Anti-abrasion and anti-explosion intelligent real-time monitoring system and method for boiler

Abstract

The embodiment of the invention discloses an anti-wear and anti-explosion intelligent real-time monitoring system for a boiler, which relates to the technical field of industrial intellectualization and comprises a boiler data acquisition module, an edge intelligent processing module, a parallel acceleration operation module, a multi-source communication module and an intelligent early warning module, the boiler data acquisition module is used for acquiring sensor data of each part of the boiler through high frequency. The invention further discloses an anti-abrasion and anti-explosion intelligent real-time monitoring method for the boiler. According to the anti-abrasion and anti-explosion intelligent real-time monitoring system and method for the boiler, data of sensors at all parts of the boiler are collected at high frequency, preliminary data processing and analysis are carried out on edge equipment, abnormal data can be recognized in real time, early warning is triggered, the safety and reliability of boiler operation are remarkably improved, and the safety and reliability of boiler operation are improved. The data is screened and preprocessed, the amount of data transmitted to the central system is reduced, the transmission delay is reduced, the data processing efficiency is improved, and the potential abnormal mode of the data can be quickly mined.

Description

Technical Field

[0001] The invention belongs to the technical field of industrial intelligence, and specifically relates to an intelligent real-time monitoring system and method for boiler wear and explosion prevention. Background Art

[0002] During the operation of the boiler, due to the long-term high temperature and high pressure working environment, various parts of the boiler are prone to wear and corrosion, which may cause serious faults such as pipe bursting, which not only affects the normal operation of the boiler, but also causes huge economic losses and safety hazards. Traditional boiler monitoring methods often rely on manual inspections and regular testing. This method is not only inefficient, but also difficult to detect and deal with potential wear and explosion risks in a timely manner; With the rapid development of the Internet of Things and big data technology, intelligent and real-time monitoring systems have gradually become an important trend in the industrial field. However, there are still many challenges in the intelligent monitoring system for boiler wear and explosion protection. For example, how to efficiently collect and process sensor data from various parts of the boiler, how to accurately identify and process abnormal data in real time, how to quickly transmit early warning information and take corresponding treatment measures, and how to comprehensively evaluate and judge abnormal situations are all issues that need to be urgently solved. In this regard, we propose an intelligent real-time monitoring system and method for boiler wear and explosion protection. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art, and discloses a boiler anti-wear and explosion-proof intelligent real-time monitoring system and method.

[0004] In a first aspect, an embodiment of the present invention provides a boiler wear and explosion proof intelligent real-time monitoring system, including: a boiler data acquisition module, an edge intelligent processing module, a parallel acceleration computing module, a multi-source communication module and an intelligent early warning module; The boiler data acquisition module is used to collect sensor data from various parts of the boiler at a high frequency; The edge intelligent processing module is electrically connected to the boiler data acquisition module. The edge intelligent processing module is used to process and analyze data on edge devices close to the sensor, use built-in intelligent algorithms to perform preliminary analysis on the collected data in real time, quickly identify and process abnormal data based on preset thresholds and rules, and trigger preliminary warnings on the spot. At the same time, the module filters and pre-processes the data to reduce the amount of data transmitted to the central system and reduce transmission delays. The parallel acceleration computing module is electrically connected to the edge intelligent processing module, and the parallel acceleration computing module is used to use parallel computing technology to decompose complex data analysis tasks into multiple subtasks to be executed simultaneously, quickly perform in-depth analysis on the data transmitted after edge processing, and mine potential abnormal patterns of the data through parallel machine learning algorithms; The multi-communication module is electrically connected to the parallel acceleration computing module. The multi-source communication module is used to construct multiple communication links of different types. Before the early warning information is transmitted, each communication link is evaluated and the optimal link is selected for transmission to the intelligent early warning module and the operation and maintenance personnel. The intelligent early warning module is electrically connected to the multi-communication module. The intelligent early warning module is used to integrate various data and analysis results, use intelligent algorithms and fault models, comprehensively evaluate and judge abnormal situations, and handle them according to preset risk levels and priorities.

