A multi-dimensional flue gas desulfurization state monitoring method and system

Abstract

This application provides a multi-dimensional flue gas desulfurization status monitoring method and system, relating to the field of intelligent monitoring technology. The method includes acquiring historical process parameters and corresponding fault type labels under historical fault conditions of the desulfurization system; establishing a process fault diagnosis model through multi-parameter joint feature extraction and association training; and continuously collecting status monitoring signals from the desulfurization target equipment. Multi-type diagnostic maps are generated through map analysis and matched with an equipment fault feature knowledge base to identify equipment faults. When a process or equipment fault exists, the method combines a process-equipment fault association mapping table to retrieve corresponding expected abnormal features within the same time window for verification and matching. Finally, the fault level and corresponding adjustment plan are determined based on the matching results. This method can achieve comprehensive identification and predictive control of desulfurization system faults.

Description

Technical Field

[0001] This application relates to the field of intelligent monitoring technology, and in particular to a multi-dimensional flue gas desulfurization status monitoring method and system. Background Technology

[0002] Limestone-gypsum wet desulfurization is currently the mainstream desulfurization technology, with a high application rate in operational desulfurization units. Its core principle is SO2 removal through counter-current contact between slurry spraying and flue gas, making it a key facility for industrial flue gas treatment. However, the desulfurization tower operates in a harsh environment of high temperature, strong corrosion, and gas-liquid two-phase flow, which easily leads to scaling, nozzle blockage, and deterioration of slurry activity, directly affecting desulfurization efficiency and system stability. Therefore, the need for monitoring the desulfurization status is becoming increasingly urgent.

[0003] Currently, the monitoring of desulfurization status is mostly done by collecting abnormal data in real time to determine the fault type. By setting safety margins, upper and lower limits are defined for the monitored items. When the real-time data collected by the sensor exceeds the boundary or the rate of change changes abruptly, the system directly retrieves the corresponding fault judgment rules to determine the corresponding fault type.

[0004] However, desulfurization systems are complex physical-chemical coupled systems. When a fault occurs, it often tends to spread from a localized area to the entire system (e.g., chemical corrosion caused by abnormal slurry pH has a cumulative effect over time, while chemical reaction imbalances caused by mechanical failures such as oxidation blowers have a delayed effect). This concealed cross-domain transmission and asynchronous response time mean that related technologies can only generate scattered, isolated single-point alarm information when faced with related faults, making it difficult to systematically monitor faults and provide accurate and appropriate systemic adjustment solutions. Summary of the Invention

[0005] This application provides a multi-dimensional flue gas desulfurization status monitoring method and system, which is used to capture the coupling correlation between the process chemical reaction and the mechanical operation of the equipment in the desulfurization system, and realize the systematic prediction of desulfurization coupling faults.

[0006] Firstly, this application provides a multi-dimensional flue gas desulfurization status monitoring method, which includes: acquiring historical process parameters and corresponding fault type labels of the desulfurization system under historical fault conditions, wherein the historical process parameters include at least slurry data, flue gas inlet and outlet temperatures, and inlet and outlet SO2 concentrations; extracting multi-parameter joint process features from the historical process parameters, associating the process features with the fault type labels to establish a process fault diagnosis model; during the operation of the desulfurization system, performing online analysis of the collected current process parameters through the process fault diagnosis model to identify the corresponding current process fault; continuously collecting equipment status monitoring signals of the desulfurization target equipment, wherein the equipment status monitoring signals include at least temperature signals of the equipment body and motor, three-dimensional vibration signals, and operating electrical parameters; and performing spectral analysis on the equipment status monitoring signals to generate... A multi-type diagnostic atlas is generated. This atlas is then matched against a pre-defined equipment fault feature knowledge base, which stores multiple atlases corresponding to various equipment fault modes. When the matching result meets the judgment criteria for any equipment fault mode, the corresponding current equipment fault is output. In response to the identified current process fault and / or current equipment fault, a pre-defined process-equipment fault association mapping table is used to retrieve expected abnormal features within a pre-defined time window. These expected abnormal features are then verified and matched with actual monitoring data features to obtain a matching result. The expected abnormal features include anticipated equipment abnormal features that may result from the occurrence of the current process fault, and / or anticipated process abnormal features that result from the occurrence of the current equipment fault. Based on this matching result, the final fault level and adjustment plan are determined.

[0007] By adopting the above technical solution, this method breaks through the limitations of traditional desulfurization monitoring that only focuses on single-point threshold alarms for process parameters. It simultaneously integrates two core monitoring dimensions: process parameters and equipment status signals. First, it uses historical fault data to complete the correlation training of the process fault model, achieving accurate identification of process-level anomalies. Then, through the spectral analysis of equipment status signals and feature knowledge base matching, it completes the independent determination of equipment-side faults. Finally, it combines the process-equipment fault correlation mapping table for cross-validation, achieving linked tracing of the two types of faults. Compared with the traditional isolated monitoring mode, this solution can accurately capture the coupling correlation between the process chemical reaction and equipment mechanical operation of the desulfurization system, effectively avoiding the problems of false alarms and missed alarms for single parameters, comprehensively improving the comprehensiveness and accuracy of fault identification. At the same time, it directly outputs the appropriate fault level and adjustment scheme, improving the operational stability and desulfurization efficiency of the desulfurization system, and avoiding system shutdowns and equipment losses caused by fault propagation.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, the step of generating multiple types of diagnostic maps by performing spectral analysis on the device status monitoring signal specifically includes: performing time-frequency domain transformation on the three-dimensional vibration signal to generate a vibration spectrum diagram and a vibration trend diagram; performing time-series analysis on the temperature signal to generate a temperature change curve diagram and a temperature rise rate diagram; performing power analysis on the operating electrical parameters to generate a current waveform diagram and a power factor diagram; and determining the multiple types of diagnostic maps based on the vibration spectrum diagram, vibration trend diagram, temperature change curve diagram, temperature rise rate diagram, current waveform diagram, and power factor diagram.

[0009] By adopting the above technical solutions, specialized graphical analysis is conducted on equipment condition monitoring signals according to their types. This abandons the crude approach of traditional single-value monitoring. Time-frequency domain transformation is performed on three-dimensional vibration signals to extract spectral and trend features; time-series analysis is performed on temperature signals to capture temperature rise rates and fluctuation patterns; and power analysis is performed on operating electrical parameters to reconstruct waveform and power factor details. Multiple graphical representations comprehensively reconstruct the hidden anomalies of equipment operation from different dimensions. Compared to simply collecting raw signal values, this graphical analysis transforms abstract equipment condition signals into intuitive and comparable visual features, accurately capturing weak signals of early equipment degradation, effectively identifying hidden faults that are difficult to detect with traditional numerical monitoring, improving the sensitivity and accuracy of equipment fault diagnosis, and laying a reliable data foundation for subsequent equipment fault feature matching, ensuring no equipment-side fault determination is missed.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, the step of responding to the identified current process fault and / or current equipment fault, in conjunction with a preset process-equipment fault association mapping table, retrieving expected abnormal features within a preset time window, and verifying and matching the expected abnormal features with actual monitoring data features to obtain a matching result, specifically includes: marking the currently determined process fault and / or current equipment fault as the first abnormal event; based on the process-equipment fault association mapping table, determining the corresponding prediction target object, the expected abnormal feature, and the fault propagation direction according to the first abnormal event; the process-equipment fault association mapping table predefines the physical characteristics of each process fault. This involves identifying multiple equipment failures that may occur at the physical level, and multiple process failures that each equipment failure may cause at the chemical level. The direction of fault propagation is used to characterize the propagation path from the initial abnormal event to the corresponding predicted target object. The predicted severity of the initial abnormal event is obtained. Based on the fault propagation direction and the predicted severity, the transmission reaction lag time offset caused by media transmission or mechanical inertia in the desulfurization system is determined. The preset time window is shifted and compensated according to the transmission reaction lag time offset to obtain a dynamic compensation time window. The actual monitoring data characteristics of the predicted target object within the dynamic compensation time window are determined. The actual monitoring data characteristics are matched with the expected abnormal characteristics to obtain the matching result.

[0011] By adopting the above technical solution, the identified process or equipment faults are first marked as initial anomalies. Then, relying on a preset process-equipment fault association mapping table, the fault propagation path, prediction target, and expected characteristics are clarified. Simultaneously, considering the severity of the fault and the characteristics of system medium transmission and mechanical inertia, a time lag shift compensation is applied to the fixed time window, forming a dynamically adapted monitoring window. This design fully considers the lag and cumulative nature of fault propagation in the desulfurization system, solving the pain point that traditional fixed time windows are difficult to adapt to delayed fault propagation. It avoids missed detection of associated faults due to misaligned time windows, accurately tracks subsequent propagation signs of the initial fault, and achieves full-cycle tracking of faults from isolated single points to system-wide associations. This improves the timeliness and reliability of associated fault identification and eliminates systemic risks caused by the hidden spread of faults.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, the step of determining the corresponding predicted target object, the expected abnormal feature, and the fault propagation direction based on the process-equipment fault association mapping table and the initial abnormal event specifically includes: extracting multiple potential fault propagation paths corresponding to the initial abnormal event from the process-equipment fault association mapping table, wherein the fault propagation path includes an initial abnormal node, an intermediate transmission medium, and a potential target abnormal node; obtaining the current medium flow direction state parameters of the desulfurization system; determining the current activation state of the intermediate transmission medium in each fault propagation path based on the medium flow direction state parameters; screening out effective fault propagation paths in which the intermediate transmission medium is in an active state and determining the fault propagation direction; determining the potential target abnormal node of the effective fault propagation path as the predicted target object, and obtaining the corresponding abnormal characterization parameters recorded in the process-equipment fault association mapping table as the expected abnormal feature.

