Boiler fault diagnosis method and system based on digital twinborn model

Abstract

The invention discloses a boiler fault diagnosis method and system based on a digital twinborn model, and belongs to the technical field of boiler fault monitoring, and the method comprises the steps: building a boiler digital twinborn model according to actual design and operation data; internal combustion and flowing of the boiler are simulated, and a monitoring model is constructed for fault analysis; early warning and diagnosis are carried out on boiler faults according to simulation and analysis results; optimizing the operation strategy and parameter setting of the boiler based on real-time feedback; the monitoring model is optimized regularly according to the collected real-time data, and an optimization result is recorded; the system comprises a model building module, a simulation analysis module, an early warning and diagnosis module, a parameter setting module and an optimization module. According to the invention, high-precision boiler heating surface corrosion and overtemperature detection can be realized, the safety is improved, faults are avoided, the energy loss is reduced, the operation efficiency of the boiler is improved, accidental shutdown is reduced, and the service life of equipment is prolonged.

Description

Technical Field

[0001] The present invention belongs to the technical field of boiler fault monitoring, and in particular relates to a boiler fault diagnosis method and system based on a digital twin model. Background Art

[0002] With the improvement of industrial automation level, the complexity of industrial boiler systems continues to increase. As a key equipment in industrial production, the safety, reliability and efficiency of boilers are crucial to industrial production. However, due to the complex operating environment of boilers, long-term high temperature and high pressure conditions can lead to various faults, such as corrosion of the heating surface, slagging, overtemperature and flame deflection. These faults will not only reduce the operating efficiency of the boiler, but may also cause safety accidents and seriously affect production. Traditional boiler fault diagnosis methods are often based on preset empirical rules or monitoring indicators, which are difficult to accurately reflect the complex operating conditions and the interaction between multiple fault factors; therefore, it is particularly important to invent a boiler fault diagnosis method based on a digital twin model.

[0003] Existing boiler fault diagnosis methods are unable to perform high-precision detection of boiler heating surface corrosion and overheating, which reduces the safety of boiler operation and easily causes the boiler to shut down unexpectedly due to faults, increasing energy loss and shortening equipment life. Summary of the invention

[0004] The purpose of the present invention is to solve the defects existing in the prior art and provide a boiler fault diagnosis method and system based on a digital twin model.

[0005] In order to achieve the above object, the present invention adopts the following technical solution: The boiler fault diagnosis method based on the digital twin model includes: 1) Establish a digital twin model of the boiler based on actual design and operation data; 2) Based on the boiler digital twin model, simulate the internal combustion and flow of the boiler and build a monitoring model to perform fault analysis and obtain simulation and analysis results; 3) Provide early warning and diagnosis of boiler failures based on simulation and analysis results; 4) Optimize boiler operation strategy and parameter settings based on diagnostic results and real-time feedback; 5) After the parameters are set, the monitoring model is optimized regularly based on the collected real-time data and the optimization results are recorded.

[0006] A further improvement of the present invention is that step 1) establishes a boiler digital twin model based on actual design and operation data, including: S1.1: Monitor the boiler operation data collected by each group of sensors, fill in the missing data caused by sensor failure or incomplete data collection by average filling, and then build a set of sliding windows according to the preset size, and repeatedly move the window until it moves to the end of the data, and perform sliding average on the data in the window each time it moves; S1.2: After S1.1 is processed, the abnormal data points of the operating information are identified by the 3-times standard deviation method, and the detected abnormal data points are processed by the interpolation method. Based on the physical model and operating experience of the boiler, the boiler features are extracted from each group of processed data, and the dimension of the input features is reduced by dimensionality reduction technology; S1.3: Obtain the geometric design data of the boiler, and perform a structured analysis of the boiler based on the extracted boiler features, decompose the boiler into different subsystems, construct a three-dimensional geometric model of the boiler, and then obtain the physical parameters of each component in each subsystem of the boiler according to the boiler design manual; S1.4: Based on the combustion information of the boiler collected by the sensor and the physical parameters of each component of the boiler, determine the heating surfaces in the boiler, calculate the heat generation and loss in the combustion process based on the law of conservation of energy, and build a thermodynamic model based on the calculation results. According to the three heat transfer modes of the boiler, namely convection, conduction and radiation, calculate the heat transfer in different areas respectively to establish a digital twin model of the boiler.

[0007] A further improvement of the present invention is that the groups of sensors described in S1.1 specifically include a temperature sensor, a pressure sensor, a flow sensor and a flue gas sensor.