[0005] In some possible embodiments, collecting sensor data of various parts of the boiler at a high frequency specifically includes: For the temperature sensor, the heat exchange rate threshold and the collection time interval threshold are set based on the previous data. If the heat exchange rate is higher than the set threshold, the collection is performed when it is lower than the collection time interval threshold; if the heat exchange rate is lower than or equal to the set threshold, the collection is performed when it is higher than the collection time interval threshold; For the pressure sensor, the pressure fluctuation degree threshold and the acquisition frequency threshold are set based on past data. If the pressure fluctuation degree is greater than the set threshold, the acquisition frequency is greater than the acquisition frequency threshold; when the pressure fluctuation degree is less than or equal to the set threshold, the acquisition frequency is less than the acquisition frequency threshold.

[0006] In some possible embodiments, quickly identifying and processing abnormal data according to preset thresholds and rules specifically includes: Temperature data: calculate the mean and standard deviation of the temperature data collected over a period of time, set an abnormal threshold range, and if the collected temperature value exceeds the threshold range, the temperature data is judged to be abnormal; Pressure data, calculate the pressure mean and standard deviation, set the pressure abnormality threshold range, and determine that the pressure data is abnormal if it exceeds this range.

[0007] In some possible embodiments, the parallel acceleration computing module allocates computing resources according to the amount of data and the complexity of the task when using parallel computing technology; Decompose complex data analysis tasks into multiple subtasks and calculate the amount of data that each subtask is expected to process; For each subtask, computing resources are allocated according to its processing complexity. The processing complexity is determined by evaluating the number of mathematical operations and data dependencies within the task, and the number of processor cores is allocated to each subtask according to the complexity.

[0008] In some possible embodiments, evaluating each communication link specifically includes: Link bandwidth, delay, and bit error rate are considered comprehensively. Link bandwidth directly obtains the link technical parameter value. Delay is calculated by sending test data packets to the link and recording the round-trip time. Bit error rate is calculated by sending data bits and counting the number of erroneous data bits. Calculate the comprehensive evaluation value for each link, use the weight allocation method, and set the weight according to actual needs; By comparing the comprehensive evaluation values ​​of each link, the link with the largest evaluation value is selected as the optimal link, and the warning information is transmitted to the intelligent warning module and operation and maintenance personnel.

[0009] In some possible embodiments, the intelligent early warning module uses the analytic hierarchy process when evaluating the risk level of abnormal situations; Determine factors affecting the risk level, including: the severity of abnormal data, the frequency of abnormality, and the degree of impact on boiler operation safety; Construct a judgment matrix and determine the matrix elements by comparing the importance of each factor; Calculate the maximum eigenvalue and eigenvector of the judgment matrix, and normalize the eigenvector to obtain the weight of each factor; According to the quantitative value of each factor, the risk level value is calculated, and different risk levels are divided according to the risk level value.

[0010] In some possible embodiments, the parallel acceleration computing module uses a parallelized random forest algorithm to mine potential abnormal patterns in data; Random forests are composed of multiple decision trees. When constructing each decision tree, some samples are randomly selected with replacement from the edge-processed data set as the training set for the decision tree. For each node split, some features are randomly selected from multiple features, and the optimal feature is selected for splitting through comparison; Each decision tree is constructed in parallel, and finally the classification result of the data is determined through a voting mechanism. If the majority of decision trees judge that a certain data is abnormal, it is determined that the data has a potential abnormal pattern.

[0011] In some possible embodiments, the intelligent early warning module uses a fault tree analysis method when constructing a fault model; The fault condition is taken as the top event, and the direct cause of the pipe burst is set as the intermediate event. Each intermediate event is further decomposed into basic events. The relationship between the events is represented by logic gates. According to the probability of occurrence of each basic event, the probability of occurrence of the intermediate event and the top event is calculated through the logic gate relationship.

[0012] In a second aspect, an embodiment of the present invention provides a method for intelligent real-time monitoring of boiler wear and explosion prevention, which is used to implement the intelligent real-time monitoring system for boiler wear and explosion prevention as described above, and the method includes: Collect sensor data from various parts of the boiler at high frequency; On edge devices close to sensors, built-in intelligent algorithms are used to conduct preliminary analysis of collected data in real time, process abnormal data according to preset thresholds and rules, trigger preliminary warnings, and filter and preprocess data at the same time; Use parallel computing technology to decompose complex data analysis tasks into multiple subtasks for simultaneous execution, conduct in-depth analysis of edge processed data, and mine potential abnormal patterns; Build multiple communication links of different types, evaluate each link before transmitting warning information, and select the best link to transmit to the intelligent warning module and operation and maintenance personnel; Integrate data and analysis results from all aspects, use intelligent algorithms and fault models, and comprehensively evaluate and judge abnormal situations according to preset risk levels and processing priorities.