[0013] By adopting the above technical solution, all potential fault propagation paths corresponding to the initial anomaly are first extracted. Then, combined with the real-time medium flow direction status parameters of the desulfurization system, the activation state of the intermediate transmission medium in the path is determined, and the paths in an effective transmission state are accurately screened, thus locking in the true prediction target and expected anomaly characteristics. This step abandons the blindness of traditional fault association judgment, no longer performing indiscriminate verification of all potential paths, but dynamically screening based on the real-time operating status of the system to eliminate invalid transmission paths, reduce the computational load and false judgment probability of fault verification, and accurately lock in the true transmission direction of the fault, enabling rapid tracing of the root cause of the fault, avoiding interference from invalid monitoring and false alarms, further enhancing the pertinence and efficiency of fault diagnosis, and adapting to the complex and ever-changing real-time operating conditions of the desulfurization system.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, the step of determining the final fault level and adjustment scheme based on the matching result specifically includes: if the matching result is a mismatch, then determining the first abnormal event as an isolated fault level, and generating a local adjustment scheme for the first abnormal event.

[0015] By adopting the above technical solution, when the expected abnormal characteristics do not match the actual monitoring data, the first abnormality is directly judged as an isolated fault, and a targeted local adjustment plan is generated. This logic achieves refined hierarchical judgment of fault levels. Compared with the traditional crude mode of fault alarm without level differentiation and uniform handling, this method can quickly distinguish between isolated single-point faults and systemic cascading faults. For isolated faults without the risk of propagation, it directly provides a precise local adjustment strategy, avoiding system operation fluctuations caused by over-control, reducing unnecessary system intervention, ensuring rapid fault handling, maximizing the maintenance of normal operating conditions of the desulfurization system, reducing energy consumption and operation and maintenance costs of fault handling, and improving the continuity and economy of system operation.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, the step of determining the final fault level and adjustment scheme based on the matching result further includes: if the matching result is a match, determining that the initial abnormal event has fault propagation and setting the fault level as a system cascaded fault level; constructing a fault propagation chain based on the effective fault propagation path, the fault propagation chain including the initial abnormal event and at least one verified associated abnormal event; calculating a comprehensive risk score for the system cascaded fault based on the impact range and propagation intensity of each associated abnormal event in the fault propagation chain; and generating a system-level adjustment scheme based on the comprehensive risk score, the system-level adjustment scheme including a collaborative adjustment strategy for multiple abnormal nodes in the fault propagation chain.

[0017] By adopting the above technical solution, when the matching results confirm that the fault is transmitted across processes and equipment, the fault is directly classified as a system cascading fault, constructing a complete fault transmission chain. A comprehensive risk score is calculated by combining the impact range and transmission intensity of each associated anomaly, thereby generating a multi-node collaborative adjustment scheme. This design breaks through the limitations of traditional single-point handling of single faults. Addressing the characteristics of fault coupling and propagation in desulfurization systems, it upgrades from "single-point alarm" to "full-chain control," enabling comprehensive control of fault propagation and system risks. It avoids the drawbacks of single-point adjustments failing to eradicate cascading faults, and rapidly blocks fault transmission through multi-node collaborative control, fundamentally mitigating systemic operational risks, ensuring stable desulfurization efficiency, reducing the probability of major equipment damage and unplanned system downtime, and extending system lifespan.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, the step of determining the actual monitoring data characteristics of the predicted target object within the dynamic compensation time window specifically includes: acquiring multi-dimensional raw monitoring data of the predicted target object within the dynamic compensation time window; extracting process parameter anomaly features and equipment status anomaly features from the multi-dimensional raw monitoring data, wherein the process parameter anomaly features include the magnitude, duration, and trend of the deviation of the process parameters from the normal range, and the equipment status anomaly features include abnormal waveform features, spectral features, and statistical features of the equipment status monitoring signal; and aligning the process parameter anomaly features and the equipment status anomaly features with timestamps to generate time-series correlated actual monitoring data features.

[0019] By adopting the above technical solution, multi-dimensional raw data of the target object within a dynamic time window are first extracted. Then, the deviation magnitude, duration, and trend of process parameters, as well as the waveform, spectrum, and statistical characteristics of equipment signals, are decomposed separately. Finally, time stamp alignment is completed to generate time-series correlation features. This step achieves spatiotemporal unification of process and equipment anomaly features, solves the matching error problem caused by asynchronous data acquisition and inconsistent feature dimensions, ensures that the comparison between expected and actual features is based on the same time dimension and the same data benchmark, improves the accuracy of feature matching, and effectively avoids mismatches caused by misaligned data timing and fragmented features. This provides a real and accurate core basis for subsequent fault level determination and adjustment scheme formulation, ensuring the logical rigor and reliability of the entire monitoring method.

[0020] In a second aspect, this application provides a state monitoring system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the state monitoring system to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, this application provides a computer-readable storage medium including instructions that, when executed on a state monitoring system, cause the state monitoring system to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, this application provides a computer program product, including a computer program that, when run on a state monitoring system, causes the state monitoring system to perform the method described in the first aspect and any possible implementation thereof.

[0023] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0024] 1. By employing multi-dimensional linkage monitoring techniques, including historical fault data association training to construct a process fault diagnosis model, equipment status signal spectrum analysis and matching fault knowledge base, and cross-validation using a process-equipment association mapping table, the system effectively solves the technical problems of related technologies that can only provide single-point threshold alarms, are difficult to identify process and equipment coupled faults, and have fragmented and isolated fault judgments. This enables comprehensive and accurate identification of faults in the desulfurization system, avoids false alarms and missed alarms based on single parameters, directly outputs scientific fault levels and adaptation adjustment schemes, improves system operation stability and desulfurization efficiency, and eliminates the risk of equipment loss and downtime caused by fault propagation.

[0025] 2. By employing timing adaptation and tracking techniques that involve marking the initial anomaly, locating the propagation path based on the fault association mapping table, and dynamically compensating for the time window based on the system's lag characteristics, the technical problems of fixed time windows being difficult to adapt to the lag in fault propagation and the failure to detect associated and propagating faults are effectively solved. This enables accurate tracking of delayed fault propagation in the desulfurization system, avoids missed detection of associated faults due to time misalignment, achieves closed-loop management of the entire fault cycle, promptly blocks fault propagation, and improves the timeliness and reliability of associated fault identification.

[0026] 3. By employing precise path tracing techniques that extract potential transmission paths, combine real-time media flow direction to screen and activate effective paths, and dynamically lock the real transmission target, the technical problems of blind judgment of related technical faults, numerous invalid verifications, and high misjudgment rates are effectively solved. This enables precise positioning of fault transmission direction and rapid root cause tracing, eliminates invalid monitoring interference, reduces diagnostic calculations, improves the pertinence of fault judgment, and adapts to the complex real-time operating conditions of desulfurization systems. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of a scenario for the multi-dimensional flue gas desulfurization status monitoring method in the embodiments of this application;

[0028] Figure 2 This is a flowchart illustrating a multi-dimensional flue gas desulfurization status monitoring method in an embodiment of this application;

[0029] Figure 3 This is another flowchart illustrating the multi-dimensional flue gas desulfurization status monitoring method in this application embodiment;

[0030] Figure 4 This is a schematic diagram of the physical device structure of a status monitoring system in an embodiment of this application. Detailed Implementation

[0031] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. The singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0032] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0033] To facilitate understanding, the method provided in this implementation is described in a scenario below. Please refer to [link / reference]. Figure 1 This is a schematic diagram of a scenario for the multi-dimensional flue gas desulfurization status monitoring method in the embodiments of this application.

[0034] exist Figure 1 In the process, the desulfurization tower and auxiliary equipment at the desulfurization system site, and the target equipment (including but not limited to pumps / motors) generate two types of core data respectively:

[0035] Process-side data: Real-time collection of current process parameters, including slurry pH, slurry density, flue gas inlet and outlet temperatures, inlet and outlet SO2 concentrations, and tower pressure difference, through online monitoring equipment. These parameters are fed into the process fault diagnosis model (trained based on historical fault data). After online analysis, the current process faults (such as nozzle blockage, slurry activity deterioration, and SO2 removal efficiency not meeting standards) are identified.

[0036] Equipment-side data: Sensors deployed on the target equipment continuously collect equipment status monitoring signals, including temperature signals of the equipment body and motor, three-dimensional vibration signals, and operating electrical parameters. These signals are processed by the spectrum analysis and generation module and transformed into various types of diagnostic spectra, such as vibration spectrum diagrams, temperature change curves, and current waveform diagrams. Subsequently, these spectra are matched with the fault mode spectra of each equipment stored in the preset equipment fault feature knowledge base. When the matching result meets the judgment conditions, the corresponding current equipment fault (such as motor bearing wear, impeller imbalance, pump leakage, etc.) is output.