[0008] A further improvement of the present invention is that the subsystem described in S1.3 specifically includes a combustion system, a flue gas flow and heat exchange system, a heating surface system, a steam-water circulation system, a smoke exhaust and air system, a fuel transportation and supply system, an exhaust gas and pollutant treatment system, and a control and monitoring system.

[0009] A further improvement of the present invention is that step 2) based on the digital twin model of the boiler, the combustion and flow inside the boiler are simulated and a monitoring model is constructed, including: S2.1: Set boiler boundary conditions, simulate the mixing of fuel and air inside the boiler, react to generate heat and combustion products through the constructed thermodynamic model, calculate the heat released by combustion through the lower calorific value of the fuel, and calculate the combustion rate according to the preset combustion parameters. Then, simulate the three heat transfer processes of conduction, convection and radiation in the boiler through the heat transfer model; S2.2: Determine the flow direction of the flue gas generated by combustion based on the flue gas flow data collected in real time by the sensor and the boiler design parameters, and build a flow model to monitor the flue gas flow through the fluid continuity equation and the Navier-Stokes equation. Simulate the turbulent characteristics of the flue gas inside the boiler through the k-ε model. S2.3: The flow model receives the simulation data of the thermodynamic model and the heat transfer model, simulates the flue gas flow generated by the boiler combustion, and calculates the movement trajectory of the fuel particles or dust inside the boiler through the Lagrangian model based on the simulated flue gas flow; S2.4: Combine the flow model, thermodynamic model and heat transfer model, and generate the corresponding monitoring model. Discretize the monitoring model equations through the CFD solver, and perform time stepping through explicit or implicit methods according to the preset time step. At the same time, decouple the boiler pressure and velocity fields in steps to update the velocity and pressure fields. Then solve the model and check whether the residuals of each group of variables are lower than the preset values ​​in each iteration process. If so, the iteration ends; otherwise, the iteration continues.

[0010] A further improvement of the present invention is that in step 2), fault analysis is performed to obtain simulation and analysis results, including: S3.1: Unify the CFD simulation data and historical operation data, eliminate the noise and outliers of each group of data, standardize or normalize each group of data, combine the pre-processed CFD simulation data and historical operation data to form a comprehensive data set, and organize each group of data in the comprehensive data set according to time series; S3.2: Based on the Bi-GRU model architecture, a set of initial fault diagnosis models is constructed, and the weight matrix and bias vector parameters of the fault diagnosis model are randomly initialized. The total comprehensive data set is divided into a training set and a test set, and the training set data in the time step is gradually input into the fault diagnosis model according to the preset time step; S3.3: The training set data is forward propagated in the fault diagnosis model. At the same time, the forward GRU layer in the model processes the input data in sequence according to the start time to the end time of each time step, and the reverse GRU layer processes the input data in sequence according to the end time to the start time of each time step, and then calculates the hidden state of the input data at each time step and outputs it through the output layer; S3,4: Calculate the difference between the model output and the true label through the cross entropy loss function, use the back propagation algorithm to calculate the gradient of the difference value calculated by the loss function with respect to the model parameters, and update the model parameters through the Adam algorithm based on the calculated gradient. After each round of iteration, use the validation set to evaluate the performance of the fault diagnosis model, and calculate various performance indicators of the fault diagnosis model on unseen data; S3.5: If the preset indicators are not met, the learning rate, number of hidden units and time step hyperparameters of the fault diagnosis model are adjusted, and the process of forward propagation, loss calculation, back propagation and parameter update is repeated until the preset training rounds are reached. After the training is completed, the weight parameters of the model are saved; S3.6: Collect and preprocess the real-time sensor data and CFD simulation data during boiler operation, organize each set of newly input data in time steps, input the real-time data into the trained fault diagnosis model, and perform forward propagation calculations. During the forward propagation process, the model calculates the forward and reverse hidden states of each time step in turn based on the input data; S3.7: The fault diagnosis model combines the status information of the current time step and the previous and next time steps to make predictions, and transmits the prediction results to the input layer. The output layer then maps the output value to the probability of abnormal boiler operation at the current time step through an activation function. If the output probability exceeds the preset threshold, it is considered that the operating status of the boiler may be abnormal, triggering an abnormal alarm, and recording the specific abnormality type and corresponding operating parameters.