[0013] In some possible embodiments, screening and preprocessing data specifically include: Data screening: define the data range by timestamp, set parameter importance weights based on boiler anti-wear and explosion-proof expertise and historical data analysis, and screen based on weights; Data preprocessing includes missing value processing, outlier processing and data normalization. Missing value processing uses adjacent known data points for linear interpolation to supplement missing values, and outlier processing is based on the 3σ principle.

[0014] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: The intelligent real-time monitoring system and method for boiler wear and explosion prevention proposed in the embodiment of the present invention can identify abnormal data in real time and trigger early warning by collecting sensor data of various parts of the boiler at a high frequency and performing preliminary data processing and analysis on the edge device, thereby significantly improving the safety and reliability of boiler operation, screening and preprocessing the data, reducing the amount of data transmitted to the central system, reducing transmission delays, and improving data processing efficiency. It can quickly mine potential abnormal patterns in data, and improve the accuracy and timeliness of data analysis. It constructs multiple communication links of different types, evaluates each link before transmitting early warning information, and selects the optimal link for transmission, thereby ensuring the reliable transmission of early warning information. It uses intelligent algorithms and fault models to comprehensively evaluate and judge abnormal situations, and handles them according to preset risk levels and processing priorities, thereby improving the pertinence and effectiveness of fault handling. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The above and other features, advantages and aspects of the embodiments of the present invention will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that components and elements are not necessarily drawn to scale.

[0016] Figure 1 This is a module framework diagram of a boiler wear and explosion proof intelligent real-time monitoring system according to an embodiment of the present invention; Figure 2 This is a flow chart of the intelligent real-time monitoring method for boiler wear and explosion prevention according to an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0018] Unless otherwise specified, the technical terms or scientific terms used in the embodiments of the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. "Including" or "comprising" used in the embodiments of the present invention neither limit the shapes, numbers, steps, actions, operations, components, originals and / or their groups mentioned, nor exclude the appearance or addition of one or more other different shapes, numbers, steps, actions, operations, components, originals and / or their groups, or the addition of these. In addition, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number and order of the indicated technical features. Thus, the features defined as "first" and "second" can explicitly or implicitly include one or more of the features. In the description of the embodiments of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0019] Unless otherwise specifically stated, the relative arrangement of the components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present invention. At the same time, it should be understood that for ease of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship, and the techniques, methods and devices known to ordinary technicians in the relevant fields may not be discussed in detail, but where appropriate, the techniques, methods and devices shown should be considered as part of the authorized specification. In all examples shown and discussed here, any specific other examples may have different values. It should be noted that similar symbols and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.

[0020] In the description of the embodiments of the present invention, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In the embodiments of the present invention, the schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in the embodiments of the present invention and the features of the different embodiments or examples, without contradiction.

[0021] Reference Figure 1 As shown, an embodiment of the present invention relates to an intelligent real-time monitoring system for boiler wear and explosion prevention, which is a comprehensive system integrating multiple functions such as data acquisition, processing, analysis, communication and early warning. Its specific composition includes: boiler data acquisition module, edge intelligent processing module, parallel acceleration computing module, multi-communication module and intelligent early warning module.

[0022] Among them, the boiler data acquisition module is responsible for obtaining first-hand information on the boiler's operating status. It collects sensor data from various parts of the boiler at a high frequency to ensure that every subtle change in the boiler's operation can be captured accurately and timely. These sensors are distributed in various key parts of the boiler, covering the monitoring of multiple important parameters such as temperature, pressure, and flow. They are like sensitive tentacles densely distributed around the boiler, providing a steady stream of basic data for the entire monitoring system.

[0023] The edge intelligent processing module is tightly electrically connected to the boiler data acquisition module. The uniqueness of this module is that it places data processing and analysis on edge devices close to the sensor. Advanced intelligent algorithms are built in here, which can conduct preliminary analysis of the collected data in real time. When faced with massive amounts of data, it can quickly and accurately identify and process abnormal data based on preset thresholds and rules. Once an abnormality is found, it triggers a preliminary warning on the spot to buy valuable time for timely discovery of potential problems. At the same time, it will also screen and pre-process the data, remove redundant and irrelevant data information, greatly reduce the amount of data subsequently transmitted to the central system, and effectively reduce transmission delays, making the data processing process of the entire system more efficient and smooth.