[0037] Once a current process fault and / or current equipment fault is identified, the expected abnormal characteristics within the same time window (including equipment anomalies that may be caused by the process fault and process anomalies that may be caused by the equipment fault) are retrieved using the process-equipment fault correlation mapping table, and verified and matched with the actual monitoring data characteristics. Finally, based on the verification and matching results, the final fault level and adjustment plan are determined, achieving predictive status monitoring and coordinated control of the desulfurization system.

[0038] The following is a flowchart describing the method provided in this implementation, based on the above description. Please refer to [link / reference]. Figure 2 This is a flowchart illustrating a multi-dimensional flue gas desulfurization status monitoring method in an embodiment of this application.

[0039] S101. Obtain the historical process parameters and corresponding fault type labels of the desulfurization system under historical fault conditions. The historical process parameters shall include at least slurry data, flue gas inlet and outlet temperatures and inlet and outlet SO2 concentrations.

[0040] The desulfurization system refers to a complete system that uses processes such as limestone-gypsum wet process to remove SO2 from industrial flue gas, including but not limited to desulfurization towers, slurry circulation pumps, oxidation fans, slurry pools and other equipment and supporting monitoring devices.

[0041] This step should be performed during the initialization phase of the condition monitoring system, and only after the desulfurization system has a certain operating history and has accumulated sufficient fault condition data.

[0042] As the executing entity, the status monitoring system first needs to review all recorded fault conditions during the past operation of the desulfurization system, clarify the operating time period corresponding to each type of fault, and then retrieve the process parameters collected by all online monitoring devices during that time period, i.e., historical process parameters. This must include at least slurry data (including but not limited to slurry temperature, pH, density, sulfite ion concentration, chloride ion concentration, etc.), flue gas inlet and outlet temperatures, and inlet and outlet SO2 concentrations. Other process parameters such as desulfurization tower inlet and outlet pressure difference, slurry circulation pump outlet pressure, and oxidation air flow rate can also be added to ensure data comprehensiveness. At the same time, for each set of historical process parameters, a clear fault type label is assigned according to the actual fault situation. The label must accurately correspond to the essence of the fault, such as "nozzle blockage," "tower scaling," "slurry activity deterioration," "insufficient oxidation," and "SO2 removal efficiency not meeting standards," to avoid ambiguous labels or classification errors. Subsequently, the system performs preliminary sorting of the retrieved historical process parameters and corresponding fault type labels, eliminating invalid data (such as abnormal values ​​caused by sensor failures or parameter groups with too much missing data) to ensure the validity and completeness of the data, ultimately forming a one-to-one correspondence dataset of "historical process parameters - fault type labels."

[0043] S102. Perform multi-parameter joint process feature extraction on the historical process parameters, associate the extracted process features with the fault type label for training, and establish a process fault diagnosis model.

[0044] This step is performed after a valid dataset of "historical process parameters - fault type labels" has been obtained, and before the desulfurization system is officially put into online monitoring. The condition monitoring system first performs multi-parameter joint process feature extraction on the historical process parameters obtained by S101:

[0045] The first step is to preprocess the original historical process parameters, including data noise reduction (removing abnormal peaks and valleys caused by sensor fluctuations), missing value completion (using interpolation, mean value and other methods to supplement missing parameter data), and time series alignment (adjusting process parameters collected by different monitoring devices and at different timestamps to a unified time interval to ensure the time series consistency of multiple parameters).

[0046] The second step is to use intelligent algorithms to explore the coupling relationship between various process parameters based on the desulfurization process principle, such as the relationship between slurry pH value and SO2 removal efficiency, the relationship between flue gas inlet temperature and slurry evaporation rate, and the relationship between circulating pump outlet pressure and slurry circulation flow rate.

[0047] The third step involves extracting process features that characterize different fault types using feature extraction algorithms. For example, for the "deterioration of slurry activity" fault, process features such as "pH fluctuation range", "pH adjustment response lag time" and "SO2 removal efficiency decrease" are extracted. For the "scaling inside the tower" fault, process features such as "rate of increase of pressure difference between inlet and outlet of the desulfurization tower" and "abnormal increase in flue gas outlet temperature" are extracted. This ensures that the extracted process features are unique and distinctive, and can effectively distinguish different types of process faults.

[0048] Subsequently, the system uses the extracted process features as input vectors and the corresponding fault type labels as output vectors, and selects an appropriate algorithm, such as a neural network, for correlation training. During training, the dataset is divided into a training set and a test set. The training set is used for the model to learn the mapping relationship between process features and fault types, while the test set is used to verify the diagnostic accuracy of the model. By continuously adjusting model parameters (such as feature weights, algorithm hyperparameters, etc.), the model performance is optimized, and the false positive rate and false negative rate are reduced. When the model's diagnostic accuracy, recall, and other indicators reach the preset standards, training is stopped, and a stable and reliable process fault diagnosis model is finally established.

[0049] This step transforms the original process parameters into process features that can accurately characterize fault features. By establishing a process fault diagnosis model through correlation training, it achieves an upgrade from "parameter monitoring" to "fault identification".

[0050] S103. During the operation of the desulfurization system, the current process parameters collected are analyzed online using the process fault diagnosis model to identify the corresponding current process faults.

[0051] This step is performed after the desulfurization system is officially put into operation and the process fault diagnosis model has been established. During the operation of the desulfurization system, the condition monitoring system collects the current process parameters in real time through online monitoring equipment deployed in key parts such as the desulfurization tower, slurry tank, and circulating pump. The collection frequency is set according to the operating scale of the desulfurization system and monitoring needs to ensure that changes in process parameters can be captured in a timely manner.

[0052] Subsequently, the system performs preprocessing on the collected current process parameters in the same manner as step S102, ensuring that the format and dimensions of the current process parameters are consistent with the input requirements of the process fault diagnosis model, thus avoiding model analysis failure due to data format mismatch. Then, the preprocessed current process parameters are input into the established process fault diagnosis model. Based on the mapping relationship between process features and fault types learned during training, the model performs online analysis on the current process parameters: extracting the process features corresponding to the current process parameters and matching them with the process features corresponding to various faults stored in the model, calculating the matching degree. When the matching degree reaches a preset threshold (e.g., above 85%), the fault type corresponding to the current process parameter is determined, i.e., the current process fault. If the matching degree does not reach the preset threshold, the current process is judged to be operating normally, with no process fault. Finally, the system outputs the identified current process faults (including fault type, fault occurrence time, and corresponding abnormal process parameters) to the operation and maintenance terminal in real time for operation and maintenance personnel to view.

[0053] This step enables real-time, online identification of process faults, allowing maintenance personnel to quickly grasp the specific type and abnormal parameters of the current process fault.

[0054] S104. Continuously collect equipment status monitoring signals of the desulfurization target equipment. The equipment status monitoring signals shall include at least the temperature signals of the equipment body and motor, three-dimensional vibration signals and operating electrical parameters.

[0055] This step is performed synchronously with step S103. That is, during the operation of the desulfurization system, while collecting the current process parameters in real time, the equipment status monitoring signals of the desulfurization target equipment are continuously collected. The application scenario is to realize the real-time monitoring of the operating status of the desulfurization target equipment, capture abnormal equipment signals, and provide data support for subsequent equipment fault identification.

[0056] Specifically, the condition monitoring system first clarifies the scope of the desulfurization target equipment. Based on the composition and structure of the desulfurization system, it screens out the core and fault-prone mechanical equipment to avoid overlooking critical equipment. Then, it deploys dedicated monitoring sensors at key locations of each desulfurization target equipment: temperature sensors are deployed at key locations such as bearings and housings of the equipment body and motor to collect temperature signals; three-dimensional vibration sensors are deployed at core vibration locations of the equipment body (such as pump body, impeller, and motor rotor) to collect vibration signals in the X, Y, and Z directions; and current and voltage sensors are deployed in the motor's power supply circuit to collect operating electrical parameters. Subsequently, the system continuously collects various equipment condition monitoring signals at a preset acquisition frequency (usually higher than the process parameter acquisition frequency) to ensure that it can capture subtle changes in the equipment's operating status, such as a slow increase in temperature or a gradual increase in vibration amplitude caused by motor bearing wear.

[0057] Meanwhile, the system performs preliminary noise reduction processing on the collected equipment status monitoring signals to eliminate abnormal signals caused by sensor interference, ensuring the authenticity and effectiveness of the signals, and stores the collected signals in real time to form a time-series dataset of equipment status monitoring signals.

[0058] S105. Perform spectral analysis on the status monitoring signals of the equipment to generate multiple types of diagnostic maps;

[0059] This step is executed after step S104 has continuously collected equipment status monitoring signals, and is performed in parallel with step S103. The application scenario is to transform abstract equipment status monitoring signals into intuitive diagnostic maps, providing visual support for subsequent equipment fault feature matching, and solving the problem that equipment status monitoring signals are abstract and difficult to directly determine equipment faults.

