A water-cooled wall thermal deviation fault online diagnosis method and system

Abstract

The present application relates to the technical field of power plant boiler intelligent operation and maintenance, and discloses a water cooling wall thermal deviation fault online diagnosis method and system, which collects water cooling wall multi-point temperature to construct a real-time temperature field, matches the current load, combustion and steam-water working condition to generate a reference temperature field; the two are compared to obtain a temperature residual field, filter abnormal nodes and merge spatial abnormal connected domains; time sequence tracking of the abnormal area constructs an event trajectory, quantifies abnormal propagation, expansion and residual characteristics; the quantified characteristics are matched with various thermal deviation fault mechanism characteristics to distinguish fault causes, combined with residual intensity, abnormal range to determine fault type, position, level and diagnosis reliability, so as to realize accurate online diagnosis of water cooling wall thermal deviation. The present application solves the defects of the prior art that the temperature field abnormal connected domain, propagation evolution and working condition lag correlation are not deeply analyzed, early fault identification is not accurate, the distinguishing degree is low and early warning is lagged, and can realize accurate early warning of thermal deviation faults.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance technology for power plant boilers, specifically to an online diagnostic method and system for water-cooled wall thermal deviation faults. Background Technology

[0002] The water-cooled wall of a power plant boiler is a core heat exchange and pressure-bearing component. It operates under complex conditions of coupled high-temperature radiation from the furnace and steam-water heat exchange within the tubes. Its working condition is highly susceptible to interference from multiple factors, including boiler combustion organization, fluctuations in coal quality, air-coal ratio, burner operation mode, coking and ash accumulation on the heating surfaces, and uneven hydrodynamic distribution. During actual boiler operation, imbalances in furnace heat load distribution or localized degradation of the water-cooled wall's heat exchange performance can induce localized thermal deviations, leading to a series of faults such as abnormally high tube wall temperature, surges in thermal stress, and deteriorated heat transfer. In severe cases, this can directly cause water-cooled wall tube rupture, threatening the operational stability and safety of the entire power plant boiler. Therefore, achieving online, accurate identification, location, and causal diagnosis of water-cooled wall thermal deviation faults is a core technical requirement for ensuring the long-term safe operation of the boiler.

[0003] Currently, some online safety evaluation technologies for water-cooled walls have been developed in the industry, such as the online safety evaluation method for water-cooled walls disclosed in patent CN101216164A. This method relies on a limited number of characteristic monitoring parameters within the water-cooled wall tubes, combined with a numerical calculation model of flow heat transfer, to online identify safety conditions such as water-cooled wall flow pulsation, wall temperature overheating, and fatigue creep fracture. It also clarifies the technical mechanism by which thermal deviations in the water-cooled wall exacerbate flow deviations, worsen heat transfer effects, and amplify wall temperature differences. This approach, to some extent, solves the industry problem of limited traditional measuring point placement and the inability to monitor the real-time operating conditions of the furnace across the entire area, providing effective technical support for the basic safety status evaluation of water-cooled walls.

[0004] However, existing mainstream technologies still have significant shortcomings. Their diagnostic logic focuses only on extrapolating heat transfer parameters within the pipe, calculating the temperature, flow rate, pressure, and lifespan of various regions of the water-cooled wall based on limited monitoring data. This only enables basic safety assessments such as wall temperature exceeding limits and flow anomalies, without conducting refined analysis of the spatial anomalies in the water-cooled wall temperature field. Consequently, it cannot meet the needs of tracing thermal deviation faults under complex actual operating conditions. In actual operation, thermal deviation faults caused by different factors exhibit differentiated characteristics: thermal deviations caused by combustion misalignment dynamically migrate along the flue gas flow direction; thermal deviations caused by coking, ash accumulation, and localized heat transfer deterioration show characteristics of fixed regions and gradual expansion of area; and thermal deviations caused by abnormal flow distribution on the steam-water side extend directionally along the working fluid flow direction.

[0005] Existing technologies lack in-depth analysis of the connectivity, propagation paths, expansion patterns, and lagging correlations between abnormal temperature field domains, operating condition disturbances, and fault evolution. In the early stages of a fault, before the wall temperature significantly exceeds limits, inaccurate fault cause identification, low fault type differentiation, and delayed early warning are common problems, making it difficult to achieve early prediction and accurate source tracing of thermal deviation faults. Therefore, there is an urgent need for an online diagnostic system based on the abnormal characteristics of the water-cooled wall temperature field. This system should accurately distinguish various types of thermal deviation faults through benchmark temperature field comparison, residual abnormal field analysis, and fault mechanism matching, thereby improving the reliability of early fault identification, location, and cause diagnosis. Summary of the Invention

[0006] In order to overcome the shortcomings of the existing technology, the present invention aims to provide an online diagnosis method and system for water-cooled wall thermal deviation faults, so as to solve the technical problems of the existing technology not conducting in-depth analysis of the abnormal connected domains, propagation paths, expansion laws and the lag correlation between operating condition disturbances and fault evolution, and the common problems of inaccurate fault cause identification, low fault type differentiation and delayed early warning in the early stage of fault and when the wall temperature has not significantly exceeded the limit.

[0007] This invention is achieved through the following technical solution: In a first aspect, the present invention provides an online diagnostic method for thermal deviation faults in water-cooled walls, comprising: Temperature data at different spatial locations of the water-cooled wall are collected, and the water-cooled wall in the furnace is spatially expanded to construct a real-time temperature field of the water-cooled wall. Based on the current operating conditions of the boiler, a reference temperature field is generated that is adapted to the current load, combustion and steam-water conditions. The real-time temperature field of the water-cooled wall is compared and calculated with the operating condition reference temperature field to obtain the temperature residual field. Based on the residual characteristics, abnormal candidate nodes are screened and merged to form a spatial abnormal connected domain. Within a continuous time window, the spatially abnormal connected domains are tracked and matched to construct abnormal event trajectories, and the abnormal propagation paths, expansion characteristics, and residual change characteristic parameters are quantified. The quantitative characteristics of abnormal events are matched and compared with the preset multi-type thermal deviation fault mechanism characteristics to distinguish different thermal deviation fault cause types. Based on the feature matching results, the thermal deviation fault type is determined, and combined with the residual strength, abnormal range, and operating condition correlation characteristics, the fault location, fault level, and diagnostic reliability are output to complete the accurate online diagnosis of water-cooled wall thermal deviation faults.