[0011] A further improvement of the present invention is that step 4) optimizes the operation strategy and parameter settings of the boiler based on the diagnosis results and real-time feedback, including: S4.1: According to the simulation and prediction results, based on maximizing boiler efficiency, minimizing emissions and optimizing various performance indicators of heat exchange efficiency, the boiler optimization goal is determined, and multiple candidate solutions, i.e., the initial operating parameter combinations of the boiler, are randomly generated. The optimization body population is initialized, where each group of individuals corresponds to a candidate solution, each parameter combination is simulated, and the fitness value of each group of optimization bodies under the current parameter combination is calculated according to the boiler optimization goal; S4.2: The parameter combination with the highest fitness value in each group of optimization bodies is taken as the current optimal solution, and then the Euclidean distance between the remaining optimization bodies and the optimal solution is detected. If the calculated distance is less than the preset threshold, the positions of the remaining optimization bodies are updated by randomly selecting to surround the prey or spiral motion; S4.3: When the optimization body position update mode is to surround the prey, the randomness of the distance between the optimization body and the optimal solution and the optimization body approach speed are calculated, and the current position of the optimization body is updated based on the randomness and approach speed. When the optimization body position update mode is spiral motion, the exponential decay function and cosine function are used to simulate the optimization body to search in a circle within the specified range of the distance to the optimal solution; S4.4: If the calculated distance is greater than the preset threshold, a random search strategy is adopted to update the positions of the optimization bodies of each group whose distance is greater than the preset threshold according to the positions of other randomly selected optimization bodies. After each round of position update, the fitness value of each optimization body under the current parameter combination is recalculated, and the current optimal solution is updated; S4.5: Repeat the iterative update multiple times until the randomness of the distance between the optimization body and the optimal solution converges to 0, stop the iteration, and use the final optimal solution as the optimized boiler operation parameter combination, and formulate the corresponding boiler operation strategy based on the parameter combination, and update each group of boiler operation parameters at the same time.

[0012] The boiler fault diagnosis system based on the digital twin model includes: Model building module, which builds a digital twin model of the boiler based on actual design and operation data; The simulation and analysis module simulates the internal combustion and flow of the boiler and builds a monitoring model based on the boiler digital twin model to perform fault analysis and obtain simulation and analysis results; Early warning and diagnosis module, which provides early warning and diagnosis of boiler failures based on simulation and analysis results; Parameter setting module, which optimizes the boiler's operating strategy and parameter settings based on diagnostic results and real-time feedback; Optimization module, after parameter setting is completed, the monitoring model is regularly optimized based on the collected real-time data and the optimization results are recorded.

[0013] A further improvement of the present invention is that, in the model building module, a boiler digital twin model is established based on actual design and operation data, including: S1.1: Monitor the boiler operation data collected by each group of sensors, fill in the missing data caused by sensor failure or incomplete data collection by average filling, and then build a set of sliding windows according to the preset size, and repeatedly move the window until it moves to the end of the data, and perform sliding average on the data in the window each time it moves; S1.2: After S1.1 is processed, the abnormal data points of the operating information are identified by the 3-times standard deviation method, and the detected abnormal data points are processed by the interpolation method. Based on the physical model and operating experience of the boiler, the boiler features are extracted from each group of processed data, and the dimension of the input features is reduced by dimensionality reduction technology; S1.3: Obtain the geometric design data of the boiler, and perform a structured analysis of the boiler based on the extracted boiler features, decompose the boiler into different subsystems, construct a three-dimensional geometric model of the boiler, and then obtain the physical parameters of each component in each subsystem of the boiler according to the boiler design manual; S1.4: Based on the combustion information of the boiler collected by the sensor and the physical parameters of each component of the boiler, determine the heating surfaces in the boiler, calculate the heat generation and loss in the combustion process based on the law of conservation of energy, and build a thermodynamic model based on the calculation results. According to the three heat transfer modes of the boiler, namely convection, conduction and radiation, calculate the heat transfer in different areas respectively to establish a digital twin model of the boiler.

[0014] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the boiler fault diagnosis method based on a digital twin model.