[0024] The parallel acceleration computing module is electrically connected to the edge intelligent processing module. This module makes full use of parallel computing technology to cleverly decompose complex data analysis tasks into multiple subtasks so that they can be executed simultaneously. In this way, it can quickly carry out in-depth analysis of the data transmitted after being processed by the edge intelligent processing module. With the help of parallelized machine learning algorithms, it can deeply explore potential abnormal patterns in the data, just like accurately finding hidden clues in a complex data maze, providing strong support for further accurate judgment of the boiler's operating status.

[0025] The multi-communication module is electrically connected to the parallel acceleration computing module. Its main function is to build multiple different types of communication links to meet the data transmission needs in different scenarios. Before the early warning information is transmitted, the module will conduct a comprehensive and detailed evaluation of each communication link, comprehensively consider the stability of the link, transmission speed, bandwidth and many other factors, and select the best link to efficiently and accurately transmit the early warning information to the intelligent early warning module and operation and maintenance personnel, ensuring the timeliness and reliability of information transmission.

[0026] The intelligent early warning module is closely connected with the multi-communication module. It undertakes the key task of integrating various data and analysis results, and comprehensively evaluates and judges abnormal situations that occur during boiler operation by using intelligent algorithms and fault models. In this process, it will classify and sort abnormal situations according to the preset risk level and processing priority, and provide clear and unambiguous processing suggestions and guidance to operation and maintenance personnel, helping them to take corresponding measures quickly and effectively to ensure the safe and stable operation of the boiler.

[0027] Reference Figure 2 As shown, the embodiment of the present invention also relates to a method for intelligent real-time monitoring of boilers against wear and explosion, which is unique in that it can effectively realize the various functions of the above-mentioned intelligent real-time monitoring system for boilers against wear and explosion. Specifically, the method first collects sensor data from various parts of the boiler at a high frequency, which is the starting point and data basis of the entire monitoring process. By collecting sensor data from various key parts of the boiler at a high frequency, the real-time operating status of the boiler can be fully understood.

[0028] On edge devices close to sensors, the collected data is preliminarily analyzed in real time using built-in intelligent algorithms. This step is crucial, as the algorithm will carefully screen and process the data in strict accordance with preset thresholds and rules. Once abnormal data is found, a preliminary warning will be triggered immediately to gain an advantage for subsequent processing. At the same time, data will be screened and preprocessed in this link to remove invalid data and optimize the data structure, reducing the burden on subsequent data processing.

[0029] Using parallel computing technology, complex data analysis tasks are systematically decomposed into multiple subtasks and executed simultaneously. After edge processing, in-depth analysis is performed to mine potential abnormal patterns. This parallel computing method can make full use of computing resources, greatly improve data analysis efficiency, and accurately understand the problems hidden behind the data.

[0030] Build multiple different types of communication links, and conduct a comprehensive assessment of each link before transmitting warning information. Consider multiple dimensions such as link stability, transmission speed, and reliability, carefully select the optimal link, and transmit warning information to the intelligent warning module and operation and maintenance personnel in a timely and accurate manner to ensure that the information transmission link is foolproof.

[0031] The intelligent early warning module integrates various data and analysis results, uses intelligent algorithms and fault models, and conducts comprehensive evaluation and judgment of abnormal situations according to the preset risk level and processing priority. Through this series of rigorous processes, it provides all-round and multi-level protection for the safe and stable operation of the boiler.

[0032] The key link of collecting sensor data from various parts of the boiler at a high frequency is to have a very detailed and targeted strategy. Specifically, for temperature sensors, the system will analyze the law of heat exchange based on long-term accumulated past data, so as to set a scientific and reasonable heat exchange rate threshold and collection time interval threshold. The heat exchange rate calculation formula is:

[0033] In the formula, R h represents the heat exchange rate, Q represents the amount of heat transferred per unit time, and ∆t represents the time interval; When the heat exchange rate is higher than the set threshold, it means that the heat exchange inside the boiler has undergone relatively active changes. At this time, in order to more accurately capture the details of temperature changes, the system will collect data at a frequency lower than the collection time interval threshold to ensure that the temperature fluctuations can be recorded in a timely and detailed manner. Conversely, if the heat exchange rate is lower than or equal to the set threshold, it indicates that the heat exchange is in a relatively stable state. At this time, the system will appropriately adjust the collection frequency and collect data at a frequency higher than the collection time interval threshold, ensuring the acquisition of valid data while avoiding excessive collection and waste of resources.