[0008] Preferably, the specific process of generating a reference temperature field that adapts to the current load, combustion, and steam-water conditions based on the current operating conditions of the boiler is as follows: The historical normal operating samples of the boiler are pre-divided into multiple operating condition clusters; Each of the aforementioned operating condition clusters corresponds to a normal temperature distribution pattern of the water-cooled wall under stable operating conditions; By matching the current boiler operating conditions with the operating condition clusters, a corresponding operating condition reference temperature field is generated.

[0009] Furthermore, the matching relationship between the current boiler operating condition and each operating condition cluster is specifically obtained by weighted calculation of the operating condition reference temperature field based on each operating condition cluster. The weight of each operating condition cluster is determined by the distance or similarity between the current boiler operating condition and the center of the corresponding operating condition cluster.

[0010] Preferably, the selection criteria for anomaly candidate nodes based on residual features are as follows: The positive residual exceeds the current operating condition residual threshold, and the residual gradient exceeds the gradient threshold, or multiple nodes in the neighborhood simultaneously exhibit abnormal positive residuals.

[0011] Preferably, within a continuous time window, the main propagation direction of the abnormal event is synthesized by weighting the migration vectors of the abnormal center in multiple consecutive time windows in the spatial abnormal connected domain. When the number of effective migration vectors is insufficient or the main propagation direction determination index is less than the lower limit of direction determination, the abnormal event is transferred to the center-preserving abnormal judgment process.

[0012] Furthermore, path stability is determined based on the degree of consistency between the migration vector of the anomaly center and the main propagation direction of the anomaly event, and expansion consistency is determined based on the degree of consistency between the major axis direction of the anomaly region and the main propagation direction of the anomaly event.

[0013] Preferably, the multi-type thermal deviation fault mechanism features include matching features of combustion skew type thermal deviation, matching features of fixed extended type thermal deviation, and matching features of abnormal flow distribution on the steam and water side type thermal deviation.

[0014] Furthermore, the matching characteristics of combustion skew thermal deviation include: the consistency between the abnormal main propagation direction and the furnace circumference, flue gas rotation direction or flue gas deflection direction, the circumferential offset ratio of the abnormal area, and the lag correlation between combustion side operating condition disturbance and residual growth. Matching characteristics of fixed extended thermal deviations include: the retention degree of the abnormal region center, the rate of abnormal area expansion, the rate of residual growth, the duration of abnormality, and the degree of abnormality that can be explained by the non-combustion side and the steam-water side conditions. The matching characteristics of abnormal thermal deviation in the steam-water side flow distribution include: consistency between the abnormal main propagation direction or the long axis direction of the abnormal region and the direction of the tube screen, the direction of the parallel loop or the direction of steam-water working fluid flow; the increasing state of the residual along the steam-water flow direction; and the lag correlation between steam-water side operating condition disturbance and residual growth. The abnormal characteristics of the temperature measurement link include abnormal dispersion, short duration, insufficient neighborhood support, and no stable propagation path.

[0015] Furthermore, the specific process for determining the type of thermal deviation fault based on feature matching results includes: The matching characteristics of combustion skew thermal deviation, fixed extended thermal deviation, abnormal flow distribution of steam and water side thermal deviation, and abnormal temperature measurement link are calculated respectively. When the maximum matching value exceeds the corresponding judgment threshold, the corresponding fault type is output.

[0016] Secondly, the present invention also provides an online diagnostic system for thermal deviation faults in water-cooled walls, comprising: The online temperature field construction unit collects temperature data at different spatial locations of the water-cooled wall, performs spatial unfolding processing on the furnace water-cooled wall, and constructs the real-time temperature field of the water-cooled wall. The operating condition reference temperature field generation unit generates an operating condition reference temperature field that is adapted to the current load, combustion and steam-water conditions based on the current operating conditions of the boiler. The residual abnormal field analysis unit compares and calculates the real-time temperature field of the water-cooled wall with the operating condition reference temperature field to obtain the temperature residual field, filters abnormal candidate nodes based on residual characteristics, and merges them to form a spatial abnormal connected domain. The anomaly propagation path tracking unit tracks and matches the spatial anomaly connected domains within a continuous time window, constructs anomaly event trajectories, and quantifies the anomaly propagation path, expansion characteristics, and residual change characteristic parameters. The thermal deviation mechanism matching unit matches and compares the quantitative characteristics of abnormal events with the preset multi-type thermal deviation fault mechanism characteristics to distinguish different thermal deviation fault cause types. The thermal deviation fault diagnosis unit determines the type of thermal deviation fault based on feature matching results, and outputs the fault location, fault level and diagnostic reliability by combining residual strength, abnormal range and operating condition correlation characteristics, thus completing the accurate online diagnosis of thermal deviation faults in water-cooled walls.

[0017] Compared with the prior art, the present invention has the following beneficial technical effects: This invention provides an online diagnostic method for thermal deviation faults in water-cooled walls. It abandons the traditional single-point temperature measurement over-limit alarm mode and generates a reference temperature field matching the load, combustion, and steam-water operation status based on the real-time boiler operating conditions. By comparing the real-time temperature field with the reference temperature field, a residual abnormal field is constructed. This method can eliminate temperature interference caused by normal operating condition fluctuations such as unit load changes and combustion strategy adjustments, ensuring that the diagnostic results are not affected by routine operational disturbances. This invention completely identifies residual abnormal connected domains and continuously tracks the main propagation direction, path stability, expansion consistency, area expansion rate, and residual growth rate of abnormal regions. It can capture the generation and expansion trend of abnormal regions before significant over-temperature occurs at a single point on the pipe wall, significantly advancing fault detection. This invention enhances early warning capabilities by identifying key nodes. It matches and distinguishes abnormal spatiotemporal evolution characteristics with four types of mechanistic features: combustion deviation, fixed coking and ash accumulation expansion, steam-water flow imbalance, and temperature measurement circuit faults. This effectively differentiates between instrument measurement point anomalies, short-term random disturbances, and actual water-cooled wall thermal deviation faults, avoiding false alarms. Based on multi-dimensional feature matching results, it outputs precise fault areas, fault levels, and diagnostic reliability, comprehensively improving the spatial positioning accuracy, fault type identification accuracy, and diagnostic reliability of water-cooled wall thermal deviation faults. This provides objective and comprehensive data support for on-site combustion optimization and adjustment of power plant boilers, planned maintenance of water-cooled wall heating surfaces, and risk prediction and control of unit operation. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the online diagnosis method for thermal deviation faults of water-cooled walls in an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating residual abnormal field analysis and abnormal connected domain formation in an embodiment of the present invention; Figure 3 This is a schematic diagram of the abnormal propagation path of the water-cooled wall unfolded surface in an embodiment of the present invention; Figure 4 This is a schematic diagram of thermal deviation cause matching and diagnostic output in an embodiment of the present invention; Figure 5 This is a schematic diagram of the online diagnostic system for thermal deviation faults of water-cooled walls in an embodiment of the present invention. Detailed Implementation