[0015] Compared with the prior art, the present invention has at least the following beneficial technical effects: The boiler fault diagnosis method and system based on the digital twin model provided by the present invention constructs a corresponding thermodynamic model, a heat transfer model and a flow model through the boiler operation data collected by each group of sensors, and simulates the boiler combustion, internal heat transfer and flue gas flow through the three groups of models constructed, and calculates the movement trajectory of fuel particles or dust inside the boiler through the Lagrangian model, and solves the three groups of models through the CDF solver, and constructs a set of initial fault diagnosis models based on the Bi-GRU model architecture, and then combines the preprocessed CFD simulation data and historical operation data to form a comprehensive data set, and divides the total comprehensive data set into a training set and a test set. The training set data is forward propagated in the fault diagnosis model to calculate the hidden state of the input data at each time step, and outputs it through the output layer, and the difference between the model output and the true label is calculated through the cross entropy loss function, and the gradient of the difference value calculated by the loss function with respect to the model parameters is calculated using the back propagation algorithm, and the model parameters are updated through the Adam algorithm. After each round of iteration, the verification set is used to evaluate the performance of the fault diagnosis model. If the preset indicators are not met, the adjustment is made. Each set of hyperparameters of the fault diagnosis model is calculated, and the process of forward propagation, loss calculation, back propagation and parameter update is repeated until the preset training round is reached. After the training is completed, the weight parameters of the model are saved, the real-time sensor data and CFD simulation data in the boiler operation are collected and preprocessed, and the newly input data groups are organized in time steps. The real-time data is input into the trained fault diagnosis model for forward propagation calculation. During the forward propagation process, the model calculates the forward and reverse hidden states of each time step in turn according to the input data. The fault diagnosis model predicts the state information of the current time step and the previous and next time steps, and transmits the prediction results to the input layer. After that, the output layer maps the output value to the probability of abnormal boiler operation at the current time step through the activation function. If the output probability exceeds the preset threshold, it is considered that the operating state of the boiler may be abnormal, triggering an abnormal alarm, and recording the specific abnormal type and corresponding operating parameters, which can achieve high-precision boiler heating surface corrosion and over-temperature detection, improve safety, avoid failures, reduce energy loss, improve boiler operation efficiency, reduce unexpected downtime, and extend equipment life. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0017] Figure 1 This is a flow chart of the boiler fault diagnosis method based on the digital twin model of the present invention.

[0018] Figure 2This is a structural block diagram of the boiler fault diagnosis system based on the digital twin model of the present invention. DETAILED DESCRIPTION

[0019] In the following, only some exemplary embodiments are briefly described. As those skilled in the art will appreciate, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and descriptions are considered to be exemplary and non-restrictive in nature.

[0020] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0021] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0022] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0023] Various structural schematic diagrams of the embodiments disclosed in the present invention are shown in the accompanying drawings. These figures are not drawn to scale, and some details are magnified and some details may be omitted for the purpose of clear expression. The shapes of various regions and layers shown in the figures and the relative sizes and positional relationships therebetween are only exemplary, and may deviate in practice due to manufacturing tolerances or technical limitations, and those skilled in the art may additionally design regions / layers with different shapes, sizes, and relative positions according to actual needs.

[0024] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0025] Example 1 The boiler fault diagnosis method based on the digital twin model provided by the present invention includes: 1) Establish a digital twin model of the boiler based on actual design and operation data; 2) Based on the boiler digital twin model, simulate the internal combustion and flow of the boiler and build a monitoring model to perform fault analysis and obtain simulation and analysis results; 3) Provide early warning and diagnosis of boiler failures based on simulation and analysis results; 4) Optimize boiler operation strategy and parameter settings based on diagnostic results and real-time feedback; 5) After the parameters are set, the monitoring model is optimized regularly based on the collected real-time data and the optimization results are recorded.

[0026] Example 2 Reference Figure 1 The boiler fault diagnosis method based on the digital twin model provided by the present invention has the following specific steps: Establish a digital twin model of the boiler based on actual design and operation data.

[0027] Specifically, monitor each group of boiler operation data collected by each group of sensors, and fill in the missing data caused by sensor failure or incomplete data collection through average filling, then build a group of sliding windows according to the preset size, and repeatedly move the window until it moves to the end of the data. Perform sliding average on the data in the window each time it moves. After the processing is completed, identify the abnormal data points of the operation information through the 3 times standard deviation method, and process the detected abnormal data points through the interpolation method. Based on the physical model and operation experience of the boiler, extract the boiler features from each group of processed data, and reduce the dimension of the input features through dimensionality reduction technology. The geometric design data of the boiler is obtained, and a structured analysis of the boiler is performed in combination with the extracted boiler features. The boiler is decomposed into different subsystems, and a three-dimensional geometric model of the boiler is constructed. Then, according to the boiler design manual, the physical parameters of each component in each subsystem of the boiler are obtained. Based on the combustion information of the boiler collected by the sensor and the physical parameters of each component of the boiler, the various heating surfaces in the boiler are determined. Based on the law of conservation of energy, the heat generation and loss during the combustion process are calculated. A thermodynamic model is constructed based on the calculation results. According to the three heat transfer modes of the boiler, namely convection, conduction and radiation, the heat transfer in different areas is calculated respectively to establish the corresponding heat transfer model.

