Intelligent monitoring and diagnosing method for running state of boiler induced draft fan in thermal power plant

Abstract

The application discloses an intelligent monitoring and diagnosing method for the operation state of a boiler induced draft fan in a thermal power plant, and relates to the technical field of power equipment monitoring. The method comprises the following steps: multi-modal heterogeneous data acquisition and space-time alignment, acoustic, temperature field and airflow signals are collected by optical fiber acoustic sensors, infrared thermal imaging arrays and the like and are correlated; improved adaptive multi-scale morphological decomposition is used to extract fault features, and feature separation is realized through dynamic adjustment of a structural element; fault mode matching and result output. The application breaks through the limitations of traditional single parameter monitoring, realizes early and accurate diagnosis of induced draft fan faults through multi-physical field fusion and adaptive algorithm optimization, and is suitable for intelligent operation and maintenance of key equipment in a thermal power plant.

Description

Technical Field

[0001] This invention relates to the field of equipment testing and diagnosis, specifically to an intelligent monitoring and diagnostic method for the operating status of induced draft fans in thermal power plant boilers. Background Technology

[0002] The induced draft fan of a thermal power plant boiler is a key piece of equipment for ensuring combustion efficiency and flue gas emissions; its operating status directly affects the safety and economy of the unit. Existing monitoring and diagnostic methods have the following shortcomings: Data acquisition is limited: it relies heavily on vibration acceleration sensors or current transmitters, only monitoring local vibration or energy consumption parameters, ignoring key fault precursors such as acoustic and thermal (e.g., abnormal noise from impeller wear, temperature field distortion due to bearing overheating). Limitations of feature extraction: Traditional methods use fixed-scale wavelet decomposition or Fourier transform, which cannot adaptively match the broadband fault features of the induced draft fan (such as the low-frequency pulsation of surge and the high-frequency impact of blade cracks), leading to feature confusion. Diagnostic lag: The accuracy rate of identifying early, minor faults (such as loose bearing housings or dust accumulation on impellers) is less than 60%, often leading to the expansion of faults due to missed diagnoses.

[0003] Therefore, there is an urgent need for an intelligent monitoring and diagnostic method that can perceive in multiple dimensions and adaptively extract features to achieve early and accurate identification of induced draft fan failures. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent monitoring and diagnostic method for the operating status of induced draft fans in thermal power plant boilers, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent monitoring and diagnostic method for the operating status of induced draft fans in thermal power plant boilers, comprising the following steps: S1: Multimodal heterogeneous data acquisition and spatiotemporal alignment: Acoustic, thermal, and airflow disturbance signals during the operation of the induced draft fan are acquired through a non-invasive composite sensor array, and spatiotemporal correlation is performed to obtain the acquired data; the specific steps are as follows: S1.1: Acoustic Signal: Six fiber optic acoustic sensors (sampling frequency 10kHz) are deployed circumferentially around the exhaust fan casing to collect broadband vibration acoustic signals. ,in For timestamps, This includes impeller rotation noise (50-500Hz) and fault impact noise (1-5kHz); temperature field distribution: the bearing housing area is photographed using an infrared thermal imaging array (resolution 320×240) to generate a temperature field matrix. ,in Spatial coordinates, unit: mm, sampling period: 1s; Airflow disturbance: Three sets of micro differential pressure sensors (spaced 120° apart) were installed in the inlet pipe of the induced draft fan to calculate the airflow disturbance coefficient. ,in This represents the standard deviation of the pressure difference. The average pressure difference reflects the intensity of airflow turbulence. S1.2: Time Alignment: Using the pulse signal of the encoder on the main shaft of the induced draft fan as the reference clock (synchronization error < 1ms), the acoustic signal (10kHz), temperature field (1Hz), and airflow disturbance (10Hz) are aligned to the same time axis. ; S1.3: Spatial Association: Establish the mapping relationship between sensor locations and the structure of the induced draft fan, including the first... Road sound sensor corresponding to impeller number Quadrant region, acoustic signal With the temperature of the area Binding, generating a spatiotemporal fusion matrix: in, There are 3 sets of airflow disturbance coefficients; S2: Fault feature extraction based on improved adaptive multi-scale morphological decomposition, adaptive decomposition and sensitive feature screening of the collected data; the specific steps are as follows: S2.1: Signal preprocessing: from the fusion matrix Extracting single-channel acoustic signals (Focusing on fault-sensitive channels), perform DC removal and normalization processing: ,in, for mean Standard deviation; S2.2: Improved AMMD algorithm decomposition: Initialization: Set minimum scale Maximum scale The unit is sampling points, and the initial shape of the structuring element is a flat cap shape. ; Calculate the kurtosis of the current signal segment ,in For the expectation operator, if (High-frequency impact characteristics), then reduce the scale. ;like (Low-frequency pulse impact characteristics), then increase the scale Combined with the typical failure frequency of induced draft fans (such as bearing failure) impeller failure Correction scale: ,in Sampling frequency, This is the frequency correction factor. It is a rounding function; Multiscale decomposition: for Perform opening-closing operations, decompose into each feature component and residual components : ,in For opening operation, For closing operations, For the first The structuring element of the characteristic component; S2.3: Sensitive Feature Screening: Calculate the fault sensitivity index of each component—corrected energy entropy. , in For the first The first characteristic component Energy percentage of each frequency band Preserve the kurtosis of this component. The component is used as a fault characteristic (normal state) ); S3: Fault mode matching with fused features. The filtered feature vectors are dynamically matched with the fault mode database, and the matching degree is calculated using cosine similarity. ,when The fault is identified as a corresponding fault, that is, the operating status and fault type of the induced draft fan are determined. S4: Diagnostic results output and early warning. Based on the structure obtained in step S3, generate an operational status assessment report and trigger graded early warnings.