[0060] The condition monitoring system performs spectral analysis on the equipment condition monitoring signals collected by S104, generating multiple types of diagnostic spectral maps. The specific execution process is as follows:

[0061] The first step is to perform a spectral analysis on the three-dimensional vibration signal: using time-frequency domain transformation algorithms (such as Fourier transform, wavelet transform, etc.), the three-dimensional vibration signal is transformed from a time-domain signal (time-amplitude relationship) to a frequency-domain signal (frequency-amplitude relationship), generating a vibration spectrum. The frequency distribution of the vibration signal can be identified through the spectrum, and it can be determined whether there are abnormal frequency components (such as the characteristic frequency corresponding to bearing failure, the frequency corresponding to impeller imbalance). At the same time, a vibration trend graph is plotted, with time as the horizontal axis and vibration amplitude as the vertical axis, to intuitively display the trend of the amplitude of the three-dimensional vibration signal changing with time in different directions, making it easy to observe the change law of vibration amplitude (such as whether it continues to rise or whether there is a sudden change).

[0062] The second step is to perform graphical analysis on the temperature signal: using a time-series analysis algorithm, a temperature change curve is plotted, with time as the horizontal axis and temperature as the vertical axis, to show the temperature changes of key parts of the equipment body and motor over time, intuitively reflecting the temperature fluctuations and trends; at the same time, the temperature rise rate per unit time is calculated and a temperature rise rate graph is plotted, with time as the horizontal axis and temperature rise rate as the vertical axis, to show the changes in the temperature rise rate. Abnormal temperature rise rates (such as sudden increases) are usually a precursor to equipment failure.

[0063] The third step is to perform graphical analysis on the operating electrical parameters: using a power analysis algorithm, a current waveform graph is plotted, with time as the horizontal axis and current value as the vertical axis, to display the waveform changes of the motor's operating current and determine whether there are any abnormalities such as current fluctuations or distortions (e.g., abnormal current peak values ​​or irregular waveforms). At the same time, a power factor graph is plotted, with time as the horizontal axis and power factor as the vertical axis, to display the changes in the power factor over time. A low power factor usually indicates low motor operating efficiency or electrical faults.

[0064] Finally, the system integrates the generated vibration spectrum, vibration trend graph, temperature change curve, temperature rise rate graph, current waveform graph, power factor graph, etc., to form a multi-type diagnostic spectrum. The normal range threshold can be marked in the spectrum, which makes it easy to intuitively compare and judge whether there is an abnormality.

[0065] This step transforms abstract equipment status monitoring signals into intuitive and easy-to-understand diagnostic graphs, enabling visualized monitoring of equipment status. This allows for the rapid identification of abnormal equipment operation characteristics (such as abnormal vibration frequency, sudden temperature rise, and current distortion), providing a clear basis for subsequent equipment fault feature matching and improving the efficiency and accuracy of equipment fault identification.

[0066] S106. Perform feature matching between the multi-type diagnostic atlas and the preset equipment fault feature knowledge base. The equipment fault feature knowledge base stores multiple atlases corresponding to each equipment fault mode in advance.

[0067] This step is executed after generating multiple types of diagnostic maps, and is performed in parallel with the process fault identification in step S103. The application scenario is to identify equipment faults through feature matching.

[0068] The condition monitoring system first calls a pre-set equipment fault feature knowledge base. This knowledge base needs to be established in advance during the system initialization phase based on historical fault data, equipment manuals, and operation and maintenance experience of the desulfurization target equipment. For each equipment fault mode (such as motor bearing wear, impeller imbalance, coil aging, pump leakage, agitator jamming, etc.), various diagnostic spectra are collected when the fault occurs, and the core features of each spectra are extracted (such as the characteristic frequency of the vibration spectrum, the temperature rise amplitude of the temperature change curve, the distortion degree of the current waveform, etc.). These features are then associated with the fault modes and stored to form a correspondence between "equipment fault mode - diagnostic spectrum features". At the same time, a feature matching threshold is set for each fault mode (such as a similarity of ≥80% is considered a match).

[0069] Subsequently, the system extracts the core features of the multiple types of diagnostic maps generated by S105, including the frequency peaks and distribution of the vibration spectrum, the amplitude change rate of the vibration trend map, the temperature fluctuation range and temperature rise amplitude of the temperature change curve, the rate peak of the temperature rise rate map, the waveform distortion degree of the current waveform map, and the numerical fluctuation range of the power factor map. Then, the extracted features of the current diagnostic map are compared and matched one by one with the map features corresponding to all equipment fault modes in the equipment fault feature knowledge base, and the similarity between the two is calculated. In the matching process, the comprehensive features of multiple types of maps need to be combined for matching, rather than a single map feature. For example, to determine whether it is a "motor bearing wear" fault, the characteristic frequency of the vibration spectrum, the temperature rise of the temperature change curve, and the fluctuation of the current waveform need to be matched simultaneously to ensure the accuracy of the matching results. Finally, the feature matching degree of each equipment fault mode is recorded to provide a basis for subsequent fault determination.

[0070] This step achieves accurate matching between equipment diagnostic maps and fault modes by using a pre-set equipment fault feature knowledge base. It breaks through the limitations of traditional equipment monitoring, which can only detect signal abnormalities and is difficult to determine the fault type, and can quickly identify the specific type of equipment fault.

[0071] In some embodiments, to further improve the computational efficiency of equipment fault diagnosis, reduce the system computing power consumption caused by the generation of visualization maps, adapt to the low latency and high response requirements of real-time online monitoring in industrial sites, and at the same time take into account the practical needs of operation and maintenance personnel for intuitive interpretation, this step can be optimized to directly perform time-frequency domain analysis on the original signals of equipment status monitoring, extract multi-dimensional quantized status feature sets, and directly match and compare the quantized status feature sets with the standard feature data in the preset equipment fault feature knowledge base. The diagnostic map is no longer used as the core basis for feature matching, but only as a visualization display result output to the operation and maintenance monitoring interface, so that on-site operation and maintenance personnel can intuitively view the trend of equipment status changes and abnormal manifestations.

[0072] S107. When the matching result meets any of the judgment conditions of the device fault mode, output the corresponding current device fault.

[0073] The condition monitoring system retrieves the feature matching results from step S106 and checks the matching degree and judgment conditions between the current diagnostic spectrum features and the fault modes of each device. The judgment conditions need to be set in combination with the severity of the equipment fault and the stability of the features. They are usually divided into single spectrum feature judgment and multi-spectrum feature comprehensive judgment. For example, for faults with obvious features that can be characterized by a single spectrum (such as impeller imbalance, which has significant vibration spectrum features), the matching degree of a single spectrum feature can be set to be ≥ a preset threshold (such as 85%) to meet the judgment conditions. For faults with complex features that require multi-spectrum collaborative characterization (such as motor coil aging, which requires the combination of temperature, current, and vibration spectra), the matching degree of multiple spectrum features can be set to reach the corresponding threshold, or the comprehensive matching degree can be ≥ a preset threshold (such as 80%) to meet the judgment conditions.

[0074] If the current matching result meets the judgment conditions of any equipment failure mode, then the equipment is determined to have a corresponding failure, i.e., the current equipment failure. At the same time, the matching degree of the failure, the corresponding failure characteristics (such as abnormal vibration frequency, abnormal temperature amplitude, etc.), the failure occurrence time, the name of the equipment failure, etc. are recorded. If the current matching result does not meet the judgment conditions of any equipment failure mode, then the desulfurization target equipment is determined to be operating normally and there is no equipment failure.

[0075] Finally, the system will output the identified current equipment fault information (if any) to the operation and maintenance terminal in real time, and store it in the system database for subsequent cross-validation and fault tracing. If the matching results of multiple equipment fault modes meet the judgment conditions, the equipment fault with the highest matching degree will be output, and other potential faults will be marked for further verification by operation and maintenance personnel.

[0076] S108. In response to the identified current process fault and / or current equipment fault, the expected abnormal features within the same preset time window are retrieved by combining the preset process-equipment fault association mapping table, and the expected abnormal features are verified and matched with the actual monitoring data features to obtain the matching result. The expected abnormal features include expected equipment abnormal features that may occur after the current process fault occurs, and / or expected process abnormal features that occur after the current equipment fault occurs.

[0077] Among them, the current process fault and / or current equipment fault refers to the abnormal events output by S103 and S107, including three scenarios: identifying only the current process fault, identifying only the current equipment fault, and identifying both the current process fault and the current equipment fault simultaneously; the process-equipment fault association mapping table refers to a pre-constructed structured data table that stores the physical-chemical transmission relationship between process and equipment faults within the desulfurization system, such as "slurry pH value continuously below 4.5 (process fault) → increased pump / pipeline corrosion (intermediate transmission medium) → increased pump vibration amplitude and bearing temperature (equipment fault)" and "abnormal increase in circulating pump motor current (equipment fault) → decreased slurry circulation flow rate (intermediate transmission medium) → decreased SO2 removal efficiency and excessive outlet SO2 concentration (process fault)"; expected abnormal features refer to the set of associated abnormal parameters that may appear within a preset time window, retrieved from the mapping table based on the current process fault and / or current equipment fault, and are divided into two categories: expected equipment abnormal features caused by the current process fault (such as pH). The abnormality includes the shift in the peak value of the pump vibration spectrum after the anomaly, and the expected process anomaly caused by equipment failure (such as the reduction of SO2 concentration difference after the pump flow rate decreases). The actual monitoring data characteristics refer to the real-time process / equipment data characteristics collected within the same preset time window that are completely consistent with the dimensions of the expected abnormal characteristics, and must be aligned with the timestamp of the expected characteristics.