[0028] In this embodiment, the sensors specifically include temperature sensors, pressure sensors, flow sensors and flue gas sensors; the subsystems specifically include combustion systems, flue gas flow and heat exchange systems, heating surface systems, steam-water circulation systems, smoke exhaust and air systems, fuel transportation and supply systems, waste gas and pollutant treatment systems, and control and monitoring systems.

[0029] Based on the digital twin model of the boiler, the internal combustion and flow of the boiler are simulated and a monitoring model is constructed to perform fault analysis and obtain simulation and analysis results.

[0030] Specifically, the boiler boundary conditions are set, and the constructed thermodynamic model is used to simulate the mixing and reaction of fuel and air inside the boiler to generate heat and combustion products. The heat released by combustion is calculated by the low calorific value of the fuel, and the combustion rate is calculated according to the preset combustion parameters. Then, the heat transfer model is used to simulate the three heat transfer processes of heat conduction, convection and radiation in the boiler. The flow direction of the flue gas generated by combustion is determined based on the flue gas flow data collected in real time by the sensor and the boiler design parameters. The flow model is constructed through the fluid continuity equation and the Navier-Stokes equation to monitor the flue gas flow. At the same time, the k-ε model is used to simulate the turbulent characteristics of the flue gas inside the boiler. The flow model receives the thermodynamic model and the heat transfer model The flue gas flow generated by boiler combustion is simulated by using each group of simulation data, and based on the simulated flue gas flow, the movement trajectory of fuel particles or dust inside the boiler is calculated through the Lagrangian model. The flow model, thermodynamic model and heat transfer model are combined to generate a corresponding monitoring model. The monitoring model is discretized by the CFD solver, and time stepping is performed according to the preset time step through explicit or implicit methods. At the same time, the boiler pressure and velocity fields are decoupled by step-by-step iterative decoupling to update the velocity and pressure fields. After that, the model is solved, and in each iteration process, it is checked whether the residual of each group of variables is lower than the preset value. If it is lower than the preset value, the iteration ends; otherwise, the iteration continues.

[0031] Specifically, the CFD simulation data and the historical operation data are uniformly processed to eliminate the noise and outliers of each group of data, and then each group of data is standardized or normalized. The preprocessed CFD simulation data and historical operation data are combined to form a comprehensive data set, and each group of data in the comprehensive data set is organized in time series. Based on the Bi-GRU model architecture, a set of initial fault diagnosis models is constructed, and the weight matrix and bias vector parameters of the fault diagnosis model are randomly initialized. The total comprehensive data set is divided into a training set and a test set, and the training set data in the time step is gradually input into the fault diagnosis model according to the preset time step. The training set data is forward propagated in the fault diagnosis model. At the same time, the forward GRU layer in the model processes the input data in sequence according to the start time to the end time of each time step, and the reverse GRU layer processes the input data in sequence according to the end time to the start time of each time step. Then the hidden state of the input data at each time step is calculated and output through the output layer. The difference between the model output and the true label is calculated by the cross entropy loss function, and the gradient of the difference value calculated by the loss function for the model parameters is calculated using the back propagation algorithm. Based on the calculated gradient, the Ad The am algorithm updates the model parameters. After each round of iteration, the validation set is used to evaluate the performance of the fault diagnosis model, and the performance indicators of the fault diagnosis model on unseen data are calculated. If the preset indicators are not met, the learning rate, number of hidden units, and time step hyperparameters of the fault diagnosis model are adjusted, and the process of forward propagation, loss calculation, back propagation and parameter update is repeated until the preset training round is reached. After the training is completed, the weight parameters of the model are saved, the real-time sensor data and CFD simulation data in the boiler operation are collected and preprocessed, and the newly input data groups are organized in the form of time steps. The real-time data is input into the trained fault diagnosis model for forward propagation calculation. During the forward propagation process, the model calculates the forward and reverse hidden states of each time step in turn according to the input data. The fault diagnosis model combines the state information of the current time step and the previous and next time steps for prediction, and transmits the prediction results to the input layer. After that, the output layer maps the output value to the probability of abnormal boiler operation at the current time step through the activation function. If the output probability exceeds the preset threshold, it is considered that the operating state of the boiler may be abnormal, triggering an abnormal alarm, and recording the specific abnormal type and corresponding operating parameters.