[0006] Preferably, the non-invasive composite sensor array in step S1 includes a fiber optic acoustic sensor, an infrared thermal imaging array, and a micro differential pressure sensor, which respectively collect the vibration acoustic signal of the induced draft fan casing, the bearing temperature field distribution, and the inlet airflow disturbance coefficient.

[0007] Preferably, step S3 is implemented as follows: S3.1 Dynamic construction and maintenance of the fault mode library: Initial template construction: Collect full lifecycle data of 12 typical faults of induced draft fans, with no less than 50 samples for each fault type (covering the three stages of minor, moderate, and severe); perform steps S1 and S2 on each fault sample to select the corresponding sensitive feature components and construct a standard feature template vector for each fault type. This yields the fault mode library, in which... Number the fault type. The feature dimension is determined by the number of sensitive features after S2 filtering, and is fixed at 10 dimensions to ensure dimension uniformity. For the first Class of faults Normalized mean of dimensional features; S3.2: Extract the retained sensitive feature components from the output of step S2, i.e., those that satisfy... Construct the original feature vector to be matched ,in For the first The normalized energy entropy value of the dimensional sensitive feature. The dimension is consistent with the template vector of the fault mode library; for the fluctuations in feature amplitude under different operating conditions, the following applies: Secondary standardization is performed to eliminate the impact of differences in operating conditions: ,in These are the minimum and maximum values ​​of the original vector to be matched, respectively, after standardization. This yields the standardized vector to be matched. ; S3.3: Cosine Similarity Calculation: Cosine similarity is used to measure the similarity between vectors to be matched. Compared with a certain type of fault template vector in the pattern library The directional consistency is such that the closer the value is to 1, the higher the feature matching degree. The calculation formula is as follows: ,in For the vector to be matched and the first Matching degree of the fault template, It is the dot product of two vectors; Let L2 norms be the L2 norms of the two vectors, respectively. For the standardized matching vector, the first... Dimensional components; For the first The first class of fault template vectors Dimensional components; Traverse the template vectors of 12 fault types in the pattern library and calculate the matching degree between the vector to be matched and each fault type. Record the maximum matching degree and corresponding fault types ; S3.4: Dynamic Matching Verification and Fault Judgment: Timing consistency verification: The fault characteristics of the induced draft fan are continuous. The high matching degree of a single sampling may be caused by random interference. Therefore, continuous sampling is necessary. Each timestamp corresponds to the sampling frequency in step S1, and the vector to be matched has a total duration of 0.5 seconds. Calculate the relationship between each vector and... Matching degree of class of faults If continuous Each sampling point satisfies If the condition is met, proceed to the next step; if there are fewer than 3 sampling points that meet the requirements, it is determined to be random interference, and "No fault at present" is output. Adaptive threshold adjustment: To address the differences in feature recognition for different fault types, a threshold correction coefficient is introduced. For faults with high identifiable characteristics (such as surge, characterized by low-frequency strong pulsations), ), correct the threshold to For faults with low feature recognition (such as slight impeller dust accumulation), ), correct the threshold to The correction threshold is obtained by statistically analyzing the similarity distribution of this type of fault samples in the pattern library (ensuring that 95% of real fault samples can meet the correction threshold). Final determination: If the corrected matching degree If three or more consecutive sampling points are detected, the induced draft fan is ultimately determined to be faulty. Class of faults; if the maximum matching degree of all fault types If so, it is determined as "there is no clear fault at present, but continuous monitoring is required" (it may be a very early fault, and the characteristics have not been fully manifested). S3.5: Matching Result Output and Confidence Calculation: Output fault determination results, including fault type (e.g., "impeller 3-quadrant wear"), maximum matching degree. The number of sampling points that continuously meet the threshold Then, the confidence level for fault determination was calculated. This confidence level will serve as the basis for the graded early warning in step S4.

[0008] Preferably, the 12 typical faults in step S3.1 include: impeller wear, bearing overheating, surge, blade cracks, loose bearing housing, impeller dust accumulation, motor imbalance, coupling misalignment, seal leakage, air passage blockage, main shaft bending, and electrical faults.

[0009] Preferably, the fault mode library obtained in step S3.1 is equipped with a dynamic update mechanism, specifically: the fault mode library is iteratively updated every 100 sets of actual operating data, including normal states and confirmed fault states; specifically, if the cosine similarity between a newly collected fault sample and an existing template vector is... If it is determined to be a new fault variant, then it is used as a supplementary template for this type of fault, and the average feature vector of this type of fault is recalculated. ,in The number of existing templates for this type of fault. This serves as the feature vector for new fault samples, ensuring that the fault mode library adapts to feature drift caused by equipment aging and changes in operating conditions.

