Fault prediction method and system for main heating surface of boiler

Abstract

The invention relates to the technical field of boiler fault prediction, in particular to a boiler main heating surface fault prediction method and system.The boiler main heating surface fault prediction method comprises the steps that collected original data are processed to obtain target data, and the target data are data related to a boiler heating surface; determining an association rule of boiler heating surface failure according to the target data by adopting a target algorithm; quantitative calculation is carried out based on association rules, and a quantitative analysis result of heating surface failure is obtained; and based on a quantitative analysis result, performing fault prediction through a fault tree analysis method. The method has the beneficial effects that downtime and economic loss caused by faults can be effectively reduced, safe and stable operation of the boiler is guaranteed, and remarkable economic benefits and social benefits are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of boiler fault prediction, and particularly to a method and system for predicting faults in the main heating surfaces of a boiler. Background Art

[0002] In the field of thermal power generation, as a core device, the main heating surfaces of a boiler (such as water wall tubes, superheater tubes, reheater tubes, economizer tubes, and boiler suspension tubes, etc.) are in a high-temperature and high-pressure environment for a long time, and are easily damaged by wear, corrosion, creep, etc., resulting in frequent failures. According to statistics, about 40% of the accident shutdowns of thermal power units are caused by boiler failures, and the boiler shutdowns caused by the damage of the heating surfaces account for about 70% of the boiler failures.

[0003] Traditional methods for diagnosing faults in boiler heating surfaces mainly rely on manual experience, and have problems such as strong subjectivity and low accuracy, making it difficult to comprehensively and accurately analyze complex fault situations. In addition, the utilization efficiency of a large amount of operation data and historical data is low, and it is impossible to effectively integrate structured data such as boiler combustion coal quality parameters, design parameters, historical maintenance data, and real-time operation parameters, as well as unstructured data such as pictures and videos, and it is difficult to extract valuable fault correlation information from them. In terms of fault prediction, there is a lack of accurate quantitative calculation models and scientific analysis methods, and often only maintenance can be carried out after a fault occurs, and it is impossible to achieve early prediction and warning, which not only causes huge economic losses, but also poses a serious threat to the personal safety of operating personnel. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides the following technical solutions:

[0005] In the first aspect, the present invention provides a method for predicting faults in the main heating surfaces of a boiler, including processing the collected original data to obtain target data, where the target data is data related to the heating surfaces of the boiler;

[0006] Using a target algorithm to determine the association rules for the failure of the boiler heating surfaces according to the target data;

[0007] Performing quantitative calculation based on the association rules to obtain a quantitative analysis result of the failure of the heating surface;

[0008] Based on the quantitative analysis result, performing fault prediction through the fault tree analysis method.

[0009] As a preferred scheme of the method for predicting faults in the main heating surfaces of the boiler of the present invention, among them: the target algorithm is an improved Apriori algorithm;

[0010] The improved Apriori algorithm includes:

[0011] During the connection process, a hash table is used to store frequent itemsets, improving the connection efficiency and reducing the amount of calculation;

[0012] During the pruning process, a pruning strategy is used to exclude in advance the candidate sets that are unlikely to become frequent itemsets.

[0013] As a preferred solution of the method for predicting faults in the main heating surfaces of a boiler according to the present invention, wherein: using a target algorithm to determine the association rules for the failure of the boiler heating surfaces based on target data includes:

[0014] Setting the minimum support and confidence thresholds;

[0015] Using the improved Apriori algorithm to mine the association rules that meet the minimum support and confidence thresholds from the target data.

[0016] As a preferred solution of the method for predicting faults in the main heating surfaces of a boiler according to the present invention, wherein: performing quantitative calculations based on the association rules to obtain the quantitative analysis results of the failure of the heating surfaces includes:

[0017] According to the association rules, determining the key factors and parameters for the failure of the heating surfaces;

[0018] Selecting a mathematical model, where the mathematical model includes but is not limited to high-temperature creep models and low-temperature corrosion models;

[0019] Using the selected mathematical model to perform quantitative calculations on the failure of the heating surfaces to obtain the quantitative analysis results of the failure.

[0020] As a preferred solution of the method for predicting faults in the main heating surfaces of a boiler according to the present invention, wherein: based on the quantitative analysis results, performing fault prediction through the fault tree analysis method includes:

[0021] Constructing a fault tree and determining the top event and basic events of the fault;

[0022] According to the quantitative analysis results, calculating the occurrence probabilities of each event in the fault tree for fault prediction.

[0023] As a preferred solution of the method for predicting faults in the main heating surfaces of a boiler according to the present invention, wherein: through the fault tree analysis method, performing fault prediction further includes:

[0024] Using a machine learning algorithm to optimize and supplement the logical relationships in the fault tree analysis method.