[0078] The purpose of this step is to predict fault propagation, that is, to determine whether a current process fault will lead to a future equipment fault, or whether a current equipment fault will lead to a future process fault, thereby realizing the shift from "post-event alarm" to "predictive maintenance". Specific details will be provided in steps S201-S207, and will not be repeated here.

[0079] S109. Based on the matching result, determine the final fault level and adjustment plan.

[0080] This step is executed after the matching result is obtained in step S108. It is the final step in the entire monitoring method. The application scenario is to clarify the urgency of the fault and the handling measures based on the fault matching result, so as to solve the problem that traditional monitoring only identifies faults and does not provide graded handling solutions, resulting in low efficiency and inaccurate fault handling by operation and maintenance personnel.

[0081] The condition monitoring system retrieves the matching results obtained in step S108, and, in conjunction with the specific type, severity, and scope of the current fault, as well as whether fault propagation exists, determines the final fault level and corresponding adjustment plan, specifically in two cases:

[0082] In the first scenario, if the matching result is "no match," it indicates that the current fault is an isolated fault, meaning that the fault has not triggered any related faults and is limited to a single fault at the process or equipment level, without affecting other parts of the system. In this case, the initial abnormal event (current process fault or current equipment fault) is determined to be an isolated fault level. This level of fault has a low urgency and a small impact range. For isolated fault levels, the system retrieves the standard handling measures corresponding to the type of the initial abnormal event from the preset fault handling knowledge base and generates a local adjustment plan. The local adjustment plan only targets the process link or equipment where the current fault occurs and does not require adjustment to the entire desulfurization system. For example, if the current fault is "fluctuation of slurry pH value" (process fault), the adjustment plan is to "fine-tune the limestone slurry replenishment amount, monitor the slurry pH value every 5 minutes, and ensure that the pH value is stable within the normal range (5.5-6.5)." If the current fault is "slight increase in motor bearing temperature" (equipment fault), the adjustment plan is to "strengthen equipment heat dissipation, monitor bearing temperature in real time, and reduce equipment load if the temperature continues to rise."

[0083] In the second scenario, if the matching result is "matched", it indicates that the current fault has a risk of fault propagation. That is, the initial abnormal event (current process fault or current equipment fault) has triggered related faults, forming a fault propagation chain. The fault has a large impact range and may affect the operational stability of the entire desulfurization system. In this case, the fault level is set as a system cascade fault level. This level of fault has a high degree of urgency and a wide range of impact. For the system cascade fault level, firstly, based on the effective fault propagation path in step S108, a complete fault propagation chain is constructed. The initial abnormal event and all verified related abnormal events in the fault propagation chain are identified, as well as the propagation order and intensity between each fault.

[0084] Then, based on the impact range (the impact range assessment is based on a comprehensive judgment of factors such as the number of equipment involved, the number of process steps, and the production indicators affected by the abnormal event) and the transmission intensity (calculated based on factors such as the matching confidence of abnormal features, the degree of parameter deviation, and the duration of abnormality), the system weights and sums the impact range and transmission intensity of each associated abnormal event to calculate the comprehensive risk score of the system cascading failure. The higher the comprehensive risk score, the greater the overall harm of the failure, and the more urgent and comprehensive the countermeasures need to be.

[0085] Finally, based on the comprehensive risk score, the system retrieves a system-level adjustment plan template from the fault handling knowledge base and makes customized adjustments according to the specific structure of the fault transmission chain. The system-level adjustment plan needs to formulate a coordinated adjustment strategy for multiple abnormal nodes in the fault transmission chain to ensure that the initial fault and related faults can be resolved at the same time, and to prevent the fault from being further transmitted and aggravated. For example, if the fault transmission chain is "nozzle blockage → increased circulation pump pressure → motor overload → decreased desulfurization efficiency", the system-level adjustment plan is to "immediately stop the machine to clean the nozzle blockage, repair the circulation pump pressure regulating device, check the motor operating status and perform cooling treatment, adjust the slurry circulation flow rate and limestone slurry replenishment, and simultaneously monitor the desulfurization efficiency and equipment parameters until all faults are resolved".

[0086] Furthermore, the system must output the final fault level and adjustment plan to the operation and maintenance terminal in real time, clearly defining the fault handling priority and specific measures. Simultaneously, it must record the fault level determination process and adjustment plan for subsequent operation and maintenance traceability and model optimization. This step, through scientific fault level classification and targeted adjustment plan generation, achieves differentiated handling of different types of faults, avoiding over-response to isolated faults and under-handling of cascading faults. By constructing a fault propagation chain and calculating a comprehensive risk score, the system can comprehensively assess the potential hazards of faults, formulate collaborative system-level adjustment strategies, effectively prevent fault propagation and cascading failures, and significantly improve the fault response capability and operational reliability of the desulfurization system.

[0087] In this embodiment, by employing multi-dimensional data acquisition of processes and equipment, coupled feature extraction, fault correlation prediction, and dynamic transmission verification techniques, the desulfurization system can be monitored in a full-process, predictive, and correlated manner. This effectively solves the problems of difficult identification of fault transmission, delayed early warning, and difficulty in precise control in traditional desulfurization monitoring, thereby achieving the technical effects of accurate identification of early faults and early prevention of cascading faults in the desulfurization system.

[0088] In light of the above scenarios, the method provided in this implementation will now be described in more detail. Please refer to [link / reference]. Figure 2 This is another flowchart illustrating the multi-dimensional flue gas desulfurization status monitoring method in this application embodiment.

[0089] S201. Mark the current process fault and / or the current equipment fault in the current judgment output as the first abnormal event;

[0090] If the condition monitoring system only identifies the current process fault and not the current process fault, then the current process fault is marked as the first abnormal event. If the condition monitoring system only identifies the current equipment fault and not the current process fault, then the current equipment fault is marked as the first abnormal event. If the condition monitoring system identifies both the current process fault and the current equipment fault, then the current process fault and the current equipment fault are marked as the first abnormal events respectively, and subsequent propagation verification is carried out separately to distinguish between concurrent faults and causal propagation faults.

[0091] S202. Based on the process-equipment fault association mapping table, determine the corresponding prediction target object, the expected abnormal characteristics, and the fault propagation direction according to the first abnormal event.

[0092] The process-equipment fault association mapping table refers to the database table mentioned in step S108 that is pre-built during the system initialization phase. It defines the potential causal relationship between process faults and equipment faults. The table predefines multiple equipment faults that each process fault may cause at the physical level, and multiple process faults that each equipment fault may cause at the chemical level. The table contains multiple potential fault propagation paths corresponding to each fault. The potential fault propagation paths are used to represent all theoretically feasible propagation paths that may be propagated from the first abnormal event to other related faults. Each path includes the fault propagation direction.

[0093] The application scenario for this step is to accurately screen out fault propagation paths with actual propagation potential, and to clearly identify the fault propagation direction and expected abnormal characteristics that fit the current system operating state.

[0094] The condition monitoring system first calls the preset process-equipment fault association mapping table. Based on the fault type (process fault or equipment fault) of the first abnormal event marked in step S201, it accurately locates all associated entries corresponding to the first abnormal event in the mapping table and extracts all possible potential fault propagation paths. Each potential fault propagation path must completely contain three core elements: the starting abnormal node (first abnormal event), the intermediate transmission medium, and the potential target abnormal node. The system needs to number and record each extracted path, clarify the core elements and propagation probability of each path (preset in the mapping table), and ensure that no potential path is missed, providing a complete basis for subsequent screening. For example, if the first abnormal event is "abnormal vibration of the circulating pump motor" (equipment fault), two potential paths are extracted from the mapping table: Path 1 "abnormal vibration of the circulating pump motor → mechanical force → wear of the circulating pump impeller" and Path 2 "abnormal vibration of the circulating pump motor → electricity → fluctuation of slurry circulation flow rate", both of which completely contain the three core elements and record the propagation probability.

[0095] The system uses online monitoring equipment in the desulfurization system to collect all current status parameters related to the intermediate conduction medium in real time, namely the medium flow direction status parameters. The collection range covers the intermediate conduction medium (such as slurry, flue gas, electricity, etc.) in all potential fault conduction paths.