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

[0020] The purpose of this invention is to provide an online diagnostic method and system for thermal deviation faults in water-cooled walls, in order to solve the technical problems of existing technologies that do not conduct in-depth analysis of the abnormal connected domains, propagation paths, expansion laws, and the lag correlation between operating condition disturbances and fault evolution, and that in the early stage of the fault, when the wall temperature has not significantly exceeded the limit, there are generally inaccurate fault cause identification, low fault type differentiation, and delayed early warning.

[0021] The present invention will now be described in further detail with reference to the accompanying drawings: Example 1 See Figure 1 In one embodiment of the present invention, an online diagnostic method for thermal deviation faults in water-cooled walls is provided, comprising: Step 1: Collect temperature data at different spatial locations of the water-cooled wall, perform spatial unfolding processing on the furnace water-cooled wall, and construct the real-time temperature field of the water-cooled wall. Specifically, temperature data at different spatial locations of the water-cooled wall is acquired. The temperature data includes water-cooled wall temperature measurement point data, distributed optical fiber temperature measurement data, infrared temperature measurement data, or wall surface temperature data reconstructed from a finite number of measurement points.

[0022] The water-cooled wall is unfolded along the height and circumference of the furnace to form a two-dimensional spatial computational domain, and the nodes in the two-dimensional spatial computational domain are denoted as follows: ,in Indicates the numbering in the direction of furnace height. Indicates the circumferential, wall-facing, or pipe / screen direction number; At any moment ,node The corresponding real-time temperature is denoted as .

[0023] For nodes where temperature sensing elements are not directly installed, spatial completion is performed based on the temperatures of adjacent measuring points, the spatial distance between measuring points, the arrangement of tube panels, and historical temperature change trends, so that a real-time temperature field corresponding to the spatial position of the water-cooled wall is formed at each sampling moment. .

[0024] Step 2: Based on the current operating conditions of the boiler, generate a reference temperature field that is suitable for the current load, combustion, and steam-water conditions. Specifically, a reference temperature field is generated based on the current operating status of the boiler. The current operating status includes at least one of the following: unit load, coal feed rate, coal mill operation combination, burner layer operation status, primary air volume, secondary air volume, furnace oxygen volume, main steam flow rate, feedwater flow rate, steam drum pressure, furnace outlet flue gas temperature, reheat steam temperature deviation, and wind box differential pressure.

[0025] Time The operating condition characteristics are expressed as Based on historical normal operation samples, a correspondence is established between operating condition characteristics and the normal temperature distribution of the water-cooled wall, and a reference temperature field under the current operating condition is output. .

[0026] The reference temperature field is not a fixed historical average temperature field, but is generated based on normal samples that match the current load level, combustion organization, and steam-water side conditions. It is used to characterize the normal temperature distribution that is allowed at various locations of the water-cooled wall under the current operating conditions.

[0027] The historical normal operation samples are divided into multiple operating condition clusters, and each operating condition cluster corresponds to a normal temperature distribution pattern of the water-cooled wall under a stable operating state. When real-time operating conditions After input, the distance or similarity between the input and the center of each operating condition cluster is calculated, and the current reference temperature field is generated by weighting the normalized similarity.

[0028] node The reference temperature is determined by the following formula:

[0029] in, Indicates the number of operating condition clusters. Indicates the first Nodes in the normal temperature field corresponding to each operating condition cluster The reference temperature Indicates the current operating condition and the first The matching weights between each set of working conditions satisfy the following:

[0030] The matching weight Determine using the following formula:

[0031] in, Indicates real-time operating condition characteristics With the The distance between the centers of each working condition cluster.

[0032] In this way, when the boiler load increases or decreases, the burner operation combination is adjusted, or the coal mill operation combination changes, the system can update the normal temperature reference according to the change of operating conditions, avoiding misjudging the overall temperature redistribution during normal adjustment as a thermal deviation fault.

[0033] Step 3: Compare and calculate the real-time temperature field of the water-cooled wall with the operating condition reference temperature field to obtain the temperature residual field. Based on the residual characteristics, filter out abnormal candidate nodes and merge them to form a spatial abnormal connected domain. Specifically, the real-time temperature field With reference temperature field By comparison, the temperature residual field is obtained. .

[0034] For nodes Its temperature residual is calculated using the following formula:

[0035] when When the value is positive and continues to increase, it indicates that the actual temperature of the node is higher than the normal reference value under the current operating conditions; When multiple adjacent nodes simultaneously form a positive residual region, it indicates that the region may have non-uniform heating enhancement or local heat transfer capacity reduction.

[0036] To identify the degree of abrupt boundary changes between anomalous residual regions and surrounding normal regions, the residual gradient is further calculated. :

[0037] in, Represents a node The residual gradient at that point.

[0038] If the overall residual in a certain area increases but the gradient is small, it is more likely that it is a gradual temperature rise caused by load changes. If a region has both a high positive residual and a significant residual gradient, it indicates that the region forms an abnormal thermal boundary relative to the neighboring region, which is more consistent with the spatial behavior of thermal deviation faults.

[0039] Specifically, according to Figure 2 As shown, the real-time temperature field is compared with the reference temperature field to obtain the temperature residual field, and anomaly connected domains are formed based on residual intensity, residual gradient and spatial connectivity. Initial outlier nodes are generated based on residual strength, and the boundaries of outlier regions are identified and isolated fluctuations are eliminated using residual gradients.