[0025] In a second aspect, the present invention provides a system for predicting faults in the main heating surfaces of a boiler, including: a data acquisition module for acquiring the operation data of the main heating surfaces of the boiler using an industrial-grade embedded acquisition device, wherein the operation data includes but is not limited to temperature, pressure, flow rate, corrosion rate, and stress parameters;

[0026] A storage module for storing and managing the collected operation data;

[0027] An analysis and prediction module for fault prediction based on operation data;

[0028] A display module for displaying the failure warning information of the main heating surfaces of the boiler, where the failure warning information includes the warning level, fault type, and possible occurrence time.

[0029] As a preferred solution of the fault prediction system for the main heating surfaces of the boiler of the present invention, wherein: the analysis and prediction module includes an expert system unit and a data mining unit; the expert system unit is used to reason and analyze the input data and provide suggestions for fault diagnosis and prediction; the data mining unit is used to mine potential fault patterns and rules.

[0030] In a third aspect, the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the steps of the above method are implemented.

[0031] In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0032] Compared with the prior art, the beneficial effects of the present invention are as follows: First, through the integration of multi-source data, a comprehensive and accurate data foundation is provided for subsequent analysis, ensuring the integrity and consistency of the data; second, by using the improved Apriori algorithm for association rule mining, the patterns and rules of the failure of the boiler heating surface can be quickly and accurately revealed, providing an important basis for fault diagnosis and prediction; third, through the high-temperature creep calculation model and the low-temperature corrosion calculation model for quantitative failure analysis, the failure degree and remaining life of the heating surface can be accurately calculated, improving the accuracy and reliability of the diagnosis; finally, by combining the fault tree analysis method and the machine learning algorithm for fault prediction, not only can the possible faults of the heating surface be judged in advance, but also the logical relationship of the fault tree can be optimized, improving the accuracy and adaptability of the prediction. Brief Description of the Drawings

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0034] Figure 1 It is a flow block diagram of the fault prediction method for the main heating surfaces of the boiler. Detailed Embodiments

[0035] To make the above objects, features, and advantages of the present invention more apparent and understandable, the following provides a detailed description of the specific embodiments of the present invention in conjunction with the accompanying drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present invention.

[0036] Example 1, referring to Figure 1 , which is the first embodiment of the present invention. This embodiment provides a method for predicting faults in the main heat-absorbing surfaces of a boiler, including:

[0037] S1. Process the collected raw data to obtain target data, where the target data is data related to the heat-absorbing surfaces of the boiler. The data related to the heat-absorbing surfaces of the boiler includes, but is not limited to, temperature, pressure, flow rate, corrosion rate, stress real-time monitoring parameters, as well as associated parameters such as the wall temperature of the heat-absorbing surface and the flow rate of steam or water. It also includes the material information of the heat-absorbing surface, manufacturing process data, as well as the number identification data of different heat-absorbing surface pipe segments and the position information in the overall structure of the boiler.

[0038] It should be noted that collecting the original data includes installing various types of sensors near the key parts of the boiler, such as water wall tubes, superheater tubes, reheater tubes, economizer tubes, and boiler suspension tubes. The temperature sensors are high-precision thermocouple or thermal resistance sensors, which can accurately measure the temperature changes of the heating surface, and their measurement accuracy can reach ±0.5°C. The pressure sensors are piezoresistive sensors, which can measure the pressure of steam or water, and the accuracy reaches ±0.1% of the full scale. The flow sensors are electromagnetic flowmeters or turbine flowmeters, which are selected according to different media and flow ranges to ensure the accuracy of flow measurement within ±2%. In addition, corrosion sensors are also installed, such as corrosion monitoring sensors based on the electrochemical principle, which can monitor the corrosion rate of the heating surface metal in real time. The stress sensors are strain gauge sensors, which can measure the stress changes of the heating surface during operation. Further, the data acquisition device uses an industrial-grade embedded device with powerful data processing and storage capabilities. Its built-in microprocessor has high-speed computing capabilities and can quickly process a large amount of collected data. The data acquisition frequency is set to 10 times per second to ensure that the dynamic changes of the heating surface parameters can be captured in time. The acquisition device is also equipped with a large-capacity cache memory. When the network transmission fails, it can cache at least 1 hour of data to achieve the function of resuming data transmission from the breakpoint and ensure the integrity of the data. Furthermore, the collected original data is transmitted to the data integration platform through wired or wireless transmission methods (such as industrial Ethernet, ZigBee, etc.). In the data integration platform, a special data receiving and parsing program is developed, which can identify the data formats transmitted by different sensors and convert them into a unified data format. For unstructured data, such as image and video data, image recognition and video processing software are used for preprocessing. The image recognition software uses a convolutional neural network algorithm based on deep learning to analyze the appearance images of the boiler heating surface, identify key information such as wear marks and deformation characteristics, and convert this information into structured data. The video processing software extracts parameters such as vibration frequency, amplitude, and flue gas flow velocity by analyzing the dynamic changes of the heating surface in the video, further enriching the data information required for fault diagnosis.