[0010] Preferably, the specific working logic of step S4 is as follows: S4.1: Extract the core information from the output of step S3: fault type (e.g., "impeller wear", "bearing overheating"), determination confidence level (e.g., 0.736), the source channel of sensitive features, that is, the sensor number corresponding to the sensitive features selected in step S2; then, based on the sensor spatial mapping relationship in step S1, the accurate anchoring of the fault location is achieved; Integrated output information: The fault type, confidence level, and precise location are integrated into a standardized diagnostic result: "Fault type: Fault location: Location result; Confidence level: (Example: "Fault type: impeller wear; Fault location: impeller 3 quadrant; Confidence level: 92%"); S4.2: Two-dimensional classification of fault levels: Definition of partitioning dimensions: Dimension 1: Confidence level of judgment ( Dimension 1: Reflects the reliability of the diagnostic results; Dimension 2: Inherent risk level of the fault. ,in Based on the preset values ​​and the degree of impact of the fault on the operation of the induced draft fan, it is divided into 3 levels: High risk: surge, blade cracks, spindle bending, electrical faults (which may lead to equipment downtime or safety accidents); Medium risk: bearing overheating, coupling misalignment, motor imbalance (long-term operation leads to the expansion of the fault and affects the unit efficiency); Low risk: Dust accumulation on the impeller, minor seal leakage, minor blockage of the air passage (no serious impact in the short term, can be operated with the fault until the maintenance window); Then, combining Dimension 1 and Dimension 2, the faults are divided into four warning levels: Level 1 warning, Level 2 warning, Level 3 warning, and Level 4 warning. S4.3: Hierarchical Early Warning Execution and Linkage Control: Early warning hardware linkage: Trigger corresponding preset hardware actions according to the fault level to ensure the timeliness and effectiveness of early warning; Warning cancellation conditions: After manual confirmation of maintenance completion, a "maintenance completed" command is sent through the operation and maintenance system; 10 sampling points are continuously collected (total duration 1 second), and if the fault matching degree of all sampling points is... And confidence level If the fault symptoms are not eliminated after maintenance, the warning will be automatically lifted; If so, the S2-S4 process will be re-executed to update the diagnostic results and warning level; S4.4: Result Recording: Standardized diagnostic results, warning level, trigger time, release time (if any), and original sensor data fragments (10s before and after the fault occurred) are stored in a distributed database and archived according to "device number-date-fault type" for subsequent fault tracing and pattern library updates.

[0011] Preferably, the precise anchoring of the fault location in step S4.1 specifically includes the following: If the sensitive feature originates from the acoustic sensor: 6 acoustic sensors are deployed in quadrants 1-6 of the circumference of the induced draft fan impeller (quadrant 1 corresponds to 0°-60°, quadrant 2 to 60°-120°, and so on), sensor numbered... Direct mapping impeller quadrant That is, "sound sensor No. 3" corresponds to "impeller 3 quadrant"; If the sensitive feature originates from an infrared thermal imaging array: based on the temperature field matrix Coordinates corresponding to medium and high energy entropy features Based on the physical dimensions of the bearing housing, i.e., the preset center coordinates of the bearing housing ), calculate fault offset , Locate the specific fault point; If the sensitive feature comes from the airflow disturbance sensor: the 3 sets of sensors correspond to the 0°, 120° and 240° positions of the inlet pipe, and the sensor number corresponds to the pipe position, and the location is "abnormal airflow disturbance at the 120° position of the inlet pipe" (such as airway blockage or seal leakage).

[0012] Compared with existing technologies, the beneficial effects of this invention are: Multi-dimensional perception enhances data effectiveness: This invention collects data through an "acoustic-thermal-airflow" multi-modal sensor array, breaking through the limitations of traditional single-parameter monitoring. Simultaneously, spatiotemporal alignment processing (synchronization error <1ms) ensures data integrity and timeliness. Compared with traditional vibration monitoring, the addition of dimensions such as temperature field gradient and airflow disturbance coefficient makes fault precursor detection more comprehensive, increasing data utilization by over 60%, and providing a reliable foundation for subsequent diagnosis.

[0013] Adaptive feature extraction enhances fault identification: The improved AMMD algorithm dynamically adjusts the scale of structural elements based on kurtosis and fault frequency, solving the problem that traditional fixed-scale algorithms struggle to separate broadband fault features. It improves the feature separation accuracy for different types of faults, such as blade cracks (high-frequency impact) and surge (low-frequency pulsation), and enhances the accuracy of identifying early, minor faults (such as slight impeller dust accumulation or loose bearing housings), effectively preventing the expansion of faults due to missed diagnoses.

[0014] Precise matching and location reduce diagnostic errors: A dynamically updated fault mode library adapts to feature drift caused by equipment aging and changes in operating conditions. Combined with cosine similarity calculation and temporal consistency verification (confirmation through three consecutive sampling points), the fault misjudgment rate is reduced to below 5%. At the same time, precise fault location is achieved based on sensor spatial mapping relationships (such as impeller quadrant, bearing housing coordinates, and pipeline orientation), providing clear targets for operation and maintenance and shortening fault troubleshooting time.

[0015] Tiered early warning and operation and maintenance linkage improve management efficiency: adopting a two-dimensional tiered early warning system of "confidence level + inherent risk" to achieve differentiated intervention (emergency shutdown, time-limited maintenance, and regular investigation) and avoid over-response or delayed processing.