[0040] For nodes If a node's positive residual exceeds the residual threshold under the current operating condition, and its residual gradient exceeds the gradient threshold, or if multiple nodes in its neighborhood simultaneously exhibit positive residual anomalies, then that node is identified as an anomalous candidate node. Anomalous candidate node set. Determine using the following formula:

[0041] in, This represents the residual threshold under the current operating conditions. This represents the gradient threshold under the current operating conditions. Represents a node The neighborhood, This represents the average residual of nodes within the neighborhood.

[0042] In this way, the system can identify thermal deviation anomalies with obvious boundary abrupt changes, as well as retain fixed extended anomalies that are locally fixed, have an overall increase in residuals, but have weak boundary gradients.

[0043] The residual threshold and gradient threshold are adaptively determined based on the residual distribution and gradient distribution of historical normal samples within the current matching condition cluster, rather than being fixed to a single empirical value.

[0044] The residual threshold Take the mean of the residuals of the normal samples in the current operating condition cluster and... The gradient threshold is the sum of multiples of the standard deviations. Take the mean of the residual gradient of the normal sample in the current operating condition cluster and... The sum of two standard deviations.

[0045] For the set of abnormal candidate nodes Spatial connectivity analysis is performed, and spatially adjacent outlier candidate nodes with the same residual sign are merged into outlier connected components. Each outlier connected component is denoted as . ,in Indicates the number of the abnormal connected component.

[0046] For abnormal connected components Calculate its abnormal area Maximum residual Mean residual Regional Center Boundary gradient mean Major axis direction and regional eccentricity .

[0047] The abnormal area is used to represent the range of thermal deviation influence, the maximum residual is used to represent the local thermal intensity, the region center is used to determine the wall surface, height range and tube screen number where the abnormality is located, the mean of the boundary gradient is used to determine whether the abnormal region has a clear thermal boundary, the major axis direction is used to determine whether the abnormal region expands along the furnace circumference, along the height direction or along the tube screen direction, and the region eccentricity is used to determine whether the abnormal temperature zone is offset relative to the geometric center of the water-cooled wall or the combustion center of the furnace.

[0048] Step 4: Track and match the spatially abnormal connected domains within a continuous time window, construct the abnormal event trajectory, and quantify the abnormal propagation path, expansion characteristics, and residual change characteristic parameters. Specifically, according to Figure 3 As shown, the generation, expansion, migration and stabilization process of anomalous connected domains are tracked within the computational domain of the water-cooled wall, and the migration sequence of the anomalous region center and the main propagation direction are formed. Track the generation, expansion, migration, and stabilization of anomalous connected components within a continuous time window.

[0049] For time abnormal connected components and time abnormal connected components If the two overlap in space, or the distance between their center regions is less than a preset distance threshold, and their average residuals have the same sign or both belong to the positive residual abnormal region, then the two will be matched as the continuous state of the same abnormal event.

[0050] Establish an abnormal trajectory for each abnormal event. The abnormal trajectory includes the abnormal start time, starting node, starting region, continuous coverage region, region center migration sequence, abnormal area change sequence, residual intensity change sequence, abnormal morphology change sequence, and propagation direction change sequence.

[0051] For abnormal events , and at time The regional center is denoted as Let its center migration vector between two adjacent sampling windows be denoted as :

[0052] in, Used to indicate the direction and distance of migration of anomaly regions between two adjacent time windows.

[0053] It also calculates the rate of anomalous area expansion. and residual growth rate :

[0054]

[0055] in, Indicates an abnormal event At any moment The abnormal area, Indicates an abnormal event At any moment The average residual.

[0056] like and This indicates that the range and intensity of the abnormal area are developing simultaneously. like Smaller but A consistently positive value indicates that the location of the abnormal area is relatively fixed, but the degree of thermal deviation is increasing. like Maintaining similar directions across multiple time windows indicates that the abnormal region has a clear propagation path.

[0057] To ensure that propagation path determination is not limited to a single center offset, anomalous events are synthesized based on center migration vectors within multiple consecutive time windows. Main direction of communication :

[0058] in, This indicates the number of time windows used for path identification. Indicates the first The weights of each historical time window are determined, and the following conditions are met:

[0059] Before calculating the main propagation direction, remove center migration vectors whose magnitude is less than the effective lower bound of migration. When the number of effective transfer vectors is insufficient, or If the value is less than the lower limit of the direction determination, direction cosine matching is not performed, and the anomaly is transferred to the center-preserving anomaly judgment process.

[0060] Path stability Used to indicate whether multiple valid transfer vectors continue along similar directions:

[0061] in, Indicates satisfaction and The effective set of migration windows, and As the effective lower bound of direction, This indicates the number of valid migration windows.

[0062] like A higher value indicates that the abnormal region propagates along a stable direction within a continuous time window; like A lower value indicates that the migration direction of the abnormal area is chaotic, and it is more likely to be due to noise disturbance, abnormal measurement point, or short-term operational fluctuation.

[0063] Extended Consistency Used to indicate whether the shape extension direction of the abnormal region is consistent with the main propagation direction:

[0064] in, Indicates the time of the abnormal event The direction of the shape's extension.

[0065] like A higher value indicates that not only has the center of the anomalous region migrated, but the anomalous region itself has also extended in the same direction, suggesting that the anomalous region has stronger thermal deviation propagation characteristics.

[0066] In calculating path stability and extended consistency If the magnitude of the center migration vector, the magnitude of the main propagation direction, or the magnitude of the major axis of the abnormal region is less than the corresponding effective lower limit, then the corresponding directional index will not be used for cause matching. For abnormal connected regions with excessively small areas, insufficient major-minor axis ratios, or near-circular shapes, expand consistency. Treat it as a low-confidence feature.

[0067] This process avoids distortion of direction matching results caused by fixed anomalies, minor anomalies, or noise at measurement points.

[0068] To avoid the calculation of abnormal propagation paths being affected by single-point jitter or fixed anomalies, an abnormal event is only identified after the abnormal connected component has persisted for a preset minimum time. .