[0032] Example 3 Reference Figure 1 The boiler fault diagnosis method based on the digital twin model provided by the present invention has the following specific steps: Based on the simulation and analysis results, early warning and diagnosis of boiler failures are carried out.

[0033] Optimize boiler operation strategy and parameter settings based on diagnostic results and real-time feedback.

[0034] Specifically, according to the simulation and prediction results, based on the maximization of boiler efficiency, minimization of emissions and optimization of heat exchange efficiency, the boiler optimization target is determined, and multiple candidate solutions are randomly generated, that is, the initial operating parameter combination of the boiler, and the optimization body population is initialized, where each group of individuals corresponds to a candidate solution, and each parameter combination is simulated. The fitness value of each group of optimization bodies under the current parameter combination is calculated according to the boiler optimization target, and the parameter combination with the highest fitness value in the current groups of optimization bodies is taken as the current optimal solution. Then, the Euclidean distance between the remaining optimization bodies and the optimal solution is detected. If the calculated distance is less than the preset threshold, the positions of the remaining optimization bodies are updated by randomly selecting to surround the prey or spiral motion. When the optimization body position update method is to surround the prey, the randomness of the distance between the optimization body and the optimal solution is calculated, and the optimization The approach speed of the optimization body is calculated, and the current position of the optimization body is updated based on randomness and the approach speed. When the position update mode of the optimization body is spiral motion, the exponential decay function and the cosine function are used to simulate the optimization body to search in a circle within the specified range of the distance to the optimal solution. If the calculated distance is greater than the preset threshold, a random search strategy is adopted to update the positions of the optimization bodies of each group whose distance is greater than the preset threshold according to the positions of other randomly selected optimization bodies. After each round of position update, the fitness value of each optimization body under the current parameter combination is recalculated, and the current optimal solution is updated. The iterative update is repeated many times until the randomness of the distance between the optimization body and the optimal solution converges to 0, then the iteration is stopped, and the final optimal solution is used as the optimized boiler operation parameter combination, and the corresponding boiler operation strategy is formulated based on this group of parameter combinations, and the operating parameters of each group of boilers are updated at the same time.

[0035] After the parameters are set, the monitoring model is regularly optimized based on the collected real-time data and the optimization results are recorded.

[0036] Example 4 Reference Figure 2 The boiler fault diagnosis system based on the digital twin model provided by the present invention includes: Model building module, which builds a digital twin model of the boiler based on actual design and operation data; The simulation and analysis module simulates the internal combustion and flow of the boiler and builds a monitoring model based on the boiler digital twin model to perform fault analysis and obtain simulation and analysis results; Early warning and diagnosis module, which provides early warning and diagnosis of boiler failures based on simulation and analysis results; Parameter setting module, which optimizes the boiler's operating strategy and parameter settings based on diagnostic results and real-time feedback; Optimization module, after parameter setting is completed, the monitoring model is regularly optimized based on the collected real-time data and the optimization results are recorded.

[0037] Example 5 The present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps of the boiler fault diagnosis method based on the digital twin model.

[0038] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0039] The present application is described with reference to the flowcharts and / or block diagrams of the methods, systems and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A system that specifies the functions of a box or multiple boxes.

[0040] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0041] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0042] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the attached claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any figure mark in the claims should not be regarded as limiting the claims involved.

[0043] In addition, it should be understood that although this specification is described in accordance with the implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation modes that can be understood by those skilled in the art. The above content is only to illustrate the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution according to the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A boiler fault diagnosis method based on a digital twin model, characterized in that: include: 1) Establish a digital twin model of the boiler based on actual design and operation data; 2) Based on the boiler digital twin model, simulate the internal combustion and flow of the boiler and build a monitoring model to perform fault analysis and obtain simulation and analysis results; 3) Provide early warning and diagnosis of boiler failures based on simulation and analysis results; 4) Optimize boiler operation strategy and parameter settings based on diagnostic results and real-time feedback; 5) After the parameters are set, the monitoring model is optimized regularly based on the collected real-time data and the optimization results are recorded.