[0039] It should be further noted that the original data includes, but is not limited to, boiler combustion coal quality parameters, boiler design parameters, boiler historical maintenance data, real-time operation parameters, and expert experience data. Among them, the sulfur content in the boiler combustion coal quality parameters directly affects the low-temperature corrosion of the boiler's tail heating surface; the temperature, pressure, corrosion rate, and stress in the real-time operation parameters directly reflect the operation status of the heating surface. The heating surface material, pipe diameter, and wall thickness in the boiler design parameters are the basic data for analyzing the performance and faults of the heating surface. The records of heating surface maintenance and replacement in the historical maintenance data are of great significance for judging the health status of the heating surface. The part of expert experience regarding the judgment and handling of heating surface faults also directly serves the diagnostic analysis of the heating surface. Further, although the content of other trace elements in the boiler combustion coal quality parameters does not directly cause heating surface faults, it may affect the stability of the combustion process and the flue gas composition, and thus indirectly affect the working conditions of the heating surface. The overall thermal efficiency and ventilation volume in the boiler design parameters are indirectly related to the heating surface by affecting the combustion and heat transfer processes. The records of the overall operation status of the boiler in the historical maintenance data can assist in analyzing the background conditions for the occurrence of heating surface faults. These indirectly related data provide a broader perspective for comprehensively understanding the operation environment and fault causes of the heating surface.

[0040] Furthermore, it should be noted that the processing of the collected original data includes extracting data related to the boiler heating surface from the integration platform and performing cleaning and preprocessing. Specifically: removing noise and outliers in the data, and using data smoothing algorithms such as the moving average method or median filtering method to smooth the data with large fluctuations; for missing values, linear interpolation or multiple imputation methods are used for supplementation according to the historical trend and correlation of the data.

[0041] S2. Use the target algorithm to determine the association rules for the failure of the boiler heating surface based on the target data.

[0042] Furthermore, the target algorithm is the improved Apriori algorithm;

[0043] The improved Apriori algorithm includes:

[0044] During the connection process, a hash table is used to store frequent item sets, improving the connection efficiency and reducing the computational amount;

[0045] It should be noted that when generating the binomial frequent set, all frequent 1-itemsets containing a certain element can be quickly found through the hash table for connection operations to generate new candidate item sets. For example, when generating the candidate k-item set Ck, the elements in Lk-1 are quickly located in the hash table through the hash function, reducing unnecessary comparison operations and shortening the connection time by about 30%.

[0046] Preferably, compared with the traditional Apriori algorithm, the optimized algorithm can generate candidate item sets faster in the joining step, reducing the time complexity of the algorithm.

[0047] During the pruning process, a pruning strategy is used to exclude candidate sets that cannot become frequent item sets in advance.

[0048] It should be noted that after generating candidate item sets, a pruning strategy is used to exclude candidate sets that cannot become frequent item sets in advance. For example, pruning is performed according to the upper bound of the support of the item set. If the upper bound of the support of a candidate item set is less than the minimum support threshold, it can be pruned in advance and no further calculation is required.

[0049] Preferably, adopting a pruning strategy can reduce the number of candidate item sets that need to be calculated, thereby improving the execution speed of the algorithm. The optimized Apriori algorithm can complete the association rule mining task in a shorter time, improving the mining efficiency.

[0050] Furthermore, the association rules for boiler heating surface failure determined by the target algorithm according to the target data include:

[0051] Set the minimum support and confidence thresholds;

[0052] Using the improved Apriori algorithm, mine association rules from the target data that meet the minimum support and confidence thresholds.

[0053] It should be noted that a minimum support threshold is set (indicating the minimum frequency of an item set appearing in the dataset). For example, setting the minimum support to 0.2 means that only item sets with a frequency greater than or equal to 20% in the dataset are considered frequent item sets. At the same time, a minimum confidence threshold is set (indicating that the confidence of the association rule must be greater than or equal to this threshold). For example, setting the minimum confidence to 0.6 means that only rules with a confidence greater than or equal to 60% are considered valid association rules.