[0069] Propagation path features are generated based on the anomaly initiation location, main propagation direction, path stability, expansion consistency, anomaly area expansion rate, residual growth rate, and regional eccentricity. :

[0070] in, Indicates the starting position of the abnormal event. Indicates the main direction of propagation. Indicates path stability. Indicates extended consistency. Indicates the rate of anomalous area expansion. Indicates the residual growth rate. Indicates the eccentricity of the region.

[0071] The propagation path characteristics are used to describe where the thermal deviation anomaly originates, where it extends, whether the propagation direction is stable, whether the anomaly morphology is consistent with the propagation direction, and whether the anomaly intensity increases synchronously.

[0072] Step 5: Match and compare the quantitative characteristics of the abnormal event with the preset multi-type thermal deviation fault mechanism characteristics to distinguish different thermal deviation fault cause types. Specifically, according to Figure 4 As shown, the propagation path characteristics are matched with the mechanistic characteristics corresponding to different causes of thermal deviation, and the diagnostic results are output using a two-level diagnostic method of first matching causes and then determining the level. Characteristics of the propagation path Matching the mechanistic characteristics with the causes of different thermal deviations.

[0073] The mechanistic features include combustion skewness, fixed expansion, abnormal flow distribution on the steam-water side, and abnormal temperature measurement link.

[0074] Since both coking and ash accumulation and localized heat transfer deterioration exhibit localized stagnation, continuous enhancement, and slow expansion in the temperature residual field, this embodiment uses them as primary causes for identification of fixed-expansion thermal deviations, and retains two candidate explanations—coking and ash accumulation and localized heat transfer deterioration—in the output results.

[0075] When performing cause matching, first perform flow segmentation judgment: like If effective, proceed to directional propagation matching, focusing on judging combustion skew thermal deviation and abnormal flow distribution thermal deviation on the steam-water side; like If the center retention, area expansion rate, and residual growth rate of the invalid but abnormal area meet the characteristics of fixed expansion anomaly, then enter the center retention matching and focus on judging the thermal deviation of coking and ash accumulation or local heat transfer deterioration. If there is neither an effective main propagation direction nor continuous expansion and residual enhancement, then the anomaly is considered a candidate for transient disturbance or temperature measurement link anomaly.

[0076] Before performing causal matching, residual strength, outlier area, residual growth rate, area expansion rate, path stability, expansion consistency, center preservation, neighborhood support, and operating condition lag correlation are normalized to map features of different dimensions to a unified value range.

[0077] For combustion skew type thermal deviation, the key is to determine whether the main propagation direction of the abnormality is consistent with the furnace circumference, flue gas rotation direction or flue gas deviation direction, and to determine whether the residual growth is lagging behind the coal feed deviation, coal mill output difference, air volume deviation or burner operation change.

[0078] The direction of the combustion-side mechanism is represented as Combustion skew direction matching degree Calculate using the following formula:

[0079] Simultaneously, calculate the circumferential offset ratio of the abnormal region. :

[0080] in, This indicates the coordinates of the center of the abnormal event in the circumferential or wall direction. The circumferential coordinates represent the moment the anomaly began. This represents the total number of circumferential nodes in the computational domain of the water-cooled wall.

[0081] If the abnormal region migrates stably along the circumference of the furnace or the direction of flue gas deflection, and its residual growth has a reasonable lag relationship with the combustion-side disturbance, then the abnormality is more consistent with the thermal deviation caused by combustion skew.

[0082] To characterize the delayed response between combustion-side disturbances and abnormal residual growth, at least one of the following variables—coal feed rate deviation, pulverizer output difference, burner operating status change, primary air volume deviation, secondary air volume deviation, and furnace oxygen content change—was selected as a combustion-side operating condition variable, and the combustion-side hysteresis correlation was calculated. :

[0083] in, Represents the set of combustion-side operating condition variables. This indicates the lag time between the operating condition disturbance and the temperature residual response. This represents the set of allowable lag times on the combustion side. This indicates the correlation calculation within a set time window.

[0084] The Determined based on boiler thermal inertia, sampling cycle, and historical operating data.

[0085] For fixed-expansion thermal deviations, the main criterion is not whether the abnormal area migrates along a certain direction, but rather whether the abnormal area has the characteristics of being locally fixed, persistent, slowly expanding in area, and continuously increasing residual.

[0086] Abnormal events Center retention Calculate using the following formula:

[0087] in, This represents the maximum reference distance in the computational domain of the water-cooled wall.

[0088] when A higher value indicates that the center of the abnormal region has changed relatively little relative to the starting position.

[0089] If the abnormal region remains fixed within a local height range for a long period of time, and the rate of expansion of the abnormal area is... and residual growth rate If the value remains positive and neither the combustion side nor the steam-water side operating conditions can explain this abnormal change, then the abnormality is more consistent with a fixed-expansion thermal deviation caused by coking and ash accumulation or local heat transfer deterioration.

[0090] For abnormal thermal deviations in the flow distribution of steam and water, the key is to determine whether the abnormal area extends along the direction of the water-cooled wall tube screen, the direction of the parallel loop, or the direction of the steam and water working fluid flow, and to determine whether the residuals show an increasing or continuously high trend along the flow direction.

[0091] When the main propagation direction When valid, determine In relation to the mechanism of steam and water Consistency; When the main propagation direction Invalid but abnormal region long axis direction When valid, determine In relation to the mechanism of steam and water Consistency.

[0092] Soft drink side-direction matching degree Calculate using the following formula:

[0093] Among them, when or If any direction is invalid, the cosine matching term corresponding to that direction is not used; only the matching term corresponding to the valid direction is used.

[0094] In this way, anomalies on the steam-water side can manifest as central migration along the steam-water flow direction, or as the formation of a stable high residual band along the pipe-screen direction or the parallel loop direction.

[0095] Increasing consistency of residuals calculated along the direction of steam flow :

[0096] in, Indicates the first [number]th [unit] along the direction of steam / water flow within the area covered by the abnormal event. The average residual at each location This indicates the number of positions involved in the calculation along the direction of the soda flow. Indicates taking when the condition is met. Otherwise take .

[0097] If the abnormal region extends along the direction of the pipe screen, the parallel loop, or the flow direction of the steam and water working fluid, and the residual shows a continuous increase or a continuous increase along the steam and water flow direction, it indicates that the abnormality is more consistent with the thermal deviation caused by abnormal flow distribution on the steam and water side.