2. The boiler fault diagnosis method based on digital twin model according to claim 1 is characterized in that: Step 1) Establish a boiler digital twin model based on actual design and operation data, including: S1.1: Monitor the boiler operation data collected by each group of sensors, fill in the missing data caused by sensor failure or incomplete data collection by average filling, and then build a set of sliding windows according to the preset size, and repeatedly move the window until it moves to the end of the data, and perform sliding average on the data in the window each time it moves; S1.2: After S1.1 is processed, the abnormal data points of the operating information are identified by the 3-times standard deviation method, and the detected abnormal data points are processed by the interpolation method. Based on the physical model and operating experience of the boiler, the boiler features are extracted from each group of processed data, and the dimension of the input features is reduced by dimensionality reduction technology; S1.3: Obtain the geometric design data of the boiler, and perform a structured analysis of the boiler based on the extracted boiler features, decompose the boiler into different subsystems, construct a three-dimensional geometric model of the boiler, and then obtain the physical parameters of each component in each subsystem of the boiler according to the boiler design manual; S1.4: Based on the combustion information of the boiler collected by the sensor and the physical parameters of each component of the boiler, determine the heating surfaces in the boiler, calculate the heat generation and loss in the combustion process based on the law of conservation of energy, and build a thermodynamic model based on the calculation results. According to the three heat transfer modes of the boiler, namely convection, conduction and radiation, calculate the heat transfer in different areas respectively to establish a digital twin model of the boiler.

3. The boiler fault diagnosis method based on digital twin model according to claim 2 is characterized in that: The sensor groups described in S1.1 specifically include a temperature sensor, a pressure sensor, a flow sensor and a flue gas sensor.

4. The boiler fault diagnosis method based on digital twin model according to claim 2 is characterized in that: The subsystems described in S1.3 specifically include the combustion system, flue gas flow and heat exchange system, heating surface system, steam-water circulation system, smoke exhaust and air system, fuel transportation and supply system, waste gas and pollutant treatment system, and control and monitoring system.

5. The boiler fault diagnosis method based on digital twin model according to claim 2 is characterized in that: Step 2) Based on the boiler digital twin model, simulate the combustion and flow inside the boiler and build a monitoring model, including: S2.1: Set boiler boundary conditions, simulate the mixing of fuel and air inside the boiler, react to generate heat and combustion products through the constructed thermodynamic model, calculate the heat released by combustion through the lower calorific value of the fuel, and calculate the combustion rate according to the preset combustion parameters. Then, simulate the three heat transfer processes of conduction, convection and radiation in the boiler through the heat transfer model; S2.2: Determine the flow direction of the flue gas generated by combustion based on the flue gas flow data collected in real time by the sensor and the boiler design parameters, and build a flow model to monitor the flue gas flow through the fluid continuity equation and the Navier-Stokes equation. Simulate the turbulent characteristics of the flue gas inside the boiler through the k-ε model. S2.3: The flow model receives the simulation data of the thermodynamic model and the heat transfer model, simulates the flue gas flow generated by the boiler combustion, and calculates the movement trajectory of the fuel particles or dust inside the boiler through the Lagrangian model based on the simulated flue gas flow; S2.4: Combine the flow model, thermodynamic model and heat transfer model, and generate the corresponding monitoring model. Discretize the monitoring model equations through the CFD solver, and perform time stepping through explicit or implicit methods according to the preset time step. At the same time, decouple the boiler pressure and velocity fields in steps to update the velocity and pressure fields. Then solve the model and check whether the residuals of each group of variables are lower than the preset values ​​in each iteration process. If so, the iteration ends; otherwise, the iteration continues.