[0054] It further needs to be noted that frequent item sets that meet the minimum support threshold are mined from the target dataset. These frequent item sets contain various combinations and patterns related to heating surface failure. Furthermore, according to the frequent item sets, association rules that meet the minimum confidence threshold are generated. These rules reveal the patterns and laws before, during, and after the heating surface failure. For example, the mined rules may indicate that when the temperature of the heating surface continuously exceeds a certain threshold and the dust content in the flue gas exceeds a certain standard, the probability of tube explosion of the heating surface increases significantly.

[0055] S3. Perform quantitative calculations based on the association rules to obtain the quantitative analysis results of the heating surface failure.

[0056] Furthermore, quantitative calculations are performed based on association rules, and the quantitative analysis results of the heating surface failure include:

[0057] According to the association rules, determine the key factors and parameters of the heating surface failure;

[0058] Select a mathematical model, where the mathematical model includes but is not limited to high-temperature creep models and low-temperature corrosion models;

[0059] It should be noted that for the calculation of the high-temperature creep model, first determine the relevant parameters of the metal material, such as the elastic modulus, yield strength, creep activation energy, etc. of the material. These parameters are obtained through material tests or by referring to relevant material manuals. Then, according to the actual temperature, pressure, and stress conditions borne by the heating surface, use finite element analysis software to calculate the creep strain. During the calculation process, consider the creep characteristics of the metal material under different stress application methods, such as high-temperature compression creep, high-temperature tensile creep, high-temperature bending creep, and high-temperature torsion creep. By simulating the creep process of the heating surface during long-term operation, predict the remaining life of the heating surface. For example, for a certain type of high-temperature superheater tube, it is found through calculation that under the current operating conditions, the creep strain increases at a rate of 0.05% per year, and it is expected that the creep strain will exceed the safety threshold within 5 years, and replacement or repair needs to be carried out in advance; in the calculation of the low-temperature corrosion model, monitor parameters such as the sulfur content in the fuel, sulfur dioxide, sulfur trioxide, and water vapor content in the flue gas in real time. According to these parameters, use the chemical kinetics model to calculate the generation amount and dew point temperature of sulfuric acid vapor. For the boiler's tail heating surface (economizer and air preheater), judge whether the wall temperature is lower than the flue gas dew point by measuring its wall temperature. If the wall temperature is lower than the dew point, then calculate the corrosion rate and corrosion depth of sulfuric acid corrosion according to the heat transfer principle and corrosion chemical reaction kinetics. For example, during a certain period, the sulfur content in the fuel is relatively high, resulting in an increase in the sulfuric acid vapor content in the flue gas. It is found through calculation that the corrosion rate of the economizer reaches 0.1 mm per year, and corresponding anti-corrosion measures need to be taken, such as increasing the soot blowing frequency, adjusting the operating parameters of the air preheater, etc.

[0060] Use the selected mathematical model to perform quantitative calculations on the heating surface failure to obtain the quantitative analysis results of the failure.

[0061] It should be noted that the purpose of the quantitative failure calculation is to predict the failure risk and remaining service life of the heating surface under different operating conditions through precise quantitative analysis of various physical parameters and operating data of the main heating surfaces of the boiler. Its core principle is based on material mechanics, heat transfer, chemical kinetics, and related physical and chemical principles, considering the comprehensive influence of various factors on the heating surface during long-term operation, such as temperature, pressure, corrosion, creep, wear, etc., and converting these factors into quantifiable indicators to evaluate the health status of the heating surface.

[0062] Specifically, 1. High-temperature creep calculation:

[0063] Creep is the slow and irreversible deformation that occurs in materials under high temperature and continuous stress. According to the creep constitutive equation, the creep strain ε can be expressed as a function of time t, stress σ, temperature T, and material constants, and is usually described by an equation in the following form:

[0064] ε = f(σ, T, A, n, Q, t)

[0065] Where: A, n, and Q are constants related to the material, representing the creep coefficient, stress exponent, and activation energy of the material, respectively.

[0066] It should be noted that for different metal materials, these constants can be obtained by fitting material tests and experimental data. For example, for common boiler steels, long-term creep tests are carried out under different temperature and stress conditions, the change of strain with time is measured, and then the values of A, n, and Q are determined by methods such as regression analysis.

[0067] Furthermore, the boiler heating surface will bear various stress states during actual operation, including axial stress, circumferential stress, and radial stress. According to the geometric shape of the boiler and stress analysis, the stresses in each direction are calculated through elastic mechanics formulas:

[0068] Axial stress σ a = (P × D) / (4 × t)

[0069] Circumferential stress σ h = (P × D) / (2 × t)

[0070] Radial stress σ r = -P

[0071] For thin-walled containers, P in the formula is the pressure, D is the pipe diameter, and t is the wall thickness.