[0098] Soft drink side lag correlation Calculate using the following formula:

[0099] in, This represents the set of steam and water side operating condition variables, which include at least one of the following: feedwater flow rate, steam drum pressure, main steam flow rate, evaporation rate, circulation flow rate, and water-cooled wall inlet flow rate. This represents the set of allowable lag times on the carbonated beverage side. If the abnormal residual growth is strongly correlated with changes in carbonated beverage side variables within a reasonable lag time, then the causal matching result for abnormal carbonated beverage flow distribution is improved.

[0100] For anomalies in the temperature measurement link, the key is to determine whether the anomaly lacks spatial continuity, neighborhood support, and a stable propagation path.

[0101] If anomalies occur only in a single node or a small number of discrete nodes, the candidate nodes for the anomalies cannot form stable connected components, or the formed anomalous connected components have small areas, short durations, and low path stability. Low, extended consistency If the temperature deviation is low and neither the residuals of adjacent nodes nor the residual gradients support the anomaly, it will be judged as a measurement point inaccuracy, temperature measurement link interference, or local data anomaly, instead of directly outputting a thermal deviation fault alarm.

[0102] Step 6: Determine the type of thermal deviation fault based on the feature matching results, and combine the residual strength, abnormal range, and operating condition correlation characteristics to output the fault location, fault level, and diagnostic reliability, thus completing the accurate online diagnosis of water-cooled wall thermal deviation faults.

[0103] Specifically, a two-level diagnostic approach is adopted, which first matches the causes and then determines the level.

[0104] First, candidate fault types for abnormal events are determined based on the combustion skew matching results, fixed extended matching results, abnormal matching results of steam-water side flow distribution, and abnormal matching results of temperature measurement link. Then, after the candidate fault types are determined, the fault level is output according to the severity index corresponding to that type.

[0105] This approach avoids directly outputting the fault type using only a uniform weighted score, making the diagnostic process more consistent with the performance differences caused by different thermal deviations.

[0106] The matching values ​​for each cause are obtained by weighted summation of the corresponding normalized features:

[0107] in, The causes are indicated, including combustion deviation, fixed expansion, abnormal steam-water flow distribution, and abnormal temperature measurement link. Indicates the first The first type of cause A normalized matching feature, Represents the corresponding weight, and satisfies , .

[0108] The weights are determined based on historically labeled fault samples, expert rules, or the principle of minimizing the misjudgment rate.

[0109] Among them, the combustion deviation cause matching value is jointly determined by the combustion deviation direction matching degree, circumferential offset ratio, path stability, expansion consistency and combustion side hysteresis correlation degree; The fixed extended cause matching value is jointly determined by the center retention degree, the rate of anomalous area expansion, the rate of residual growth, the duration of anomalous events, and the degree of lack of explanation for combustion-side or steam-water-side conditions. The matching value of the cause on the steam-water side is jointly determined by the directional matching degree on the steam-water side, the consistency of residual increase, the number of abnormal coverage tubes, the residual growth rate, and the hysteretic correlation degree on the steam-water side. The anomaly matching value of the temperature measurement link is determined by the anomaly dispersion, neighborhood support, anomaly duration, path stability, and extension consistency.

[0110] When a certain cause matching value is the maximum value and exceeds the corresponding judgment threshold, the corresponding fault type is output; When multiple matching values ​​of causes exceed their respective judgment thresholds and the difference is less than the preset difference threshold, the composite thermal deviation status is output, and the candidate causes are sorted according to the size of the matching values.

[0111] The determination threshold can be determined based on the mean and standard deviation of the corresponding cause matching values ​​in historical normal samples, or it can be corrected based on historical labeled fault samples or manual confirmation results.

[0112] The mean of the corresponding cause-matching values ​​in the normal operating samples and The sum of multiples of the standard deviations is used as the initial judgment threshold; When a manually confirmed thermal deviation fault sample occurs during unit operation, the judgment threshold and the weights of each matching feature are adjusted according to the principle of minimizing the false judgment rate.

[0113] After determining the fault type, the fault level is determined based on the severity index of the corresponding fault type.

[0114] Calculate the severity value based on the severity index of the corresponding fault type. :

[0115] in, Indicates the fault type. Indicates the first Class of faults A normalized severity characteristic, This represents the corresponding severity weight, and satisfies... , .

[0116] when When the value is less than the first-level threshold, a slight thermal deviation is output. when When the value is greater than or equal to the first-level threshold and less than the second-level threshold, a moderate thermal deviation is output. when When the value is greater than or equal to the second-level threshold, a severe thermal deviation is output.

[0117] For combustion-skewed thermal deviations, the severity of the fault is primarily determined by the circumferential offset ratio. Path stability Residual growth rate Mean residual Determining the correlation between combustion-side hysteresis and the degree of hysteresis; For fixed extended thermal deviations, the failure level is primarily determined by center retention. Abnormal area expansion rate Mean residual Residual growth rate And the duration is determined; For thermal deviations caused by abnormal flow distribution on the steam and water sides, the fault level is primarily determined by the matching degree of the steam and water side direction. Increasing consistency of residuals Number of abnormal coverage screens, average residual The correlation between the lag and the soda side was determined.

[0118] For abnormal temperature measurement links, the output of measurement point or temperature measurement link check prompts will not be regarded as the actual water-cooled wall thermal deviation fault level output.