[0034] For the pressure sensor, the pressure fluctuation degree threshold and the acquisition frequency threshold are also set based on the previous data. The pressure fluctuation degree calculation formula is:

[0035] In the formula, σ P Indicates the degree of pressure fluctuation, that is, the pressure standard deviation, P irepresents the pressure value collected for the i-th time, represents the average pressure, and n represents the number of acquisitions; When the pressure fluctuation is greater than the set threshold, it means that the internal pressure of the boiler is in an unstable state and there is a greater risk of change. At this time, the system will be alert and set the collection frequency to be greater than the collection frequency threshold so as to monitor pressure changes more frequently and find potential problems in time. When the pressure fluctuation is less than or equal to the set threshold, the pressure is relatively stable, and the system will reduce the collection frequency accordingly to less than the collection frequency threshold, achieving a balance between data collection and resource utilization.

[0036] In terms of quickly identifying and processing abnormal data based on preset thresholds and rules, the system has a rigorous judgment method for different types of data. Taking temperature data as an example, the system will calculate the mean and standard deviation of the temperature data collected over a period of time. The mean calculation formula for temperature data is:

[0037] In the formula, μ T Represents the mean of temperature data, T i represents the temperature value collected for the i-th time; The standard deviation of temperature data is calculated as:

[0038] In the formula, σ T Indicates the standard deviation of temperature data; The mean reflects the average temperature level during the period, and the standard deviation reflects the degree of dispersion of the temperature data. Through the analysis of a large amount of historical data and professional knowledge, the abnormal threshold range is set as follows: If the collected temperature value exceeds the set threshold range, the temperature data can be judged as abnormal, which means that the temperature state of the boiler in this part may deviate from normal operation and needs attention.

[0039] The processing method for pressure data is similar. The pressure mean and standard deviation are calculated first, and the pressure abnormality threshold range is set based on this. Once the collected pressure data exceeds this range, the pressure data is considered abnormal, indicating that the pressure inside the boiler does not meet the normal operating requirements, which may pose a threat to the safe operation of the boiler.

[0040] The parallel acceleration computing module demonstrates a high degree of intelligence and scientificity when using parallel computing technology. It will reasonably and accurately allocate computing resources according to the amount of data and the complexity of the task itself. In specific operations, the complex data analysis task is first decomposed into multiple subtasks in an orderly manner. In this process, the system will calculate in detail the amount of data that each subtask is expected to process, so as to have a clear understanding of the subsequent resource allocation.

[0041] For each subtask, the evaluation of its processing complexity is a key step. This complexity is determined by comprehensively considering multiple factors, including the number of mathematical operations within the task. The more operations, the more complex the task; and data dependencies. If the dependencies between data in the subtasks are complex, it will also increase the difficulty of processing. Through a comprehensive evaluation of these factors, the processing complexity of each subtask is determined. Then, the number of processor cores is reasonably allocated to each subtask based on the complexity. Subtasks with high complexity will be allocated more processor cores to ensure that they can complete the processing efficiently and quickly, while subtasks with lower complexity will be allocated fewer cores to achieve optimal use of resources.

[0042] When evaluating each communication link, the system considers multiple key dimensions. The first is link bandwidth, which directly reflects the link's ability to transmit data. Its specific value is determined by directly obtaining the link's technical parameter values. In terms of latency, the system calculates latency by sending a test data packet to the link and accurately recording the time it takes for the data packet to go back and forth. The expression is:

[0043] In the formula, L represents the delay, RTT represents the data packet and records the round-trip time; The bit error rate is calculated by sending a certain number of data bits and carefully counting the number of erroneous data bits, thereby measuring the accuracy of the link transmitting data.