[0072] Furthermore, when calculating creep, different stress application methods will have different effects on the creep behavior of materials, so the comprehensive influence of the stress state needs to be considered. For complex stress states, the von Mises equivalent stress may be used for unified description:

[0073]

[0074] Where: τ is the shear stress, which can be ignored in simple cases.

[0075] Substitute the calculated equivalent stress into the creep equation, and consider the creep characteristics of the material at different temperatures to calculate the change of creep strain with time of the material under high temperature and complex stress.

[0076] 2. Low-temperature corrosion calculation:

[0077] Low-temperature corrosion mainly occurs in the tail heating surface of the boiler, usually due to the condensation and corrosion of sulfuric acid vapor. According to the principle of chemical kinetics, the formation of sulfuric acid vapor is related to the sulfur content in the fuel, and the contents of sulfur dioxide (SO2), sulfur trioxide (SO3) and water vapor (H2O) in the combustion products.

[0078] According to the sulfur content of the fuel and the chemical equilibrium equation of the combustion process, calculate the amounts of SO2 and SO3 generated:

[0079] Based on the reaction equilibria of S + O2 → SO2 and 2SO2 + O2 → 2SO3, considering the oxygen content and reaction equilibrium constant during the combustion process, the concentration of SO3 in the flue gas can be calculated.

[0080] The generation amount of sulfuric acid vapor is related to the concentrations of SO3 and water vapor. According to the reaction SO3 + H2O → H2SO4, calculate the dew point temperature T of sulfuric acid vapor through chemical equilibrium and mass transfer principles d 。

[0081] When the wall temperature of the heating surface is lower than T d sulfuric acid will condense on the heating surface and a corrosion reaction will occur. The corrosion rate can be described by the following empirical formula or kinetic equation:

[0082] V = k × [H2SO4] n

[0083] In the formula: V is the corrosion rate, k is the reaction rate constant, n is the reaction order, [H2SO4] is the sulfuric acid concentration. Among them, k and n are usually obtained by fitting experimental data through corrosion tests in the laboratory, placing metal specimens in an environment containing sulfuric acid vapor with different concentrations, measuring the corrosion rate at different temperatures.

[0084] Calculating the wall temperature of the heating surface is the key. According to the basic formula of heat transfer:

[0085] Q = h × A × (T1 - T2)

[0086] In the formula: Q is the heat transfer amount, h is the heat transfer coefficient, A is the heat transfer area, T1 is the flue gas temperature, and T2 is the wall temperature.

[0087] By measuring parameters such as the flue gas temperature, flow rate, and composition, combined with the geometric shape and heat transfer characteristics of the heating surface, calculate the wall temperature of the heating surface, thereby judging whether it is lower than the dew point of sulfuric acid vapor, and further evaluating the risk of low-temperature corrosion.

[0088] 3. Calculation of wall thickness reduction:

[0089] For the wall thickness reduction caused by corrosion, calculate according to the corrosion rate and operation time:

[0090] Δt = V × t

[0091] Where: Δt is the wall thickness reduction, V is the corrosion rate, and t is the operation time.

[0092] Furthermore, for different types of corrosion, such as uniform corrosion, local corrosion, etc., different evaluation methods will be adopted. For local corrosion, the local corrosion depth and area also need to be considered. Through the regular inspection data of the heating surface, key evaluation can be carried out on the areas with severe local corrosion.

[0093] Wear is mainly caused by the erosion of the heating surface by solid particles in the flue gas. According to the fluid mechanics and solid particle erosion theory, the wear amount Δw can be expressed as:

[0094] Δw = k1 × C × v n × t

[0095] Where: k1 is the wear coefficient, C is the particle concentration, v is the particle velocity, n is the velocity exponent, and t is the operation time.

[0096] By setting wear monitoring devices at different positions, measuring the particle concentration and velocity, and combining with the anti-wear performance of the material, the wall thickness reduction caused by wear can be calculated.

[0097] 4. Application of Finite Element Analysis (FEA) in creep calculation:

[0098] Geometric modeling of the structure of the boiler heating surface is carried out, and the mesh is divided to form a finite element model. According to the actual boundary conditions, such as pressure, temperature, and constraint conditions, they are applied to the finite element model.

[0099] For complex geometric shapes and load conditions, a three-dimensional finite element model can more accurately describe the stress and strain distribution. For example, for complex structures such as the bent pipes and headers of the heating surface, tetrahedral or hexahedral elements are used for mesh division.

[0100] In the finite element software, the creep constitutive equation of the material is used as the nonlinear model of the material. The creep equation mentioned above is input into the finite element software as the material property, and through the nonlinear solver of the software, the creep strain at different time steps is calculated.

[0101] In finite element software such as ABAQUS or ANSYS, the time integration algorithm is set, and the accumulation of creep strain over time is calculated step by step, considering different temperature fields and stress fields, to simulate the creep deformation of the heating surface during long-term operation.