[0096] The system retrieves the intermediate transmission medium for each potential fault transmission path, combines it with previously acquired medium flow direction parameters, and compares it against preset medium activation criteria to determine the current activation status (activated / inactive) of the intermediate transmission medium. The criteria are pre-set based on the desulfurization system's process principles and medium transmission characteristics, with the core principle being "the medium possesses normal flow and transmission capabilities, enabling fault transmission, thus indicating an activated state." For example, for a path where the intermediate transmission medium is "slurry," if the slurry circulation flow rate is within the normal range, the pressure is stable, the flow direction is normal, and there are no abnormalities such as slurry stagnation or leakage, it is determined to be in an activated state; if the slurry flow rate is 0, the pressure is abnormally low, or there is slurry leakage, making fault transmission difficult, it is determined to be inactive. For a path where the intermediate transmission medium is "flue gas," if the flue gas flow rate is normal, the flow direction is correct, and there are no blockages, it is determined to be in an activated state; if the flue gas flow rate is abnormal, the flue is blocked, and normal flow is difficult, it is determined to be inactive. The system must clearly determine the medium activation status of each potential path and record the determination basis (such as parameter values ​​and whether they meet the determination criteria).

[0097] The system filters all potential fault propagation paths, eliminating paths where the intermediate propagation medium is inactive, and retaining only paths where the intermediate propagation medium is active. These are considered valid fault propagation paths—paths that possess the actual conditions for fault propagation and are the paths where the initial abnormal event may actually trigger related faults. After filtering, the system integrates all valid fault propagation paths, clarifying the starting abnormal node, intermediate propagation medium, potential target abnormal node, and propagation sequence for each valid path, thereby determining the fault propagation direction: the fault propagation direction starts from the initial abnormal event and ends at the potential target abnormal node, clarifying the transmission process of the intermediate propagation medium. If multiple valid paths exist, multiple fault propagation directions are corresponding to them, and the fault propagation directions are prioritized according to the preset propagation probabilities in the mapping table (the direction with the higher propagation probability is the primary propagation direction) for subsequent steps to focus on.

[0098] The system retrieves all potential target anomaly nodes along valid fault propagation paths as prediction targets (i.e., associated faults that may be triggered by the initial anomaly event). For each potential target anomaly node, it obtains the corresponding anomaly characterization parameters recorded in the process-equipment fault association mapping table and extracts its corresponding core anomaly features as expected anomaly features. If the potential target anomaly node is an equipment fault, it extracts equipment status anomaly features (such as the anomaly thresholds and trends of temperature, vibration, and electrical parameters); if the potential target anomaly node is a process fault, it extracts process parameter anomaly features (such as the anomaly thresholds and trends of slurry and flue gas parameters). Subsequently, the anomaly features corresponding to all potential target anomaly nodes are integrated, duplicate features are removed, and a complete set of expected anomaly features is formed. This set serves as the core benchmark for feature matching in the subsequent S207 step, ensuring that the expected anomaly features accurately correspond to the associated faults that may occur in the current system.

[0099] For scenarios where two types of initial anomalies are marked simultaneously in S201, the system will independently execute the above-mentioned retrieval, filtering, and determination process for the initial anomalies of the process type and the equipment type, respectively, to obtain two sets of corresponding predicted target objects, expected anomaly characteristics, and fault propagation directions, and then conduct verification separately.

[0100] This step breaks through the limitations of traditional methods that directly use all potential fault propagation paths. By combining the current medium flow direction state parameters of the desulfurization system, it selects effective fault propagation paths with actual propagation potential, ensuring that the determination of fault propagation direction is consistent with the current operating state of the system and avoiding interference from invalid paths to subsequent steps.

[0101] In some embodiments, in order to further address the hidden physical-chemical coupling failure problem that may occur in the desulfurization system under complex operating conditions (such as power grid transient disturbances and deep peak shaving of the unit), this application further introduces a chemical buffer inertia assessment mechanism and a scaling potential tracing mechanism in some preferred embodiments to improve the accuracy of predictive condition monitoring and the engineering fault tolerance rate.

[0102] Preferred Implementation Example 1: Fault Propagation Blocking Based on Chemical Buffer Inertia Assessment. Following the step of "determining the corresponding expected abnormal characteristics and fault propagation direction based on the initial abnormal event according to the process-equipment fault association mapping table," if the initial abnormal event is determined to be an equipment fault (e.g., a sudden drop in slurry circulation pump speed, fluctuations in flow rate of some spray layers), and the potential target abnormal node is a process fault (e.g., excessive SO2 outlet concentration), the system executes the following steps:

[0103] First, obtain the macroscopic chemical state parameters within the current desulfurization absorption tower, including real-time liquid level, slurry density, and pH value. Based on the real-time liquid level and the cross-sectional area of ​​the desulfurization tower bottom, calculate the absolute volume of the current slurry pool. Combine this with the real-time density and pH value to calculate the amount of unreacted calcium carbonate in the current slurry pool, which serves as the current chemical buffer. Based on the magnitude of the decrease in medium flow rate caused by the initial abnormal event (equipment failure) and the expected duration of the failure, calculate the theoretical additional chemical buffer required during the expected failure duration, i.e., the failure impact equivalent. Compare this chemical buffer with the failure impact equivalent. If the chemical buffer is greater than the failure impact equivalent multiplied by a preset safety margin coefficient, it indicates that the macroscopic chemical buffer capacity of the desulfurization system is sufficient to completely absorb the transient disturbance from the mechanical equipment. At this point, the system actively blocks the transmission path from the initial abnormal event to the process failure, downgrading the initial abnormal event to an isolated mechanical transient failure for local alarm purposes, without triggering matching verification or system cascading alarms for the process failure.

[0104] Through the above embodiments, the system can identify and accommodate transient disturbances caused by power grid fluctuations or brief mechanical jamming, and use the large chemical inertia of the absorption tower itself to offset false alarms. This solves the technical problem in traditional algorithms where local transient anomalies cause a chain of alarms across the entire domain, leading to alarm overload and interference with normal operation and maintenance.

[0105] Preferred Implementation Example 2: Under actual unit deep peak shaving and other variable load conditions, some equipment will be shut down. In the step of screening out the effective fault transmission paths in which the intermediate transmission medium is in an active state, for fault transmission paths determined to be "inactive" (e.g., the shut-down standby circulating pump and its corresponding spray pipeline), the system does not directly ignore them, but performs the following dormant tracking steps:

[0106] The physical pipelines corresponding to the inactive conduction paths are added to the dormant tracking queue, and the initial density of residual slurry and the initial timestamp of shutdown are recorded at the moment the pipeline stops operating. During the dormant period when the pipeline is inactive, the real-time temperature change of the environment in which the pipeline is located is continuously acquired, and the initial density of residual slurry, the decrease in ambient temperature, and the duration of dormancy are integrated according to the preset slurry solid phase deposition rate formula, and the scaling potential index of the dormant pipeline is dynamically output in real time. The scaling potential index is used to quantify the degree of hardening of gypsum crystals precipitating from the liquid phase and adhering to the pipe wall in the stagnant state. When the status monitoring system receives a reactivation command for the equipment where the dormant pipeline is located (such as receiving a DCS command to start the standby pump), the system first retrieves the scaling potential index of the pipeline at the current moment. If the scaling potential index exceeds the preset scaling density safety threshold, it indicates that destructive hard scale has been formed in the pipeline. At this time, the system intercepts the reactivation command, prevents the main equipment from starting, and prioritizes sending a pulse flushing activation command to the DCS control system; after the high-pressure pulse flushing disperses the dormant deposits and verifies that the pipeline is unobstructed, the system releases the command interception, allowing the conductive medium to resume its active state and be reintegrated into the normal fault conduction monitoring logic, and executes the normal fault conduction path determination and monitoring process.

[0107] This technical solution addresses the blind spots of conventional monitoring systems, which only monitor operating equipment and neglect the status of shut-down equipment and related pipelines. By leveraging the fluid characteristics of limestone-gypsum wet desulfurization slurry—which readily deposits solids and hardens scale—it transforms the non-operational downtime of shut-down pipelines into a quantitatively monitorable period for scale risk management. This allows for proactive assessment and intervention of scale risks before equipment restart. By intercepting high-risk restart commands in advance and coordinating pre-flushing operations, it effectively avoids serious secondary equipment failures such as overload and burnout of circulating pumps and clogging and tearing of spray nozzles caused by blindly starting scaled pipelines. This significantly improves the hardware operational safety of the desulfurization system under varying unit loads and frequent equipment start-ups and shutdowns, enhancing the foresight and risk control capabilities of condition monitoring.

[0108] S203. Obtain the predicted severity of the initial abnormal event;

[0109] The condition monitoring system predicts the severity of the initial anomaly based on its core abnormal characteristics, pre-defined severity assessment criteria, and historical severity classification data and development patterns of similar faults. For example, the severity of a process fault is determined by the magnitude and duration of parameter deviations and their impact on desulfurization efficiency; the severity of an equipment fault is determined by the amplitude and rate of change of abnormal equipment characteristics and their impact on equipment functionality. After assessment, the system assigns a specific predicted severity level to the initial anomaly and records the assessment criteria (such as parameter deviation and anomaly duration), synchronously linking it to the fault propagation direction to support the calculation of the propagation reaction lag time offset in step S204.

[0110] For the third scenario in S201 where two types of initial anomalies are marked simultaneously, the system independently calculates and predicts the severity of process-related and equipment-related anomalies.

[0111] S204. Based on the direction of fault propagation and the predicted severity, determine the transmission reaction lag time offset caused by the medium transmission or mechanical inertia of the desulfurization system.