[0044] After obtaining these key indicators, a comprehensive evaluation value is calculated for each link. A weight distribution method is used here, that is, reasonable weights are set for link bandwidth, delay, and bit error rate according to actual application scenarios and requirements. For example, in scenarios where data real-time requirements are extremely high, a higher weight may be set for delay; and in cases where data accuracy requirements are stringent, the weight of the bit error rate may be relatively large. In this way, the impact of various indicators on link performance is comprehensively considered. Finally, by comparing the comprehensive evaluation values ​​of each link, the link with the largest evaluation value is selected as the optimal link, and the critical warning information is accurately transmitted to the intelligent warning module and operation and maintenance personnel to ensure the reliability and efficiency of information transmission. The expression of the comprehensive evaluation value is:

[0045] In the formula, S represents the comprehensive evaluation value, ω1, ω2, and ω3 represent weights, B represents the link bandwidth, and E represents the bit error rate. The bit error rate is calculated by sending a number of data bits in a certain period of time and counting the number of erroneous data bits. The formula is: , where m represents the number of erroneous data bits; The intelligent early warning module uses the scientific and effective method of hierarchical analysis to evaluate the risk level of abnormal situations. First, the key factors that affect the risk level are clearly determined. These factors include the severity of abnormal data, that is, the degree to which the abnormal data deviates from the normal range. The greater the deviation, the higher the severity; the frequency of abnormality. Frequent abnormalities are obviously more worthy of attention than occasional abnormalities; and the degree of impact on the safety of boiler operation. For example, some abnormal situations may directly threaten the structural safety of the boiler, and their impact is relatively high.

[0046] Next, construct a judgment matrix. In this process, the matrix elements are determined by comparing the importance of each factor. For example, the severity of abnormal data is compared with the frequency of abnormal occurrence to determine which factor is more important in assessing the risk level and how important it is, so as to determine the element value of the corresponding position in the matrix.

[0047] After that, the maximum eigenvalue and eigenvector of the judgment matrix are calculated. This step requires the use of a specific mathematical algorithm to obtain the maximum eigenvalue and the corresponding eigenvector by operating the matrix. Then, the eigenvector is normalized so that the sum of the weights of each factor is 1, thereby obtaining the weight of each factor in the risk level assessment.

[0048] Finally, the risk level value is calculated based on the quantitative value of each factor and the determined weight. Through the analysis of a large amount of historical data and actual operation conditions, reasonable standards are formulated, and different risk levels are divided according to the risk level value, providing clear risk instructions for operation and maintenance personnel so that they can take corresponding treatment measures according to different risk levels.

[0049] The system has a clear and definite process for screening and preprocessing data. In the data screening stage, the data range is accurately defined by timestamps to ensure that the screened data is related to a specific time period and meets the actual analysis needs. At the same time, based on the professional knowledge in the field of boiler wear and explosion prevention and in-depth analysis of a large amount of historical data, the parameter importance weights are set. Different parameters have different degrees of importance in judging whether there are wear and explosion problems in the boiler. For example, some parameters related to the temperature and pressure of key components may have higher weights. The data is screened according to these weights to retain the data that is more valuable for analysis.

[0050] Data preprocessing includes several key steps. The first is missing value processing. When missing values ​​appear in the data, the system uses adjacent known data points for linear interpolation. In this way, according to the changing trend of adjacent data, the missing values ​​are reasonably supplemented to make the data more complete without affecting subsequent analysis. Outlier processing is based on the 3σ principle, that is, when the distance between a data point and the mean exceeds 3 times the standard deviation, it is treated as an outlier and processed. Usually, correction or elimination is used to ensure the accuracy and reliability of the data. In addition, data normalization is performed to unify data of different magnitudes into the same range to facilitate subsequent data analysis and model training.

[0051] The parallel acceleration computing module uses a parallelized random forest algorithm to mine potential abnormal patterns in data. Random forests are composed of many decision trees. When constructing each decision tree, some samples are randomly selected with replacement from the edge-processed data set as the training set for the decision tree. This random sampling method increases the diversity of the model and avoids overfitting of the model.

[0052] For each node split, some features are randomly selected from multiple features, and then the optimal features are selected for splitting by comparing these features. In this process, the contribution of each feature to data classification is considered, and features that can make the data more distinguishable after splitting are selected. Each decision tree is constructed in parallel to make full use of computing resources and improve construction efficiency. Finally, the classification result of the data is determined through a voting mechanism. If the majority of decision trees judge that a certain data is abnormal, it is determined that the data has a potential abnormal pattern. This method combines the judgment results of multiple decision trees and improves the accuracy and reliability of the judgment.

[0053] When constructing the fault model, the intelligent early warning module uses the fault tree analysis method. Fault conditions, such as serious faults such as boiler tube bursts, are taken as top events, and the direct causes of tube bursts, such as tube wall overheating and excessive pressure, are set as intermediate events. Each intermediate event is further decomposed into more basic events, such as equipment aging and sensor failure. The relationship between events is represented by logic gates. For example, the "AND gate" indicates that the simultaneous occurrence of multiple basic events will lead to the occurrence of intermediate events, and the "OR gate" indicates that the intermediate event will be triggered as long as one basic event occurs. According to the probability of each basic event, the probability of the occurrence of intermediate events and top events is calculated through the logic gate relationship. In this way, a fault model with clear hierarchy and rigorous logic is constructed, which provides strong support for accurately judging the probability and cause of faults.