[0102] 5. Application of time series analysis in predicting the wall thickness reduction:

[0103] Collect wall thickness measurement data, which can be obtained regularly by means such as ultrasonic thickness measurement and radiographic testing. Arrange the wall thickness data in a time series to form a time series data set.

[0104] Smooth the data to remove noise. The moving average method or exponential smoothing method can be used. For example, using simple moving average, for the time series x(t), its moving average value x′(t) = (x(t) + x(t - 1) + … + x(t - n + 1)) / n.

[0105] Use the ARIMA (Autoregressive Integrated Moving Average) model for prediction. First, perform a stationarity test on the time series data. If it is not stationary, perform differencing to make it stationary. Then determine the order (p, d, q) of the ARIMA model according to the autocorrelation function and partial autocorrelation function.

[0106] Establish the ARIMA(p, d, q) model:

[0107]

[0108] Where: c is a constant, φ and θ are model parameters, and ε is white noise.

[0109] Estimate the model parameters by the least squares method or maximum likelihood estimation method, and predict the future wall thickness reduction amount based on historical data.

[0110] 6. Chemical kinetics calculation algorithm:

[0111] For the chemical reaction equilibrium constant in the combustion and corrosion processes, according to the thermodynamic principle, through the formula:

[0112] ΔG = -RT × ln(K)

[0113] Where: ΔG is the change in Gibbs free energy, R is the gas constant, T is the temperature, and K is the equilibrium constant.

[0114] Calculate ΔG based on the standard enthalpy of formation and standard entropy change of the chemical reaction, and then calculate the equilibrium constant K at different temperatures to determine the equilibrium state of the reaction.

[0115] It should be noted that in the calculation of the formation and corrosion of sulfuric acid vapor, an iterative algorithm is used to solve the nonlinear equations. For example, for the formation reaction of sulfuric acid vapor, given the sulfur content in the fuel, the reaction equilibrium equation is solved iteratively to continuously update the concentrations of SO2, SO3, and H2SO4 until the convergence condition is met to determine the concentration and dew point temperature of sulfuric acid vapor.

[0116] S4. Based on the quantitative analysis results, perform fault prediction through the fault tree analysis method.

[0117] Furthermore, based on the quantitative analysis results, fault prediction by the fault tree analysis method includes:

[0118] Construct a fault tree to determine the top event and basic events of the fault;

[0119] It should be noted that taking the boiler heating surface fault as the top event, factors that may cause the fault such as overheating, corrosion, wear, creep, etc. are used as intermediate events, and these intermediate events are further decomposed into basic events such as too high temperature, excessive dust content in flue gas, poor water quality, etc.

[0120] According to the quantitative analysis results, calculate the occurrence probability of each event in the fault tree for fault prediction.

[0121] It should be noted that through the quantitative calculation results such as wall thickness corrosion and creep loss obtained in real-time online, as well as by means of maintenance experience and other data, continuously fit and improve the logical relationship of the fault tree. Using fault tree analysis software, calculate the occurrence probability and importance of each event. For example, through analysis, it is found that under a certain working condition, the probability of the heating surface fault caused by overheating is 0.3, while the probability of the fault caused by corrosion is 0.2, and the importance of the overheating event is relatively high. Therefore, in fault prevention and diagnosis, the overheating problem should be focused on, and corresponding measures such as optimizing the combustion process and increasing the cooling water volume should be taken.

[0122] It further needs to be explained that according to the set failure conditions, such as the wall thickness reduction of carbon steel pipes being greater than 30%, the wall thickness reduction of alloy steel pipes being greater than 25%, or the calculated remaining life being less than one major overhaul interval, etc., write a program to compare and analyze the real-time data and historical data of the heating surface. When it is found that the parameters of the heating surface meet a certain failure condition, the system automatically issues a warning message to remind the operator to perform maintenance or replacement.

[0123] Furthermore, fault prediction by the fault tree analysis method also includes:

[0124] Adopt machine learning algorithms to optimize and supplement the logical relationship in the fault tree analysis method.

[0125] It should be noted that use the regression analysis model in machine learning algorithms to predict the remaining life of the heating surface. Through learning a large amount of historical data, establish a mathematical model between parameters such as wall thickness reduction, operation time, temperature, pressure, etc. and the remaining life to improve the accuracy of prediction.