[0016] Intelligent operation and maintenance drives industrial upgrading: This invention integrates technologies such as multi-physics field data fusion, adaptive algorithm optimization, intelligent matching and operation and maintenance linkage, breaking away from the traditional monitoring mode that relies on manual experience, and promoting the intelligent transformation of thermal power plant induced draft fan operation and maintenance towards "data-driven + precise intervention". Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall method flow of the present invention; Figure 2 This is a schematic diagram of the multimodal heterogeneous data acquisition and spatiotemporal alignment workflow of the present invention; Figure 3 This is a schematic diagram illustrating the adaptive decomposition and sensitive feature screening workflow of the collected data in this invention. Figure 4 This is a schematic diagram illustrating the workflow for determining the operating status and fault type of the induced draft fan according to the present invention. Figure 5 This is a schematic diagram of the diagnostic result output and early warning workflow of the present invention. Detailed Implementation

[0018] 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.

[0019] Please see Figure 1-5 This invention provides a technical solution: an intelligent monitoring and diagnostic method for the operating status of induced draft fans in thermal power plant boilers, comprising the following steps: S1: Multimodal heterogeneous data acquisition and spatiotemporal alignment: Acoustic, thermal, and airflow disturbance signals during the operation of the induced draft fan are acquired using a non-invasive composite sensor array, and spatiotemporal correlation is performed to obtain the acquired data. The non-invasive composite sensor array includes a fiber optic acoustic sensor, an infrared thermal imaging array, and a micro-differential pressure sensor, which respectively acquire the vibration acoustic signals of the induced draft fan casing, the bearing temperature field distribution, and the inlet airflow disturbance coefficient. Specific steps are as follows: S1.1: Acoustic Signal: Six fiber optic acoustic sensors (sampling frequency 10kHz) are deployed circumferentially around the exhaust fan casing to collect broadband vibration acoustic signals. ,in For timestamps, This includes impeller rotation noise (50-500Hz) and fault impact noise (1-5kHz); temperature field distribution: the bearing housing area is photographed using an infrared thermal imaging array (resolution 320×240) to generate a temperature field matrix. ,in Spatial coordinates, unit: mm, sampling period: 1s; Airflow disturbance: Three sets of micro differential pressure sensors (spaced 120° apart) were installed in the inlet pipe of the induced draft fan to calculate the airflow disturbance coefficient. ,in This represents the standard deviation of the pressure difference. The average pressure difference reflects the intensity of airflow turbulence. S1.2: Time Alignment: Using the pulse signal of the encoder on the main shaft of the induced draft fan as the reference clock (synchronization error < 1ms), the acoustic signal (10kHz), temperature field (1Hz), and airflow disturbance (10Hz) are aligned to the same time axis. ; S1.3: Spatial Association: Establish the mapping relationship between sensor locations and the structure of the induced draft fan, including the first... Road sound sensor corresponding to impeller number Quadrant region, acoustic signal With the temperature of the area Binding, generating a spatiotemporal fusion matrix: in, There are 3 sets of airflow disturbance coefficients; S2: Fault feature extraction based on improved adaptive multi-scale morphological decomposition, adaptive decomposition and sensitive feature screening of the collected data; the specific steps are as follows: S2.1: Signal preprocessing: from the fusion matrix Extracting single-channel acoustic signals (Focusing on fault-sensitive channels), perform DC removal and normalization processing: ,in, for mean Standard deviation; S2.2: Improved AMMD algorithm decomposition: Initialization: Set minimum scale Maximum scale The unit is sampling points, and the initial shape of the structuring element is a flat cap shape. ; Calculate the kurtosis of the current signal segment ,in For the expectation operator, if (High-frequency impact characteristics), then reduce the scale. ;like (Low-frequency pulse impact characteristics), then increase the scale Combined with the typical failure frequency of induced draft fans (such as bearing failure) impeller failure Correction scale: ,in Sampling frequency, This is the frequency correction factor. It is a rounding function; Multiscale decomposition: for Perform opening-closing operations, decompose into each feature component and residual components : ,in For opening operation, For closing operations, For the first The structuring element of the characteristic component; S2.3: Sensitive Feature Screening: Calculate the fault sensitivity index of each component—corrected energy entropy. , in For the first The first characteristic component Energy percentage of each frequency band Preserve the kurtosis of this component. The component is used as a fault characteristic (normal state) ); S3: Fault mode matching with fused features. The filtered feature vectors are dynamically matched with the fault mode database, and the matching degree is calculated using cosine similarity. ,when The fault is identified as a corresponding fault, that is, the operating status and fault type of the induced draft fan are determined; the specific implementation steps of step S3 are as follows: S3.1 Dynamic construction and maintenance of the fault mode library: Initial template construction: Collect full lifecycle data of 12 typical faults of induced draft fans, including: impeller wear, bearing overheating, surge, blade cracks, bearing housing loosening, impeller dust accumulation, motor imbalance, coupling misalignment, seal leakage, air passage blockage, main shaft bending, and electrical faults. Collect no less than 50 samples for each fault type (covering three stages: minor, moderate, and severe). Perform steps S1 and S2 on each fault sample to select the corresponding sensitive feature components and construct a standard feature template vector for each fault type. This yields the fault mode library, in which... Number the fault type. The feature dimension is determined by the number of sensitive features after S2 filtering, and is fixed at 10 dimensions to ensure dimension uniformity. For the first Class of faults Normalized mean of dimensional features; The fault mode library is dynamically updated. Specifically, it is iteratively updated every 100 sets of actual operational data, including both normal and confirmed fault states. Specifically, if the cosine similarity between a newly collected fault sample and an existing template vector is... If it is determined to be a new fault variant, then it is used as a supplementary template for this type of fault, and the average feature vector of this type of fault is recalculated. ,in The number of existing templates for this type of fault. This serves as the feature vector for new fault samples, ensuring that the fault mode library adapts to feature drift caused by equipment aging and changes in operating conditions. S3.2: Extract the retained sensitive feature components from the output of step S2, i.e., those that satisfy... Construct the original feature vector to be matched ,in For the first The normalized energy entropy value of the dimensional sensitive feature. The dimension is consistent with the template vector of the fault mode library; for the fluctuations in feature amplitude under different operating conditions, the following applies: Secondary standardization is performed to eliminate the impact of differences in operating conditions: ,in These are the minimum and maximum values ​​of the original vector to be matched, respectively, after standardization. This yields the standardized vector to be matched. ; S3.3: Cosine Similarity Calculation: Cosine similarity is used to measure the similarity between vectors to be matched. Compared with a certain type of fault template vector in the pattern library The directional consistency is such that the closer the value is to 1, the higher the feature matching degree. The calculation formula is as follows: ,in For the vector to be matched and the first Matching degree of the fault template, It is the dot product of two vectors; Let L2 norms be the L2 norms of the two vectors, respectively. For the standardized matching vector, the first... Dimensional components; For the first The first class of fault template vectors Dimensional components; Traverse the template vectors of 12 fault types in the pattern library and calculate the matching degree between the vector to be matched and each fault type. Record the maximum matching degree and corresponding fault types ; S3.4: Dynamic Matching Verification and Fault Judgment: Timing consistency verification: The fault characteristics of the induced draft fan are continuous. The high matching degree of a single sampling may be caused by random interference. Therefore, continuous sampling is necessary. Each timestamp corresponds to the sampling frequency in step S1, and the vector to be matched has a total duration of 0.5 seconds. Calculate the relationship between each vector and... Matching degree of class of faults If continuous Each sampling point satisfies If the condition is met, proceed to the next step; if there are fewer than 3 sampling points that meet the requirements, it is determined to be random interference, and "No fault at present" is output. Adaptive threshold adjustment: To address the differences in feature recognition for different fault types, a threshold correction coefficient is introduced. For faults with high identifiable characteristics (such as surge, characterized by low-frequency strong pulsations), ), correct the threshold to For faults with low feature recognition (such as slight impeller dust accumulation), ), correct the threshold to The correction threshold is obtained by statistically analyzing the similarity distribution of this type of fault samples in the pattern library (ensuring that 95% of real fault samples can meet the correction threshold). Final determination: If the corrected matching degree If three or more consecutive sampling points are detected, the induced draft fan is ultimately determined to be faulty. Class of faults; if the maximum matching degree of all fault types If so, it is determined as "there is no clear fault at present, but continuous monitoring is required" (it may be a very early fault, and the characteristics have not been fully manifested). S3.5: Matching Result Output and Confidence Calculation: Output fault determination results, including fault type (e.g., "impeller 3-quadrant wear"), maximum matching degree. The number of sampling points that continuously meet the threshold Then, the confidence level for fault determination was calculated. This confidence level will serve as the basis for the graded early warning in step S4.