[0119] During the diagnosis of thermal deviation faults, if a continuous positive residual region first appears in the upper middle part of the water-cooled wall front wall, and then this abnormal region migrates along the furnace circumference towards the left wall within multiple sampling windows, the main propagation direction is... Direction of flue gas deflection Consistency, path stability and extended consistency All meet the requirements for determining combustion skewness, and the residual growth rate is... If the delay time is set and the output difference between adjacent coal mills increases or the air-coal ratio deviation increases, then a combustion skew type thermal deviation fault will be output, and the wall surface, height range and pipe screen number covered by the abnormal connected domain will be used as the fault location result. If a certain abnormal region remains fixed within a local height range of the water-cooled wall for a long period of time, the center of the region... Relative to the starting center The migration distance is small, and the center retention is high. It meets the criteria for fixed expansion, but has an abnormal area. and mean residual If the increase is slow and no combustion-side or water-side hysteresis condition correlation that can explain the abnormal change is detected, a fixed extended thermal deviation fault will be output, with coking and ash accumulation and local heat transfer deterioration as candidate causes. If the abnormal region gradually extends along the same tube screen direction or parallel loop direction, the main propagation direction is... or the long axis direction of the abnormal region Flow direction of the working fluid Consistent, and the residuals increase in consistency along the direction of soda flow. The requirements for determining the quality of the soda / water side are met, and the residual growth rate is also met. If there is a hysteresis relationship with fluctuations in feedwater flow, steam drum pressure, evaporation rate, or circulation flow, then the output steam-water side flow distribution is an abnormal thermal deviation fault.

[0120] To improve diagnostic stability during variable operating conditions, when boiler load changes rapidly, burner operation status changes, or coal mill combination is adjusted, the reference temperature field is updated first, and the residual threshold is temporarily raised during the transition period of the operating condition. and gradient threshold ; Reduce the reliability of faults in abnormal regions that appear during the transition period but have not formed a stable main propagation direction, have low path stability, or low expansion consistency.

[0121] Once the operating conditions enter a stable range, the system will restore the normal discrimination criteria for the corresponding operating condition cluster.

[0122] Therefore, the system can avoid misjudging the overall temperature field redistribution during normal operation and adjustment as a thermal deviation fault.

[0123] The output diagnostic results include the location of the fault area, fault level, fault type, propagation path, main propagation direction, cause matching results, lagged correlation variables, diagnostic confidence, and handling suggestions.

[0124] The location of the fault area is determined by the abnormal connectivity domain coverage node and the center of the area, which can correspond to the front wall, rear wall, left wall, right wall, specific height range and tube panel number of the water-cooled wall; The fault type is determined based on the cause matching results; The fault level is determined based on the severity index under the corresponding fault type; The reliability of the diagnosis is determined comprehensively based on the spatial continuity of the abnormal area, duration, path stability, expansion consistency, measurement point consistency, and the degree of interpretation of operating condition lag.

[0125] Example 2 according to Figure 5As shown, this embodiment also provides an online diagnostic system for thermal deviation faults in water-cooled walls, including: The online temperature field construction unit collects temperature data at different spatial locations of the water-cooled wall, performs spatial unfolding processing on the furnace water-cooled wall, and constructs the real-time temperature field of the water-cooled wall. The operating condition reference temperature field generation unit generates an operating condition reference temperature field that is adapted to the current load, combustion and steam-water conditions based on the current operating conditions of the boiler. The residual abnormal field analysis unit compares and calculates the real-time temperature field of the water-cooled wall with the operating condition reference temperature field to obtain the temperature residual field, filters abnormal candidate nodes based on residual characteristics, and merges them to form a spatial abnormal connected domain. The anomaly propagation path tracking unit tracks and matches the spatial anomaly connected domains within a continuous time window, constructs anomaly event trajectories, and quantifies the anomaly propagation path, expansion characteristics, and residual change characteristic parameters. The thermal deviation mechanism matching unit matches and compares the quantitative characteristics of abnormal events with the preset multi-type thermal deviation fault mechanism characteristics to distinguish different thermal deviation fault cause types. The thermal deviation fault diagnosis unit determines the type of thermal deviation fault based on feature matching results, and outputs the fault location, fault level and diagnostic reliability by combining residual strength, abnormal range and operating condition correlation characteristics, thus completing the accurate online diagnosis of thermal deviation faults in water-cooled walls.

[0126] The system is used to identify uneven heating of the water-cooled wall, combustion deviation, coking and ash accumulation, local heat exchange deterioration, abnormal steam-water flow distribution, and abnormal temperature measurement link during boiler operation.

[0127] The system does not use the exceeding of a fixed threshold at a single wall temperature measurement point as a direct basis for judging thermal deviation faults. Instead, it generates a water-cooled wall reference temperature field based on the current boiler operating conditions, transforms the deviation of the real-time temperature field relative to the reference temperature field into a residual abnormal field, and further tracks the starting position, main propagation direction, expansion consistency, path stability, and operating condition lag correlation of the residual abnormal region within a continuous time window. This allows the system to determine the location, type, level, and reliability of the thermal deviation fault. This embodiment uses positive residual abnormalities as the object of thermal deviation fault diagnosis. For negative residual regions, they can be used as auxiliary information for assessing diagnostic reliability, such as weakened combustion, enhanced local cooling, or abnormal temperature measurement.

[0128] In summary, the online diagnostic method and system for water-cooled wall thermal deviation faults provided by this invention no longer rely on single wall temperature measuring points exceeding limits or fixed temperature difference thresholds for water-cooled wall thermal deviation alarms. Instead, it generates a reference temperature field based on the current boiler operating conditions and compares the real-time temperature field with the reference temperature field to form a residual abnormal field. This eliminates the interference of normal operating condition fluctuations such as load changes and combustion organization adjustments on the diagnostic results. By identifying residual abnormal connected domains and tracking their main propagation direction, path stability, expansion consistency, area expansion rate, and residual growth rate, the system can identify the formation and expansion trend of abnormal regions before the thermal deviation develops into a significant single-point overheating, improving the early fault detection capability. The system matches the abnormal propagation path with the mechanistic characteristics of combustion skew, fixed expansion anomalies, steam-water side flow distribution anomalies, and temperature measurement link anomalies, enabling the differentiation of different causes of thermal deviation and avoiding misjudging measuring point anomalies or short-term disturbances as real thermal deviation faults. Therefore, this invention can improve the online positioning accuracy, type identification accuracy, and diagnostic reliability of water-cooled wall thermal deviation faults, providing a more reliable basis for judgment in boiler combustion adjustment, heating surface maintenance, and operational risk control.