[0126] In summary, the beneficial effects of the method for predicting faults in the main heating surfaces of a boiler according to the present invention are as follows: Firstly, through the integration of multi-source data, a comprehensive and accurate data basis is provided for subsequent analysis, ensuring the integrity and consistency of the data; Secondly, by using the improved Apriori algorithm for association rule mining, the patterns and laws of the failure of the boiler heating surfaces can be quickly and accurately revealed, providing an important basis for fault diagnosis and prediction; Thirdly, through the high-temperature creep calculation model and the low-temperature corrosion calculation model for quantitative failure analysis, the degree of failure and the remaining life of the heating surfaces can be accurately calculated, improving the accuracy and reliability of the diagnosis; Finally, by combining the fault tree analysis method and the machine learning algorithm for fault prediction, not only can the possible faults of the heating surfaces be judged in advance, but also the logical relationship of the fault tree can be optimized, improving the accuracy and adaptability of the prediction.

[0127] Embodiment 2 is the second embodiment of the present invention. This embodiment provides a system for predicting faults in the main heating surfaces of a boiler, including a data acquisition module for acquiring the operation data of the main heating surfaces of the boiler by using an industrial-grade embedded acquisition device, where the operation data includes but is not limited to temperature, pressure, flow rate, corrosion rate, and stress parameters.

[0128] It should be noted that the acquisition device adopts a modular design, which is convenient for installation, maintenance, and upgrade. Each acquisition module has independent data acquisition and processing capabilities and can operate stably in a harsh industrial environment. At the same time, the acquisition device also has a self-check function, which can detect the working state of the sensor and the accuracy of data acquisition in real time. Once an abnormality is found, an alarm message will be sent immediately.

[0129] A storage module for storing and managing the acquired operation data.

[0130] It should be noted that the storage module combines a SQLServer database and a PI real-time database. The SQLServer database is used to store structured data such as the historical data, design parameters, and maintenance information of the boiler. By establishing a reasonable data table structure and index, the storage and query efficiency of the data are improved. The PI real-time database is mainly used to store and process the real-time acquired data. By using its patented swing door compression technology and unique secondary filtering technology, the efficient compression and storage of real-time data are realized. Further, in the storage module, a data synchronization program is developed to ensure the data consistency between the SQLServer database and the PI real-time database. When the data in the PI real-time database is updated, the synchronization program will synchronize the updated data to the SQLServer database in a timely manner. At the same time, data encryption technology is used to encrypt sensitive data to protect the security of the data.

[0131] An analysis and prediction module for predicting faults according to the operation data.

[0132] Display module: used to display the failure warning information of the main heating surfaces of the boiler, where the failure warning information includes the warning level, fault type, and possible occurrence time.

[0133] It should be noted that the display module adopts a B / S architecture, and users can access the system through a browser. On the home page, the failure warning information of the main heating surfaces of the boiler is displayed in a prominent manner, including the warning level, fault type, and possible occurrence time, etc.

[0134] Furthermore, the analysis and prediction module includes an expert system unit and a data mining unit; the expert system unit is used to reason and analyze the input data and provide suggestions for fault diagnosis and prediction; the data mining unit is used to mine potential fault patterns and rules.

[0135] It should be noted that the expert system unit is jointly constructed by knowledge engineers and domain experts. By sorting out and summarizing a large number of fault cases, maintenance experiences, and theoretical knowledge, it is transformed into rules and knowledge recognizable by computers. The expert system unit is organized in a way that combines a hierarchical structure and a semantic network, which is convenient for knowledge storage, retrieval, and update. The inference engine adopts a hybrid inference method that combines forward inference and backward inference. According to the known data and problems provided by users, it searches for relevant knowledge and rules in the knowledge base, conducts reasoning and judgment, and obtains solutions; the data mining unit uses algorithms such as association rule mining algorithms and clustering analysis algorithms to mine and analyze the data in the database, and discovers potential fault patterns and rules. For example, through the clustering analysis algorithm, the heating surfaces with similar operating parameters and fault characteristics are grouped together, providing a reference for fault diagnosis and prediction.

[0136] It further needs to be noted that the display module also includes an overhaul scatter unit, an overhaul summary unit, a loss analysis unit, and an over-temperature analysis unit. Among them, the overhaul scatter unit uses maps or charts to display the maintenance history distribution at the specific locations of the heating surfaces, and users can intuitively understand the maintenance times and times of each part. The overhaul summary unit summarizes the overhaul records and displays them in two forms: fault statistics and overhaul records, which is convenient for users to conduct data analysis and management. The loss analysis unit uses 3D image technology to visually display the loss information of the heating surfaces, and users can view the loss situation of each part in detail through operations such as rotation and zoom. The over-temperature analysis unit displays the over-temperature history of each heating surface device and the losses caused under over-temperature conditions in the form of a time series chart, and at the same time provides an over-temperature trend prediction function to help users take measures in advance. The association rules mined by the association analysis and prediction unit provide specific guidance for users in operation, such as how to adjust the operating conditions when certain parameters change, etc.