[0112] The transmission response lag time offset refers to the time offset calculated based on the medium transmission characteristics and mechanical inertia of the fault transmission direction, combined with the predicted severity of the initial abnormal event. It is used to correct a preset time window. Its core function is to compensate for the time delay during fault transmission, ensuring that the time window accurately covers the occurrence time of the associated fault. Example: The fault transmission direction is "nozzle blockage → slurry circulation obstruction → increased circulation pump pressure," the transmission medium is slurry, the predicted severity is "moderate," the slurry transmission delay time is 15 minutes, the mechanical inertia delay time is 5 minutes, and the calculated transmission response lag time offset is 20 minutes. This means the associated fault (increased circulation pump pressure) will occur approximately 20 minutes after the initial abnormal event (nozzle blockage).

[0113] This step is executed to correct time deviations within the preset same time window, addressing the problem that traditional cross-validation, which uses a fixed time window, struggles to adapt to the lag characteristics of fault propagation, leading to missed detection of associated faults. The condition monitoring system retrieves the fault propagation direction and predicted severity, analyzing the core transmission medium (slurry, flue gas, etc.) and mechanical components (circulating pumps, motors, etc.) along the fault propagation direction: For fault paths using slurry as the transmission medium, the medium transmission lag time is calculated based on parameters such as slurry flow rate, pipe length, and slurry concentration; for fault paths using flue gas as the transmission medium, the medium transmission lag time is calculated based on parameters such as flue gas flow rate, flue gas temperature, and desulfurization tower structure; for fault paths involving mechanical components, the mechanical inertia lag time is calculated based on equipment inertial parameters, operating speed, and wear degree.

[0114] Subsequently, based on the predicted severity of the initial abnormal event, the aforementioned lag time is corrected: when the predicted severity is "severe," the fault propagation speed is fast, and the lag time is shortened (correction coefficient is 0.8); when the predicted severity is "moderate," the lag time remains unchanged (correction coefficient is 1.0); when the predicted severity is "minor," the fault propagation speed is slow, and the lag time is extended (correction coefficient is 1.2). Finally, the corrected media transmission lag time is added to the mechanical inertia lag time to obtain the transmission response lag time offset. This offset can be positive or negative (positive offset indicates that the associated fault occurred later than the preset time, and negative offset indicates that the associated fault occurred earlier than the preset time), and is stored in the system for time window shift compensation in step S205.

[0115] This step accurately calculates the time lag offset during fault propagation, avoiding the shortcomings of fixed time windows that are difficult to adapt to different fault propagation characteristics. It provides an accurate basis for the generation of subsequent dynamic compensation time windows, ensuring that the abnormal characteristics of associated faults can be accurately captured.

[0116] S205. Based on the transmission response lag time offset, the preset same time window is shifted and compensated to obtain a dynamic compensation time window.

[0117] This step ensures that the time window accurately covers the occurrence time of the associated fault. The condition monitoring system retrieves the preset identical time window (with its start and end times and duration specified) from step S108 and the transmission response lag time offset calculated in step S204. Then, based on the positive or negative value of the transmission response lag time offset, the preset identical time window is shifted for compensation: if the transmission response lag time offset is positive, it indicates that the associated fault occurred later than the preset time window, and the preset time window is shifted backward by the offset amount, keeping the time window duration unchanged; if the transmission response lag time offset is negative, it indicates that the associated fault occurred earlier than the preset time window, and the preset time window is shifted forward by the offset amount, keeping the time window duration unchanged; if the offset is 0, then the preset identical time window is the dynamic compensation time window.

[0118] After the translation compensation is completed, the system generates a dynamic compensation time window, specifying its start and end times, and correlates it with the initial abnormal event and the direction of fault propagation. This ensures that the dynamic compensation time window accurately covers the possible time range of associated faults—neither prematurely capturing irrelevant anomalies nor lagging behind and omitting associated fault characteristics. Simultaneously, the system records the translation compensation process of the time window (preset time window, offset, dynamic compensation time window) for subsequent maintenance traceability and parameter optimization. This step achieves dynamic optimization of the time window, enabling it to adapt to the propagation lag characteristics of different faults and accurately cover the occurrence time of associated faults.

[0119] S206. Determine the actual monitoring data characteristics of the predicted target object within the dynamic compensation time window;

[0120] The status monitoring system first locks the predicted target object determined in S202, retrieves the full-dimensional raw real-time monitoring data of the object within the dynamic compensation time window generated in S205, including time-domain signals, numerical data, waveform data, etc. collected by various sensors, and preprocesses the raw data first.

[0121] After preprocessing, the system extracts corresponding actual features according to the dimensions of expected abnormal characteristics: if the target object is a process, it focuses on extracting the deviation of parameters such as slurry pH, density, temperature, and inlet / outlet SO2 concentration from the baseline value, the duration of continuous abnormality, the linear change trend, and the fluctuation amplitude, quantifying each indicator into specific values ​​to form a set of actual monitoring data features for the process; if the target object is equipment, it performs time-frequency domain analysis on the three-dimensional vibration signal to extract spectral features such as peak value, kurtosis, and waveform factor, extracts statistical features such as the temperature rise rate per unit time and the highest temperature value from the temperature signal, and extracts waveform features such as the current waveform distortion rate and power factor fluctuation value from the operating electrical parameters to form a set of actual monitoring data features for the equipment.

[0122] After extraction, the system aligns all actual features one by one according to their timestamps to ensure that all features are under the same temporal reference and completely correspond to the temporal and dimensional aspects of the expected abnormal features, thus avoiding subsequent matching errors due to temporal misalignment. For dual dynamic window scenarios, the system extracts the actual features of the corresponding predicted target object in each of the two windows, stores and processes them independently, and avoids mixing them.

[0123] The technical effect of this step is to accurately extract the core quantitative features of the target object within the effective time period, realize the refined transformation from raw data to feature data, ensure the dimensional consistency and temporal alignment of the actual features and expected features, and provide a real, accurate and comparable data foundation for subsequent feature matching.

[0124] S207. Match the actual monitoring data features with the expected abnormal features to obtain the matching result.

[0125] This step is used to determine whether the initial abnormal event has led to fault propagation through feature matching. The status monitoring system first retrieves the actual monitoring data feature set from step S206 and the expected abnormal feature set from step S202, ensuring that the types and dimensions of the two types of features are consistent. Then, a preset feature similarity algorithm is used to compare the two types of features one by one, calculating the similarity of each corresponding feature. Based on the weights of each feature (higher weight for core features, lower weight for secondary features), a comprehensive similarity is calculated. Subsequently, based on a preset similarity threshold, the matching result is determined: if the comprehensive similarity reaches the preset threshold, and the occurrence time and trend of the actual monitoring data features are consistent with the expected abnormal features, and also conform to the temporal pattern of fault propagation, the matching result is "match," indicating that the initial abnormal event has triggered a related fault and fault propagation exists. If the comprehensive similarity does not reach the preset threshold, or the occurrence time and trend of the actual monitoring data features are inconsistent with the expected abnormal features, or do not conform to the temporal pattern of fault propagation, the matching result is "no match," indicating that the initial abnormal event has not triggered a related fault and is an isolated fault.

[0126] For scenarios where two types of initial anomalies exist simultaneously, the system performs independent matching processes on the two sets of actual features and expected features, outputs two sets of matching results, and determines the propagation of the two types of initial anomalies respectively.

[0127] In some embodiments, if two types of initial anomalies exist simultaneously, i.e., the current process fault and the current equipment fault are marked as initial anomaly events, it is possible to verify whether the current process fault is caused by the current equipment fault, or whether the current equipment fault is caused by the current process fault, thereby determining whether the current process fault and the current equipment fault have had derivative effects. Specifically, in step S201, the status monitoring system can first retrieve the timestamp information corresponding to the current process fault and the current equipment fault respectively, accurately compare the initial occurrence sequence of the two types of anomalies, confirm the first occurrence as the initial anomaly event, and then combine the process-equipment fault association mapping table, fault propagation direction, dynamic compensation time window and feature matching results to carry out bidirectional causal verification, rather than single-direction judgment, to clarify the causal relationship and derivative propagation relationship between the two types of anomalies.