[0054] The use process of the present invention is: high-frequency collection of boiler sensor data, preliminary edge analysis and early warning, parallel computing and deep mining, optimal link transmission and early warning, and comprehensive assessment of abnormal risks.

[0055] In summary, the advantages of the embodiments of the present invention are: by collecting boiler sensor data at high frequency and performing preliminary processing and analysis on edge devices, anomalies are identified in real time and early warnings are triggered, thereby improving boiler safety and reliability. The system screens and preprocesses data to reduce transmission volume, reduce latency, and improve data processing efficiency. Through parallel computing, potential abnormal patterns are mined to improve data analysis accuracy. Multiple communication links are evaluated and the optimal link is selected to ensure reliable transmission of early warning information. Intelligent algorithms are combined with fault models to comprehensively evaluate anomalies, and they are processed according to risk level and priority, thereby improving the pertinence and effectiveness of fault handling.

[0056] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present invention, but the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A boiler anti-wear and explosion-proof intelligent real-time monitoring system, characterized in that: include: Boiler data acquisition module, edge intelligent processing module, parallel acceleration computing module, multi-source communication module and intelligent early warning module; The boiler data acquisition module is used to collect sensor data from various parts of the boiler at a high frequency; The edge intelligent processing module is electrically connected to the boiler data acquisition module. The edge intelligent processing module is used to process and analyze data on edge devices close to the sensor, use built-in intelligent algorithms to perform preliminary analysis on the collected data in real time, quickly identify and process abnormal data based on preset thresholds and rules, and trigger preliminary warnings on the spot. At the same time, the module filters and pre-processes the data to reduce the amount of data transmitted to the central system and reduce transmission delays. The parallel acceleration computing module is electrically connected to the edge intelligent processing module, and the parallel acceleration computing module is used to use parallel computing technology to decompose complex data analysis tasks into multiple subtasks to be executed simultaneously, quickly perform in-depth analysis on the data transmitted after edge processing, and mine potential abnormal patterns of the data through parallel machine learning algorithms; The multi-communication module is electrically connected to the parallel acceleration computing module. The multi-source communication module is used to construct multiple communication links of different types. Before the early warning information is transmitted, each communication link is evaluated and the optimal link is selected for transmission to the intelligent early warning module and the operation and maintenance personnel. The intelligent early warning module is electrically connected to the multi-communication module. The intelligent early warning module is used to integrate various data and analysis results, use intelligent algorithms and fault models, comprehensively evaluate and judge abnormal situations, and handle them according to preset risk levels and priorities.

2. The intelligent real-time monitoring system for boiler wear and explosion prevention according to claim 1 is characterized in that: The data collected by the sensors in various parts of the boiler at a high frequency include: For the temperature sensor, the heat exchange rate threshold and the collection time interval threshold are set based on the previous data. If the heat exchange rate is higher than the set threshold, the collection is performed when it is lower than the collection time interval threshold; if the heat exchange rate is lower than or equal to the set threshold, the collection is performed when it is higher than the collection time interval threshold; For the pressure sensor, the pressure fluctuation degree threshold and the acquisition frequency threshold are set based on past data. If the pressure fluctuation degree is greater than the set threshold, the acquisition frequency is greater than the acquisition frequency threshold; when the pressure fluctuation degree is less than or equal to the set threshold, the acquisition frequency is less than the acquisition frequency threshold.

3. The intelligent real-time monitoring system for boiler wear and explosion prevention according to claim 1 is characterized in that: Rapidly identify and process abnormal data based on preset thresholds and rules, including: Temperature data: calculate the mean and standard deviation of the temperature data collected over a period of time, set an abnormal threshold range, and if the collected temperature value exceeds the threshold range, the temperature data is judged to be abnormal; Pressure data, calculate the pressure mean and standard deviation, set the pressure abnormality threshold range, and determine that the pressure data is abnormal if it exceeds this range.