[0137] During use, a large amount of boiler operation data is collected, including normal operation data and fault data. These data cover different working conditions, fuel types, and operation times to ensure the diversity and representativeness of the test data. At the same time, domain experts are invited to annotate and analyze some of the data to provide a reference for the evaluation of the test results.

[0138] The system is tested using the cross-validation method. The test data is divided into a training set and a test set, which are divided according to a certain ratio (such as 8:2). The training set is used to train and optimize the system, and then the test set is used to evaluate the performance of the system. The main test indicators include the accuracy rate of fault diagnosis, the lead time of fault prediction, the false alarm rate, and the missed alarm rate, etc.

[0139] Based on the test results, the system is analyzed and optimized. If the accuracy rate of fault diagnosis is low, check the parameter settings of the association rule mining algorithm, the accuracy of the failure diagnosis calculation model, and the perfection of the knowledge base of the expert system, etc., and adjust and improve the parts with problems. If the false alarm rate is high, re-evaluate the threshold setting of the fault prediction model and the data preprocessing method to reduce unnecessary alarms. Through continuous testing and optimization, the performance and reliability of the system are improved to enable it to meet the requirements of practical applications.

[0140] Embodiment 3 is the third embodiment of the present invention. What is different from the previous two embodiments is:

[0141] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that makes a contribution to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. And the aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0142] The logic and / or steps represented in the flowchart or otherwise described herein can, for example, be considered as a definable list of executable instructions for implementing logical functions, which can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with the instruction execution system, apparatus, or device.

[0143] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which a program can be printed, as the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or other suitable processing as necessary, and then stored in a computer memory.

[0144] It should be understood that the various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), and the like.

[0145] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A method for predicting failure of a main heating surface of a boiler, characterized in that: include, Processing the collected raw data to obtain target data, wherein the target data is data related to the boiler heating surface; Adopting the target algorithm to determine the association rules of boiler heating surface failure according to the target data; Quantitative calculation is performed based on association rules to obtain quantitative analysis results of heating surface failure; Based on the quantitative analysis results, fault prediction is carried out through fault tree analysis method.

2. The method for predicting failure of the main heating surface of a boiler according to claim 1, characterized in that: The target algorithm is an improved Apriori algorithm; The improved Apriori algorithm includes: During the connection process, hash tables are used to store frequent itemsets to improve connection efficiency and reduce the amount of calculation; During the pruning process, pruning strategies are used to exclude candidate sets that are unlikely to become frequent item sets in advance.

3. The method for predicting failure of the main heating surface of a boiler according to claim 2, characterized in that: The association rules for determining the failure of the boiler heating surface according to the target data using the target algorithm include: Set minimum support and confidence thresholds; The improved Apriori algorithm is used to mine association rules that meet the minimum support and confidence thresholds from the target data.

4. The method for predicting failure of the main heating surface of a boiler according to claim 3, characterized in that: The quantitative analysis results of the failure of the heating surface obtained by performing quantitative calculation based on association rules include: According to the association rules, the key factors and parameters of the failure of the heating surface are determined; Selecting a mathematical model, wherein the mathematical model includes but is not limited to a high temperature creep model and a low temperature corrosion model; The selected mathematical model is used to perform quantitative failure calculation on the heated surface and obtain quantitative analysis results of the failure.

5. The method for predicting failure of the main heating surface of a boiler according to any one of claims 1 to 4, characterized in that: The fault prediction based on the quantitative analysis results by the fault tree analysis method includes: Construct a fault tree to determine the top event and basic event of the fault; According to the quantitative analysis results, the occurrence probability of each event in the fault tree is calculated to predict the fault.

6. The method for predicting failure of the main heating surface of a boiler according to claim 5, characterized in that: The method of performing fault prediction by using the fault tree analysis method further includes: Machine learning algorithms are used to optimize and supplement the logical relationships in the fault tree analysis method.

7. A boiler main heating surface fault prediction system, characterized in that: include: A data acquisition module, used to collect the operating data of the main heating surface of the boiler using industrial-grade embedded acquisition equipment, wherein the operating data includes but is not limited to temperature, pressure, flow, corrosion rate and stress parameters; A storage module, used for storing and managing the collected operation data; Analysis and prediction module, used to predict faults based on operating data; Display module: used to display the failure warning information of the main heating surface of the boiler, where the failure warning information includes the warning level, fault type and possible occurrence time.

8. The boiler main heating surface fault prediction system according to claim 7, characterized in that: The analysis and prediction module includes an expert system unit and a data mining unit; the expert system unit is used to reason and analyze input data and provide suggestions for fault diagnosis and prediction; the data mining unit is used to mine potential fault modes and laws.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images