[0128] More specifically, for scenarios where the current equipment failure timestamp is earlier than the current process failure timestamp, and the time difference is within the propagation response lag time offset range calculated by S204, the status monitoring system initiates a positive causal verification of "equipment failure leading to process failure": taking the current equipment failure as the core initial anomaly, it retrieves the corresponding effective fault propagation path and expected process anomaly characteristics, verifies whether the specific type and parameter anomaly characteristics of the current process failure completely match the expected process anomaly characteristics, and checks whether the intermediate transmission medium in the propagation path is in an active state and whether the medium flow direction conforms to the propagation logic. If the feature matching degree meets the standard, the propagation path is effective, and the timing conforms to the lag time rule, it is determined that the current process failure is caused by the current equipment failure. The two form a one-way derivative effect of "equipment failure → process failure", with the current equipment failure being the root cause failure and the current process failure being the derivative failure. For scenarios where the current process failure timestamp is earlier than the current equipment failure timestamp, and the time difference falls within the corresponding transmission reaction lag time offset range, the system initiates a reverse causal verification of "process failure leading to equipment failure": taking the current process failure as the core initial anomaly, it retrieves the corresponding effective equipment transmission path and expected equipment anomaly characteristics, verifies whether the vibration, temperature, and electrical characteristics of the current equipment failure match the expected equipment anomaly characteristics, and checks whether the operating status of intermediate transmission media such as slurry and flue gas supports the transmission path. If all the judgment conditions are met, it is determined that the current equipment failure is caused by the transmission of the current process failure, and the two form a one-way derivative impact of "process failure → equipment failure". The current process failure is the root cause failure, and the current equipment failure is the derivative failure. This can fundamentally resolve the risk of continuous failure spread and avoid repeated failures caused by only dealing with derivative failures without addressing the root cause, effectively achieving accurate source tracing and closed-loop management of desulfurization system failures.

[0129] In this embodiment, by employing a full-process linkage technology that includes initial anomaly marking, fault association path location, dynamic lag time compensation, adaptive time window adjustment, and bidirectional feature matching verification, it is possible to accurately track the cross-domain transmission patterns of process and equipment faults in the desulfurization system, adapt to the system's time delay characteristics to carry out predictive monitoring, and effectively solve the problems of isolated fault judgment, difficulty in identifying cascading fault transmission, poor time delay adaptability, missed or misjudged associated faults, and difficulty in predicting the risk of propagation in advance in traditional desulfurization monitoring. Thus, it achieves the technical effect of accurate location of the source of faults in the desulfurization system and early warning of cascading transmission.

[0130] The status monitoring system in the embodiments of this invention is described below from the perspective of hardware processing. Please refer to [link / reference]. Figure 4 This is a schematic diagram of the physical device structure of a status monitoring system in an embodiment of this application.

[0131] It should be noted that, Figure 4The structure of the status monitoring system shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0132] like Figure 4 As shown, the status monitoring system includes a Central Processing Unit (CPU) 401, which can perform various appropriate actions and processes based on programs stored in Read-Only Memory (ROM) 402 or programs loaded from storage section 408 into Random Access Memory (RAM) 403, such as performing the methods described in the above embodiments. The RAM 403 also stores various programs and data required for system operation. The CPU 401, ROM 402, and RAM 403 are interconnected via a bus 404. An Input / Output (I / O) interface 405 is also connected to the bus 404.

[0133] The following components are connected to I / O interface 405: input section 406 including audio input devices, push-button switches, etc.; output section 407 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 408 including a hard disk, etc.; and communication section 409 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 409 performs communication processing via a network such as the Internet. Drive 410 is also connected to I / O interface 405 as needed. Removable media 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 410 as needed so that computer programs read from them can be installed into storage section 408 as needed.

[0134] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing computer programs for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit (CPU) 401, it performs the various functions defined in the present invention.

[0135] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0136] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those shown in the drawings.

Claims

1. A multi-dimensional flue gas desulfurization status monitoring method, characterized in that, The method includes: Obtain historical process parameters and corresponding fault type labels of the desulfurization system under historical fault conditions. The historical process parameters include at least slurry data, flue gas inlet and outlet temperatures, and inlet and outlet SO2 concentrations. Multi-parameter joint process feature extraction is performed on the historical process parameters, and the extracted process features are associated with the fault type labels for training to establish a process fault diagnosis model. During the operation of the desulfurization system, the collected current process parameters are analyzed online using the process fault diagnosis model to identify the corresponding current process faults. Continuously collect equipment status monitoring signals of the desulfurization target equipment. The equipment status monitoring signals include at least the temperature signals of the equipment body and motor, three-dimensional vibration signals, and operating electrical parameters. The device status monitoring signals are analyzed to generate multiple types of diagnostic maps; The multi-type diagnostic atlases are matched with a preset equipment fault feature knowledge base, which stores multiple atlases corresponding to each equipment fault mode in advance. When the matching result meets the determination condition of any of the aforementioned device failure modes, the corresponding current device failure is output; In response to the identified current process fault and / or current equipment fault, the expected abnormal features within the same preset time window are retrieved in conjunction with a preset process-equipment fault association mapping table, and the expected abnormal features are verified and matched with the actual monitoring data features to obtain a matching result. The expected abnormal features include expected equipment abnormal features that may occur after the current process fault occurs, and / or expected process abnormal features that occur after the current equipment fault occurs. The final fault level and adjustment plan are determined based on the matching results.

2. The method according to claim 1, characterized in that, The step of generating multiple types of diagnostic maps by performing spectral analysis on the equipment status monitoring signals specifically includes: The three-dimensional vibration signal is subjected to time-frequency domain transformation to generate a vibration spectrum diagram and a vibration trend diagram; Perform time-series analysis on the temperature signal to generate a temperature change curve and a temperature rise rate graph; Power analysis is performed on the operating electrical parameters to generate current waveforms and power factor diagrams; The multi-type diagnostic maps are determined based on the vibration spectrum diagram, vibration trend diagram, temperature change curve diagram, temperature rise rate diagram, current waveform diagram, and power factor diagram.

3. The method according to claim 1, characterized in that, The step of responding to the identified current process fault and / or current equipment fault by retrieving expected abnormal features within a preset time window, combining a preset process-equipment fault association mapping table, and verifying and matching the expected abnormal features with actual monitoring data features to obtain a matching result specifically includes: Mark the current process fault and / or the current equipment fault in the current judgment output as the first abnormal event; Based on the process-equipment fault association mapping table, the corresponding prediction target object, the expected abnormal characteristics, and the fault propagation direction are determined according to the first abnormal event. The process-equipment fault association mapping table predefines multiple equipment faults that each process fault may cause at the physical level, and multiple process faults that each equipment fault may cause at the chemical level. The fault propagation direction is used to characterize the propagation path from the first abnormal event to the corresponding prediction target object. Obtain the predicted severity of the initial abnormal event; The transmission reaction lag time offset caused by the medium transmission or mechanical inertia loss of the desulfurization system is determined based on the fault transmission direction and the predicted severity. Based on the transmission response lag time offset, the preset same time window is shifted and compensated to obtain a dynamic compensation time window; Determine the actual monitoring data characteristics of the predicted target object within the dynamic compensation time window; The actual monitoring data features are matched with the expected anomaly features to obtain the matching result.

4. The method according to claim 3, characterized in that, The step of determining the corresponding prediction target object, the expected anomaly characteristics, and the fault propagation direction based on the process-equipment fault association mapping table and the initial abnormal event specifically includes: Extract multiple potential fault propagation paths corresponding to the first abnormal event from the process-equipment fault association mapping table. The fault propagation path includes the initial abnormal node, intermediate propagation medium and potential target abnormal node. Obtain the current medium flow direction status parameters of the desulfurization system; Based on the medium flow direction state parameters, determine the current activation state of the intermediate conduction medium in each fault conduction path; Valid fault conduction paths in which the intermediate conductive medium is in an active state are selected, and the fault conduction direction is determined; The potential target anomaly node of the effective fault propagation path is determined as the prediction target object, and the corresponding anomaly characterization parameters recorded in the process-equipment fault association mapping table are obtained as the expected anomaly features.

5. The method according to claim 1, characterized in that, The step of determining the final fault level and adjustment plan based on the matching results specifically includes: If the matching result is a mismatch, the first abnormal event is determined to be an isolated fault, and a local adjustment plan is generated for the first abnormal event.

6. The method according to claim 4, characterized in that, The step of determining the final fault level and adjustment plan based on the matching results further includes: If the matching result is a match, it is determined that the initial abnormal event has fault propagation, and the fault level is set to the system cascading fault level; A fault propagation chain is constructed based on the effective fault propagation path, and the fault propagation chain includes the initial abnormal event and at least one verified associated abnormal event. Based on the impact range and transmission intensity of each associated abnormal event in the fault propagation chain, calculate the comprehensive risk score of the system cascading fault; A system-level adjustment plan is generated based on the comprehensive risk score. The system-level adjustment plan includes a coordinated adjustment strategy for multiple abnormal nodes in the fault propagation chain.

7. The method according to claim 3, characterized in that, The step of determining the actual monitoring data characteristics of the predicted target object within the dynamic compensation time window specifically includes: Obtain multi-dimensional raw monitoring data of the predicted target object within the dynamic compensation time window; The abnormal features of process parameters and abnormal features of equipment status are extracted from the multi-dimensional raw monitoring data. The abnormal features of process parameters include the magnitude of deviation of process parameters from the normal range, the duration of deviation, and the trend of change. The abnormal features of equipment status include the abnormal waveform features, spectral features, and statistical features of equipment status monitoring signals. The abnormal features of the process parameters and the abnormal features of the equipment status are timestamped to generate the actual monitoring data features with time-series correlation.

8. A condition monitoring system, characterized in that, The status monitoring system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the status monitoring system to perform the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that, When the instruction is executed on the condition monitoring system, the condition monitoring system performs the method as described in any one of claims 1-7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is run on the condition monitoring system, it causes the condition monitoring system to perform the method as described in any one of claims 1-7.

Images