[0020] S4: Diagnostic Result Output and Early Warning. Based on the structure obtained in step S3, an operational status assessment report is generated and a graded early warning is triggered. The specific working logic of step S4 is as follows: S4.1: Extract the core information from the output of step S3: fault type (e.g., "impeller wear", "bearing overheating"), determination confidence level (e.g., 0.736), the sensitive feature source channel, i.e., the sensor number corresponding to the sensitive feature selected in step S2; then, based on the sensor spatial mapping relationship in step S1, the accurate anchoring of the fault location is achieved, including the following: If the sensitive feature originates from the acoustic sensor: 6 acoustic sensors are deployed in quadrants 1-6 of the circumference of the induced draft fan impeller (quadrant 1 corresponds to 0°-60°, quadrant 2 to 60°-120°, and so on), sensor numbered... Direct mapping impeller quadrant That is, "sound sensor No. 3" corresponds to "impeller 3 quadrant"; If the sensitive feature originates from an infrared thermal imaging array: based on the temperature field matrix Coordinates corresponding to medium and high energy entropy features Based on the physical dimensions of the bearing housing, i.e., the preset center coordinates of the bearing housing ), calculate fault offset , Locate the specific fault point; If the sensitive feature comes from the airflow disturbance sensor: the 3 sets of sensors correspond to the 0°, 120° and 240° positions of the inlet pipe, and the sensor number corresponds to the pipe position, and the location is "abnormal airflow disturbance at the 120° position of the inlet pipe" (such as airway blockage or seal leakage). Integrated output information: The fault type, confidence level, and precise location are integrated into a standardized diagnostic result: "Fault type: Fault location: Location result; Confidence level: (Example: "Fault type: impeller wear; Fault location: impeller 3 quadrant; Confidence level: 92%"); S4.2: Two-dimensional classification of fault levels: Definition of partitioning dimensions: Dimension 1: Confidence level of judgment ( Dimension 1: Reflects the reliability of the diagnostic results; Dimension 2: Inherent risk level of the fault. ,in Based on the preset values ​​and the degree of impact of the fault on the operation of the induced draft fan, it is divided into 3 levels: High risk: surge, blade cracks, spindle bending, electrical faults (which may lead to equipment downtime or safety accidents); Medium risk: bearing overheating, coupling misalignment, motor imbalance (long-term operation leads to the expansion of the fault and affects the unit efficiency); Low risk: Dust accumulation on the impeller, minor seal leakage, minor blockage of the air passage (no serious impact in the short term, can be operated with the fault until the maintenance window); Then, combining Dimension 1 and Dimension 2, the faults are divided into four warning levels: Level 1, Level 2, Level 3, and Level 4. The specific classification table is shown below: Level definition explanation: Level 1 Warning: Emergency malfunction, immediate shutdown required; Level 2 warning: Critical fault, maintenance will be arranged within 1 hour; Level 3 warning: General fault, to be investigated within 24 hours; Level 4 warning: Minor fault, to be addressed during the next scheduled maintenance; Continuous monitoring: Early-stage fault characteristics are identified, and diagnostic results are updated every 5 minutes until the characteristics become more pronounced or disappear.

