. Real-time data is collected and the deviation between the actual results and the corresponding digital scenarios is calculated. Valid scenarios with deviation values ​​below the threshold are retained. The comprehensive control score is calculated based on the flow equalization gain, drag reduction benefit, and fuel stability margin of each effective scenario. The score is used to correct the drag parameters in the duct mapping model, and the scheme corresponding to the scenario with the highest score is selected as the current air distribution adjustment strategy. In the next control cycle, based on the corrected duct mapping model and the updated real-time data, the step of statistically analyzing the air volume deviation of each unit in adjacent cycles is returned to form adaptive control.

2. The adaptive control method for boiler air distribution flow equalization and resistance reduction according to claim 1, characterized in that, The process of calculating the air volume deviation value based on the degree of deviation between the flow capacity corresponding to the damper opening and the actual air volume is as follows: Air volume measuring devices, pressure measuring devices and oxygen detection devices are installed in each air supply branch, air box and furnace area of ​​the boiler to synchronously collect real-time air volume, air box static pressure, damper opening, furnace oxygen content, flame center position and furnace outlet flue gas temperature of each branch. Based on the structural dimensions, opening position, and design flow characteristic curve of each damper, the flow capacity under the corresponding opening condition is calculated, and the air volume data corresponding to each branch is generated. The actual air volume collected in real time is compared with the corresponding air volume, and the difference and deviation ratio between the two are calculated to obtain the air volume deviation value corresponding to each branch. By combining the static pressure change trend of the air box and the combustion characteristic parameters of the furnace, the air volume deviation value is corrected according to the operating conditions, and a standardized air volume deviation result reflecting the degree of deviation of the current air distribution state is formed.

3. The adaptive control method for boiler air distribution flow equalization and resistance reduction according to claim 1, characterized in that, The process of constructing a wind tunnel mapping model in digital space that corresponds to the physical wind tunnel is as follows: A digital duct topology network consistent with the physical duct is established based on the structure of the boiler air supply system, and each branch damper, air box, distributor and burner inlet is defined as the corresponding duct unit node. The standardized airflow deviation results are mapped to the corresponding duct unit nodes according to the collection location, and the correspondence between the airflow deviation value and the duct structure location is established. Extract the connection relationships, design resistance parameters, and flow direction relationships between each air duct unit to form a topological correlation matrix describing the airflow transmission path; Based on the air volume deviation, static pressure state, and topological correlation matrix of each air duct unit, an air duct mapping model that operates synchronously with the actual boiler air duct is constructed in the digital space. The synchronous operation continuously updates the real-time collected data on air volume, static pressure, and damper opening of each node in the physical duct to the corresponding node in the duct mapping model according to a preset refresh cycle, and recalculates the dynamic air volume distribution and pressure drop status of each node based on the mass conservation relationship and pressure balance relationship.

4. The adaptive control method for boiler air distribution flow equalization and resistance reduction according to claim 1, characterized in that, The process of combining units that continuously deviate in the same direction and have direct airflow connectivity in the duct topology into an abnormal flow region is as follows: Call the air duct mapping model to continuously obtain the air volume deviation value sequence of each air duct unit in multiple adjacent sampling periods; The direction and magnitude of increase or decrease between adjacent periods are calculated based on the sequence of each air volume deviation value, forming the corresponding deviation evolution characteristics; Filter out air duct units that maintain the same deviation direction and whose change exceeds a preset change threshold in multiple consecutive sampling periods to determine the set of continuously deviating units. The preset change threshold is dynamically set according to the standard deviation of the air volume deviation value of each air duct unit in the historical normal operation period. Based on the topological correlation matrix, the airflow connectivity between the continuously deviating units is retrieved, and the continuously deviating units with direct airflow transmission paths and consistent deviation directions are combined to form an abnormal flow region. In the duct topology, there are no intermediate nodes with independent airflow regulation functions or abrupt cross-sectional changes between the outlet of the first duct unit and the inlet of the second duct unit.

5. The adaptive control method for boiler air distribution flow equalization and resistance reduction according to claim 1, characterized in that, The process of establishing the influence transmission relationship between abnormal flow regions is as follows: Extract information about each abnormal flow region and the included air duct units, and determine the airflow outlet boundary corresponding to each abnormal flow region. The transmission path from each abnormal flow region to the downstream duct unit is traced along the airflow direction in the duct mapping model, and the airflow attenuation ratio and static pressure change along the path are recorded. Based on the airflow deviation, path length, and pressure loss coefficient in each transmission path, the cumulative impact of the abnormal flow area on the downstream airflow distribution is calculated. The cumulative impact is calculated as follows: the airflow deviation at the outlet of the abnormal flow area is taken as the initial impact, multiplied by the attenuation coefficient of the corresponding pipe segment along the transmission path segment by segment, and the impact contribution of all downstream nodes on the path is summed. The attenuation coefficient is determined by normalization based on the product of the pressure loss coefficient of the corresponding pipe segment and the path length. The intensity and direction of influence formed between abnormal flow regions through airflow paths are statistically analyzed to construct a directional influence transmission network between abnormal flow regions.