[0129] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for online diagnosis of thermal deviation faults in water-cooled walls, characterized in that, include: Temperature data at different spatial locations of the water-cooled wall are collected, and the water-cooled wall in the furnace is spatially expanded to construct a real-time temperature field of the water-cooled wall. Based on the current operating conditions of the boiler, a reference temperature field is generated that is adapted to the current load, combustion and steam-water conditions. The real-time temperature field of the water-cooled wall is compared and calculated with the operating condition reference temperature field to obtain the temperature residual field. Based on the residual characteristics, abnormal candidate nodes are screened and merged to form a spatial abnormal connected domain. Within a continuous time window, the spatially abnormal connected domains are tracked and matched to construct abnormal event trajectories, and the abnormal propagation paths, expansion characteristics, and residual change characteristic parameters are quantified. The quantitative characteristics of abnormal events are matched and compared with the preset multi-type thermal deviation fault mechanism characteristics to distinguish different thermal deviation fault cause types. Based on the feature matching results, the thermal deviation fault type is determined, and combined with the residual strength, abnormal range, and operating condition correlation characteristics, the fault location, fault level, and diagnostic reliability are output to complete the accurate online diagnosis of water-cooled wall thermal deviation faults.

2. The online diagnostic method for thermal deviation faults in water-cooled walls according to claim 1, characterized in that, The specific process of generating a reference temperature field that adapts to the current load, combustion, and steam-water conditions based on the boiler's current operating conditions is as follows: The historical normal operating samples of the boiler are pre-divided into multiple operating condition clusters; Each of the aforementioned operating condition clusters corresponds to a normal temperature distribution pattern of the water-cooled wall under stable operating conditions; By matching the current boiler operating conditions with the various operating condition clusters, a corresponding operating condition reference temperature field is generated.

3. The online diagnostic method for thermal deviation faults in water-cooled walls according to claim 2, characterized in that, The matching relationship between the current boiler operating condition and each operating condition cluster is specifically obtained by weighted calculation of the operating condition reference temperature field based on each operating condition cluster. The weight of each operating condition cluster is determined by the distance or similarity between the current boiler operating condition and the center of the corresponding operating condition cluster.

4. The online diagnostic method for thermal deviation faults in water-cooled walls according to claim 1, characterized in that, The selection criteria for identifying anomalous candidate nodes based on residual features are as follows: The positive residual exceeds the current operating condition residual threshold, and the residual gradient exceeds the gradient threshold, or multiple nodes in the neighborhood simultaneously exhibit abnormal positive residuals.

5. The online diagnostic method for thermal deviation faults in water-cooled walls according to claim 1, characterized in that, Within a continuous time window, the abnormal connected domain is tracked and matched, and the main propagation direction of the abnormal event is synthesized by weighting the migration vectors of the abnormal center in multiple consecutive time windows. When the number of effective migration vectors is insufficient or the main propagation direction judgment index is less than the lower limit of direction judgment, the abnormal event is transferred to the center-preserving abnormal judgment process.

6. The online diagnostic method for thermal deviation faults in water-cooled walls according to claim 5, characterized in that, Path stability is determined by the degree of consistency between the migration vector of the anomaly center and the main propagation direction of the anomaly event, and expansion consistency is determined by the degree of consistency between the major axis direction of the anomaly region and the main propagation direction of the anomaly event.

7. The online diagnostic method for thermal deviation faults in water-cooled walls according to claim 1, characterized in that, The various thermal deviation fault mechanism characteristics include the matching characteristics of combustion skew type thermal deviation, the matching characteristics of fixed extended type thermal deviation, and the matching characteristics of abnormal flow distribution on the steam and water side type thermal deviation.

8. The online diagnostic method for thermal deviation faults in water-cooled walls according to claim 7, characterized in that, The matching characteristics of combustion skew thermal deviation include: the consistency between the abnormal main propagation direction and the furnace circumference, flue gas rotation direction or flue gas deflection direction, the circumferential offset ratio of the abnormal area, and the lag correlation between combustion side operating condition disturbance and residual growth. Matching characteristics of fixed extended thermal deviations include: the retention degree of the abnormal region center, the rate of abnormal area expansion, the rate of residual growth, the duration of abnormality, and the degree of abnormality that can be explained by the non-combustion side and the steam-water side conditions. The matching characteristics of abnormal thermal deviation in the steam-water side flow distribution include: consistency between the abnormal main propagation direction or the long axis direction of the abnormal region and the direction of the tube screen, the direction of the parallel loop or the direction of steam-water working fluid flow; the increasing state of the residual along the steam-water flow direction; and the lag correlation between steam-water side operating condition disturbance and residual growth. The abnormal characteristics of the temperature measurement link include abnormal dispersion, short duration, insufficient neighborhood support, and no stable propagation path.

9. The online diagnostic method for thermal deviation faults in water-cooled walls according to claim 8, characterized in that, The specific process for determining the type of thermal deviation fault based on feature matching results includes: The matching characteristics of combustion skew thermal deviation, fixed extended thermal deviation, abnormal flow distribution of steam and water side thermal deviation, and abnormal temperature measurement link are calculated respectively. When the maximum matching value exceeds the corresponding judgment threshold, the corresponding fault type is output.

10. An online diagnostic system for thermal deviation faults in water-cooled walls, characterized in that, include: The online temperature field construction unit collects temperature data at different spatial locations of the water-cooled wall, performs spatial unfolding processing on the furnace water-cooled wall, and constructs the real-time temperature field of the water-cooled wall. The operating condition reference temperature field generation unit generates an operating condition reference temperature field that is adapted to the current load, combustion and steam-water conditions based on the current operating conditions of the boiler. The residual abnormal field analysis unit compares and calculates the real-time temperature field of the water-cooled wall with the operating condition reference temperature field to obtain the temperature residual field, filters abnormal candidate nodes based on residual characteristics, and merges them to form a spatial abnormal connected domain. The anomaly propagation path tracking unit tracks and matches the spatial anomaly connected domains within a continuous time window, constructs anomaly event trajectories, and quantifies the anomaly propagation path, expansion characteristics, and residual change characteristic parameters. The thermal deviation mechanism matching unit matches and compares the quantitative characteristics of abnormal events with the preset multi-type thermal deviation fault mechanism characteristics to distinguish different thermal deviation fault cause types. The thermal deviation fault diagnosis unit determines the type of thermal deviation fault based on feature matching results, and outputs the fault location, fault level and diagnostic reliability by combining residual strength, abnormal range and operating condition correlation characteristics, thus completing the accurate online diagnosis of thermal deviation faults in water-cooled walls.

Images