[0021] S4.3: Hierarchical Early Warning Execution and Linkage Control: Early warning hardware linkage: Triggers corresponding preset hardware actions based on the fault level to ensure the timeliness and effectiveness of early warnings; for example: Level 1 warning (e.g., "blade crack, confidence level 0.95"): Activate the audible and visual alarm (red warning light stays on + high-frequency buzzer, volume ≥ 85dB); Send a shutdown command (e.g., command format: "STOP_REQUEST=1") to the induced draft fan PLC control system to trigger the emergency shutdown procedure (according to the unit safety regulations, first reduce the load of the induced draft fan to 0, and then cut off the power supply). Send SMS and APP push notifications (including fault information, location, and emergency handling suggestions) to the operation and maintenance center via industrial Ethernet.

[0022] Level 2 warning (e.g., "bearing overheating, confidence level 0.88"): Activate the audible and visual alarm (yellow warning light flashing + medium frequency buzzer, volume ≥75dB); Send a load limit command ("LOAD_LIMIT=50%) to the PLC to prevent the fault from worsening; The Operations and Maintenance Center APP pushes notifications and simultaneously generates maintenance work orders (including fault location and a list of required tools).

[0023] Level 3 warning (e.g., "Dust accumulation on the impeller, confidence level 0.82"): The green warning light is flashing, but there is no buzzer alarm. The maintenance system generates to-do reminders, which are displayed on the monitoring interface in the central control room. Level 4 warning (e.g., "minor seal leak, confidence level 0.75"): There are no hardware alarms; fault information is only displayed on the monitoring interface. Include it in the regular maintenance plan and automatically associate it with the next maintenance time.

[0024] Warning cancellation conditions: After manual confirmation of maintenance completion, a "maintenance completed" command is sent through the operation and maintenance system; 10 sampling points are continuously collected (total duration 1 second), and if the fault matching degree of all sampling points is... And confidence level If the fault symptoms are not eliminated after maintenance, the warning will be automatically lifted; If so, the S2-S4 process will be re-executed to update the diagnostic results and warning level; S4.4: Result Recording: Standardized diagnostic results, warning level, trigger time, release time (if any), and original sensor data fragments (10s before and after the fault occurred) are stored in a distributed database and archived according to "device number-date-fault type" for subsequent fault tracing and pattern library updates.

[0025] The following is an extension of the present invention, which implements a linkage feedback mechanism for the operation and maintenance structure: Automatically synchronize fault information to the thermal power plant's CMMS (equipment management system) and associate it with the induced draft fan's equipment file (such as runtime and historical maintenance records). Based on historical data statistics, generate "fault-maintenance" related suggestions for similar faults (such as "impeller 3 quadrant wear, historical maintenance plan: replace impeller blades, average maintenance time 2 hours"). If the same fault occurs ≥2 times within 3 months, a pattern library optimization request will be automatically triggered, prompting operations and maintenance personnel to supplement fault sample data (which can be used for iterative updates of the S3 pattern library).

[0026] This invention discloses an intelligent monitoring and diagnostic method for the operating status of induced draft fans in thermal power plant boilers, belonging to the field of power equipment monitoring technology. It aims to address the shortcomings of traditional monitoring methods, such as single parameters, feature confusion, and diagnostic lag. The method includes: S1 Multimodal heterogeneous data acquisition and spatiotemporal alignment, using fiber optic acoustic sensors, infrared thermal imaging arrays, and micro-differential pressure sensors to collect and correlate acoustic, temperature field, and airflow disturbance signals, generating a spatiotemporal fusion matrix (synchronization error <1ms); S2 Extracting fault features based on improved adaptive multi-scale morphological decomposition (AMMD), dynamically adjusting the scale of structural elements by kurtosis and fault frequency to separate broadband fault features and filter sensitive components (retaining components with corrected energy entropy >0.6); S3 Matching the fused features with a dynamically updated fault mode library, combining cosine similarity calculation and temporal consistency verification to accurately determine the fault type (false positive rate <5%); S4 Based on a dual-dimensional hierarchical early warning system of "confidence + inherent risk," linking hardware intervention and the operation and maintenance system to achieve targeted fault handling. This invention improves the accuracy of early minor fault identification from <60% to over 60%, extends the mean fault-free operation time of induced draft fans, reduces unplanned downtime losses, and is suitable for intelligent operation and maintenance of key equipment in thermal power plants, taking into account both safety and economy.