6. The adaptive control method for boiler air distribution flow equalization and resistance reduction according to claim 1, characterized in that, The process of simulating the air volume changes under different damper adjustment amounts is as follows: The directed influence transmission network is invoked, and the influence intensity of each abnormal flow region is used as the air volume adjustment constraint input into the duct mapping model. Select the dampers corresponding to each abnormal flow area as the adjustment targets, and generate multiple combinations of different damper adjustment amounts according to the preset adjustment step size; For each combination of damper adjustment amounts, update the local resistance coefficient and flow capacity parameters of the corresponding duct unit; Based on the updated duct mapping model, the air volume redistribution calculation is performed to obtain the branch air volume distribution results, pressure distribution results, and airflow transmission status under each damper adjustment scheme.

7. The adaptive control method for boiler air distribution flow equalization and resistance reduction according to claim 1, characterized in that, The process of creating multiple digital operation scenarios corresponding to the adjustment plan is as follows: Obtain the branch air volume distribution and pressure distribution results corresponding to each damper adjustment scheme; The air volume uniformity coefficient is calculated based on the degree of deviation between the air volume of each branch and the average air volume, and the total pressure drop of the air duct is obtained by calculating the pressure difference between the inlet and outlet of the air duct. Based on the current combustion characteristic parameters of the furnace and the coupled influence of the adjusted air volume distribution on the oxygen distribution in the furnace, the oxygen prediction value corresponding to each measuring point is calculated through the airflow and combustion coupling relationship in the air duct mapping model. The standard deviation and coefficient of variation of the oxygen distribution in the furnace section are calculated to obtain the oxygen dispersion. By correlating the air volume uniformity coefficient, the total pressure drop in the air duct, the oxygen dispersion, and the corresponding damper adjustment, multiple digital operating scenarios are formed.

8. The adaptive control method for boiler air distribution flow equalization and resistance reduction according to claim 1, characterized in that, The process of loading the candidate scenario damper adjustment amount into the actual system in stages is as follows: Multiple digital operating scenarios were screened, and scenarios where the air volume uniformity coefficient was higher than the current operating state and the total duct pressure was lower than the current operating state were selected as candidate scenarios. The candidate scenarios are sorted from highest to lowest based on their comprehensive control scores to form an adjustment priority sequence; Based on the adjustment priority sequence, the target damper opening change is divided into multiple progressive loading stages, and a stable observation period is set for each stage. The damper opening in the actual boiler system is then adjusted step by step according to the progressive loading stages.

9. The adaptive control method for boiler air distribution flow equalization and resistance reduction according to claim 1, characterized in that, The process of retaining valid scenarios with deviation values ​​below the threshold is as follows: After each stage of loading is completed, real-time data on branch air volume, wind box static pressure, furnace oxygen content and flue gas temperature distribution are collected for the corresponding time period. Extract the predicted air volume distribution, predicted pressure distribution, and predicted oxygen distribution results for the corresponding stages in the digital operation scenario; The air volume deviation, pressure deviation, and oxygen deviation values ​​between the actual operating results and the predicted results of the digital operating scenario are calculated separately. Each deviation value is normalized by dividing it by the corresponding rated value or the upper limit of the design range to obtain dimensionless air volume deviation factors, pressure deviation factors, and oxygen deviation factors. The deviation factors are then weighted and summed according to preset weights to form a comprehensive scenario deviation value. The preset weights are set according to the degree of influence of each parameter on the air distribution control effect. Candidate scenarios with a comprehensive scenario deviation value lower than the scenario verification threshold are identified as valid scenarios, and the remaining scenarios are removed from the digital operation scenario library.

10. The adaptive control method for boiler air distribution flow equalization and resistance reduction according to claim 1, characterized in that, The process of selecting the scenario with the highest score as the current air distribution adjustment strategy is as follows: Extract the air volume uniformity coefficient, total duct pressure drop, and oxygen dispersion index for each effective scenario; The flow equalization gain is calculated based on the ratio of the increase in airflow uniformity coefficient of the effective scenario relative to the current operating state to the current airflow uniformity coefficient. The drag reduction gain is calculated based on the ratio of the difference between the current total pressure drop in the duct and the total pressure drop in the effective scenario to the current total pressure drop in the duct. The combustion margin is calculated by weighting and combining the reciprocal of the oxygen dispersion and the reciprocal of the flue gas temperature fluctuation amplitude, and then normalizing the result. The combustion margin is negatively correlated with both the oxygen dispersion and the flue gas temperature fluctuation amplitude. After normalizing the current equalization gain, drag reduction benefit and fuel stability margin by dividing them by their respective maximum value ranges, the comprehensive control score is obtained by weighting and summing them according to preset weights. The comprehensive control score of each effective scenario is compared with the benchmark score of the current operating state. The score difference is calculated, and the effective scenarios are sorted from largest to smallest according to the score difference. The effective scenario with the largest score difference, i.e. the highest comprehensive control score, is selected as the optimal scenario. Extract the adjustment amount of each damper corresponding to the optimal scenario. Based on the deviation between the current wind resistance parameter of each wind duct unit in the wind duct mapping model and the expected wind resistance parameter under the target operating state, solve the optimization problem with the wind duct unit node mass conservation and loop pressure balance as constraints and the maximization of comprehensive control score as the objective function. Iteratively calculate the wind resistance correction value and flow distribution correction value of each wind duct unit. The calculated correction value is then superimposed on the wind resistance parameter and flow distribution parameter of the corresponding unit in the wind duct mapping model. Select the effective scenario with the highest comprehensive control score, determine the corresponding damper adjustment scheme as the current boiler air distribution regulation strategy, and use it as the initial operation scheme for the next control cycle.

Images