4. The intelligent real-time monitoring system for boiler wear and explosion prevention according to any one of claims 1 to 3, characterized in that: When using parallel computing technology, the parallel acceleration computing module allocates computing resources according to the amount of data and task complexity; Decompose complex data analysis tasks into multiple subtasks and calculate the amount of data that each subtask is expected to process; For each subtask, computing resources are allocated according to its processing complexity. The processing complexity is determined by evaluating the number of mathematical operations and data dependencies within the task, and the number of processor cores is allocated to each subtask according to the complexity.

5. The intelligent real-time monitoring system for boiler wear and explosion prevention according to any one of claims 1 to 3, characterized in that: The evaluation of each communication link includes: Link bandwidth, delay, and bit error rate are considered comprehensively. Link bandwidth directly obtains the link technical parameter value. Delay is calculated by sending test data packets to the link and recording the round-trip time. Bit error rate is calculated by sending data bits and counting the number of erroneous data bits. Calculate the comprehensive evaluation value for each link, use the weight allocation method, and set the weight according to actual needs; By comparing the comprehensive evaluation values ​​of each link, the link with the largest evaluation value is selected as the optimal link, and the warning information is transmitted to the intelligent warning module and operation and maintenance personnel.

6. The intelligent real-time monitoring system for boiler wear and explosion prevention according to any one of claims 1 to 3, characterized in that: The intelligent early warning module uses the analytic hierarchy process when evaluating the risk level of abnormal situations; Determine factors affecting the risk level, including: the severity of abnormal data, the frequency of abnormality, and the degree of impact on boiler operation safety; Construct a judgment matrix and determine the matrix elements by comparing the importance of each factor; Calculate the maximum eigenvalue and eigenvector of the judgment matrix, and normalize the eigenvector to obtain the weight of each factor; According to the quantitative value of each factor, the risk level value is calculated, and different risk levels are divided according to the risk level value.

7. The intelligent real-time monitoring system for boiler wear and explosion prevention according to any one of claims 1 to 3, characterized in that: The parallel acceleration computing module uses a parallelized random forest algorithm to mine potential abnormal patterns in data; Random forests are composed of multiple decision trees. When constructing each decision tree, some samples are randomly selected with replacement from the edge-processed data set as the training set for the decision tree. For each node split, some features are randomly selected from multiple features, and the optimal feature is selected for splitting through comparison; Each decision tree is constructed in parallel, and finally the classification result of the data is determined through a voting mechanism. If the majority of decision trees judge that a certain data is abnormal, it is determined that the data has a potential abnormal pattern.

8. The intelligent real-time monitoring system for boiler wear and explosion prevention according to any one of claims 1 to 3, characterized in that: When constructing the fault model, the intelligent early warning module uses the fault tree analysis method; The fault condition is taken as the top event, and the direct cause of the pipe burst is set as the intermediate event. Each intermediate event is further decomposed into basic events. The relationship between the events is represented by logic gates. According to the probability of occurrence of each basic event, the probability of occurrence of the intermediate event and the top event is calculated through the logic gate relationship.

9. A method for real-time intelligent monitoring of boiler wear and explosion prevention, characterized in that: Used to implement the boiler anti-wear and explosion-proof intelligent real-time monitoring system according to any one of claims 1 to 8, the method comprises: Collect sensor data from various parts of the boiler at high frequency; On edge devices close to sensors, built-in intelligent algorithms are used to conduct preliminary analysis of collected data in real time, process abnormal data according to preset thresholds and rules, trigger preliminary warnings, and filter and preprocess data at the same time; Use parallel computing technology to decompose complex data analysis tasks into multiple subtasks for simultaneous execution, conduct in-depth analysis of edge processed data, and mine potential abnormal patterns; Build multiple communication links of different types, evaluate each link before transmitting warning information, and select the best link to transmit to the intelligent warning module and operation and maintenance personnel; Integrate data and analysis results from all aspects, use intelligent algorithms and fault models, and comprehensively evaluate and judge abnormal situations according to preset risk levels and processing priorities.

10. The intelligent real-time monitoring method for boiler wear and explosion prevention according to claim 9 is characterized in that: Screening and preprocessing data specifically include: Data screening: define the data range by timestamp, set parameter importance weights based on boiler anti-wear and explosion-proof expertise and historical data analysis, and screen based on weights; Data preprocessing includes missing value processing, outlier processing and data normalization. Missing value processing uses adjacent known data points for linear interpolation to supplement missing values, and outlier processing is based on the 3σ principle.

Images