[0027] 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 method for intelligent monitoring and diagnosis of the operating status of induced draft fans in thermal power plant boilers, characterized in that, Includes the following steps: S1: Multimodal heterogeneous data acquisition and spatiotemporal alignment: Acoustic, thermal, and airflow disturbance signals during the operation of the induced draft fan are acquired through a non-invasive composite sensor array, and spatiotemporal correlation is performed to obtain the acquired data; the specific steps are as follows: S1.1: Multimodal signal acquisition: Acoustic signals: Six fiber optic acoustic sensors are deployed circumferentially around the fan casing to acquire broadband vibration acoustic signals. ,in For timestamps, This includes impeller rotation noise and fault impact noise; temperature field distribution: the bearing housing area is photographed using an infrared thermal imaging array to generate a temperature field matrix. ,in Spatial coordinates, unit: mm, sampling period: 1 s; Airflow disturbance: Three sets of micro differential pressure sensors were installed in the inlet pipe of the induced draft fan to calculate the airflow disturbance coefficient. ,in This represents the standard deviation of the pressure difference. The average pressure difference reflects the intensity of airflow turbulence. S1.2: Time Alignment: Using the pulse signal of the encoder on the main shaft of the induced draft fan as the reference clock, the acoustic signal, temperature field, and airflow disturbance are aligned to the same time axis. ; S1.3: Spatial Association: Establish the mapping relationship between sensor locations and the structure of the induced draft fan, including the first... Road sound sensor corresponding to impeller number Quadrant region, acoustic signal With the temperature of the area Binding, generating a spatiotemporal fusion matrix: in, There are 3 sets of airflow disturbance coefficients; S2: Fault feature extraction based on improved adaptive multi-scale morphological decomposition, performing adaptive decomposition and sensitive feature screening on the collected data; the specific steps are as follows: S2.1: Signal preprocessing: from the fusion matrix Extracting single-channel acoustic signals Perform DC removal and normalization processing: ,in, for mean Standard deviation; S2.2: Improved AMMD algorithm decomposition: Initialization: Set minimum scale Maximum scale The unit is: sampling points, and the initial shape of the structuring element is a flat cap shape. ; Calculate the kurtosis of the current signal segment ,in For the expectation operator, if Then reduce the scale ;like Then increase the scale Combined with the typical failure frequency of induced draft fans Correction scale: ,in Sampling frequency, This is the frequency correction factor. It is a rounding function; Multiscale decomposition: for Perform opening-closing operations, decompose into each feature component and residual components : ,in For opening operation, For closing operations, For the first The structuring element of the characteristic component; S2.3: Sensitive Feature Screening: Calculate the fault sensitivity index of each component—corrected energy entropy. , in For the first The first characteristic component Energy percentage of each frequency band Preserve the kurtosis of this component. The components are used as fault characteristics; S3: Fault mode matching with fused features. The filtered feature vectors are dynamically matched with the fault mode database, and the matching degree is calculated using cosine similarity. ,when The fault is identified as a corresponding fault, that is, the operating status and fault type of the induced draft fan are determined. S4: Diagnostic results output and early warning. Based on the structure obtained in step S3, generate an operational status assessment report and trigger graded early warnings.

2. The intelligent monitoring and diagnosis method for the operating status of induced draft fans in thermal power plant boilers according to claim 1, characterized in that: The non-invasive composite sensor array mentioned in step S1 includes a fiber optic acoustic sensor, an infrared thermal imaging array, and a micro differential pressure sensor, which respectively collect the vibration acoustic signal of the induced draft fan casing, the bearing temperature field distribution, and the inlet airflow disturbance coefficient.

3. The intelligent monitoring and diagnosis method for the operating status of induced draft fans in thermal power plant boilers according to claim 1, characterized in that: The specific implementation steps of step S3 are as follows: S3.1 Dynamic construction and maintenance of the fault mode library: Initial template construction: Collect full lifecycle data of 12 typical faults of the induced draft fan, with no less than 50 samples collected for each fault type; perform steps S1 and S2 on each fault sample to filter out the corresponding sensitive feature components and construct a standard feature template vector for each fault type. This yields the fault mode library, in which... Number the fault type. The feature dimension is determined by the number of sensitive features after S2 filtering, and is fixed at 10 dimensions to ensure dimension uniformity. For the first Class of faults Normalized mean of dimensional features; S3.2: Extract the retained sensitive feature components from the output of step S2, i.e., those that satisfy... Construct the original feature vector to be matched ,in For the first The normalized energy entropy value of the dimensional sensitive feature. The dimension is consistent with the template vector of the fault mode library; for the fluctuations in feature amplitude under different operating conditions, the following applies: Secondary standardization is performed to eliminate the impact of differences in operating conditions: ,in These are the minimum and maximum values ​​of the original vector to be matched, respectively, after standardization. This yields the standardized vector to be matched. ; S3.3: Cosine Similarity Calculation: Cosine similarity is used to measure the similarity between vectors to be matched. Compared with a certain type of fault template vector in the pattern library The directional consistency is such that the closer the value is to 1, the higher the feature matching degree. The calculation formula is as follows: ,in For the vector to be matched and the first Matching degree of the fault template, It is the dot product of two vectors; Let L2 norms be the L2 norms of the two vectors, respectively. For the standardized matching vector, the first... Dimensional components; For the first The first class of fault template vectors Dimensional components; Traverse the template vectors of 12 fault types in the pattern library and calculate the matching degree between the vector to be matched and each fault type. Record the maximum matching degree and corresponding fault types ; S3.4: Dynamic Matching Verification and Fault Judgment Timing consistency verification: The fault characteristics of the induced draft fan are continuous. The high matching degree of a single sampling may be caused by random interference. Therefore, continuous sampling is necessary. Each timestamp corresponds to the sampling frequency in step S1, and the vector to be matched has a total duration of 0.5 seconds. Calculate the relationship between each vector and Matching degree of class of faults ; If continuous Each sampling point satisfies If the condition is met, proceed to the next step; if there are fewer than 3 sampling points that meet the requirements, it is determined to be accidental interference, and "No fault at present" is output. Adaptive threshold adjustment: To address the differences in feature recognition for different fault types, a threshold correction coefficient is introduced. ; For faults with high feature recognition, the correction threshold is: ; For faults with low feature recognition, the correction threshold is: ; The correction threshold is obtained by statistically analyzing the similarity distribution of this type of fault samples in the pattern library; Final determination: If the corrected matching degree If three or more consecutive sampling points are detected, the induced draft fan is ultimately determined to be faulty. Class of faults; if the maximum matching degree of all fault types If so, it is determined that "there is no clear fault at present, but continuous monitoring is required"; S3.5: Matching Result Output and Confidence Calculation: Output fault determination results, including fault type and maximum matching degree. The number of sampling points that continuously meet the threshold Then, the confidence level for fault determination was calculated. This confidence level will serve as the basis for the graded early warning in step S4.

4. The intelligent monitoring and diagnosis method for the operating status of induced draft fans in thermal power plant boilers according to claim 3, characterized in that: The 12 typical faults in step S3.1 include: impeller wear, bearing overheating, surge, blade cracks, loose bearing housing, impeller dust accumulation, motor imbalance, coupling misalignment, seal leakage, air passage blockage, main shaft bending, and electrical faults.

5. The intelligent monitoring and diagnosis method for the operating status of induced draft fans in thermal power plant boilers according to claim 3, characterized in that: The fault mode library obtained in step S3.1 is equipped with a dynamic update mechanism, specifically: every 100 sets of actual operating data, including normal states and confirmed fault states, the fault mode library is iteratively updated; specifically, if the cosine similarity between a newly collected fault sample and an existing template vector is... If a new fault variant is identified, it is used as a supplementary template for that type of fault, and the average feature vector of that type of fault is recalculated. ,in The number of existing templates for this type of fault. This serves as the feature vector for new fault samples, ensuring that the fault mode library adapts to feature drift caused by equipment aging and changes in operating conditions.

6. The intelligent monitoring and diagnosis method for the operating status of induced draft fans in thermal power plant boilers according to claim 3, characterized in that: The specific working logic of step S4 is as follows: S4.1: Extract the core information from the output of step S3: fault type Determine the confidence level The source channel of sensitive features is the sensor number corresponding to the sensitive features selected in step S2; then, based on the sensor spatial mapping relationship in step S1, the accurate anchoring of the fault location is achieved. Integrated output information: Fault type, confidence level, and precise location are integrated into a standardized diagnostic result: "Fault type: Fault location: Location results; Confidence level: C×100%; S4.2: Two-dimensional classification of fault levels: Definition of partitioning dimensions: Dimension 1: Confidence level of judgment Dimension 1: Reflects the reliability of the diagnostic results; Dimension 2: Inherent risk level of the fault ,in Based on the preset values ​​and the degree of impact of the fault on the operation of the induced draft fan, it is divided into 3 levels: High-risk: surge, blade cracks, spindle bending, electrical failure; Medium risk: bearing overheating, coupling misalignment, motor imbalance; Low risk: Dust accumulation on the impeller, minor seal leakage, minor blockage of the air passage; Then, combining Dimension 1 and Dimension 2, the faults are divided into four warning levels: Level 1 warning, Level 2 warning, Level 3 warning, and Level 4 warning. S4.3: Hierarchical Early Warning Execution and Linkage Control: Early warning hardware linkage: Trigger corresponding preset hardware actions according to the fault level to ensure the timeliness and effectiveness of early warning; Warning cancellation conditions: After manual confirmation of maintenance completion, a "maintenance complete" command is sent through the operation and maintenance system; 10 sampling points are continuously collected, and the fault matching degree of all sampling points is... And confidence level If the fault symptoms are not eliminated after maintenance, the warning will be automatically lifted; If so, the S2-S4 process will be re-executed to update the diagnostic results and warning level; S4.4: Result Recording: Standardized diagnostic results, warning levels, trigger times, release times, and raw sensor data fragments are stored in a distributed database and archived according to "device number-date-fault type" for subsequent fault tracing and pattern library updates.

7. The intelligent monitoring and diagnosis method for the operating status of induced draft fans in thermal power plant boilers according to claim 6, characterized in that: The precise anchoring of the fault location in step S4.1 specifically includes the following: If the sensitive feature originates from the acoustic sensor: 6 acoustic sensors are deployed in quadrants 1-6 of the circumferential direction of the induced draft fan impeller, and the sensor numbers are as follows: Direct mapping impeller quadrant ; If the sensitive feature originates from an infrared thermal imaging array: based on the temperature field matrix Coordinates corresponding to medium and high energy entropy features Based on the physical dimensions of the bearing housing, i.e., the preset center coordinates of the bearing housing Calculate the fault offset , Locate the specific fault point; If the sensitive feature comes from the airflow disturbance sensor: 3 sets of sensors correspond to the 0°, 120° and 240° orientation of the inlet pipe, and the sensor number corresponds to the pipe orientation.

Images