6. The boiler fault diagnosis method based on digital twin model according to claim 5 is characterized in that: In step 2), fault analysis is performed to obtain simulation and analysis results, including: S3.1: Unify the CFD simulation data and historical operation data, eliminate the noise and outliers of each group of data, standardize or normalize each group of data, combine the pre-processed CFD simulation data and historical operation data to form a comprehensive data set, and organize each group of data in the comprehensive data set according to time series; S3.2: Based on the Bi-GRU model architecture, a set of initial fault diagnosis models is constructed, and the weight matrix and bias vector parameters of the fault diagnosis model are randomly initialized. The total comprehensive data set is divided into a training set and a test set, and the training set data in the time step is gradually input into the fault diagnosis model according to the preset time step; S3.3: The training set data is forward propagated in the fault diagnosis model. At the same time, the forward GRU layer in the model processes the input data in sequence according to the start time to the end time of each time step, and the reverse GRU layer processes the input data in sequence according to the end time to the start time of each time step, and then calculates the hidden state of the input data at each time step and outputs it through the output layer; S3,4: Calculate the difference between the model output and the true label through the cross entropy loss function, use the back propagation algorithm to calculate the gradient of the difference value calculated by the loss function with respect to the model parameters, and update the model parameters through the Adam algorithm based on the calculated gradient. After each round of iteration, use the validation set to evaluate the performance of the fault diagnosis model, and calculate various performance indicators of the fault diagnosis model on unseen data; S3.5: If the preset indicators are not met, the learning rate, number of hidden units and time step hyperparameters of the fault diagnosis model are adjusted, and the process of forward propagation, loss calculation, back propagation and parameter update is repeated until the preset training rounds are reached. After the training is completed, the weight parameters of the model are saved; S3.6: Collect and preprocess the real-time sensor data and CFD simulation data during boiler operation, organize each set of newly input data in time steps, input the real-time data into the trained fault diagnosis model, and perform forward propagation calculations. During the forward propagation process, the model calculates the forward and reverse hidden states of each time step in turn based on the input data; S3.7: The fault diagnosis model combines the status information of the current time step and the previous and next time steps to make predictions, and transmits the prediction results to the input layer. The output layer then maps the output value to the probability of abnormal boiler operation at the current time step through an activation function. If the output probability exceeds the preset threshold, it is considered that the operating status of the boiler may be abnormal, triggering an abnormal alarm, and recording the specific abnormality type and corresponding operating parameters.

7. The boiler fault diagnosis method based on digital twin model according to claim 6 is characterized in that: Step 4) Optimize the boiler’s operating strategy and parameter settings based on the diagnostic results and real-time feedback, including: S4.1: According to the simulation and prediction results, based on maximizing boiler efficiency, minimizing emissions and optimizing various performance indicators of heat exchange efficiency, the boiler optimization goal is determined, and multiple candidate solutions, i.e., the initial operating parameter combinations of the boiler, are randomly generated. The optimization body population is initialized, where each group of individuals corresponds to a candidate solution, each parameter combination is simulated, and the fitness value of each group of optimization bodies under the current parameter combination is calculated according to the boiler optimization goal; S4.2: The parameter combination with the highest fitness value in each group of optimization bodies is taken as the current optimal solution, and then the Euclidean distance between the remaining optimization bodies and the optimal solution is detected. If the calculated distance is less than the preset threshold, the positions of the remaining optimization bodies are updated by randomly selecting to surround the prey or spiral motion; S4.3: When the optimization body position update mode is to surround the prey, the randomness of the distance between the optimization body and the optimal solution and the optimization body approach speed are calculated, and the current position of the optimization body is updated based on the randomness and approach speed. When the optimization body position update mode is spiral motion, the exponential decay function and cosine function are used to simulate the optimization body to search in a circle within the specified range of the distance to the optimal solution; S4.4: If the calculated distance is greater than the preset threshold, a random search strategy is adopted to update the positions of the optimization bodies of each group whose distance is greater than the preset threshold according to the positions of other randomly selected optimization bodies. After each round of position update, the fitness value of each optimization body under the current parameter combination is recalculated, and the current optimal solution is updated; S4.5: Repeat the iterative update multiple times until the randomness of the distance between the optimization body and the optimal solution converges to 0, stop the iteration, and use the final optimal solution as the optimized boiler operation parameter combination, and formulate the corresponding boiler operation strategy based on the parameter combination, and update each group of boiler operation parameters at the same time.

8. The boiler fault diagnosis system based on the digital twin model is characterized by: include: Model building module, which builds a digital twin model of the boiler based on actual design and operation data; The simulation and analysis module simulates the internal combustion and flow of the boiler and builds a monitoring model based on the boiler digital twin model to perform fault analysis and obtain simulation and analysis results; Early warning and diagnosis module, which provides early warning and diagnosis of boiler failures based on simulation and analysis results; Parameter setting module, which optimizes the boiler's operating strategy and parameter settings based on diagnostic results and real-time feedback; Optimization module, after parameter setting is completed, the monitoring model is regularly optimized based on the collected real-time data and the optimization results are recorded.

9. The boiler fault diagnosis system based on digital twin model according to claim 8 is characterized in that: In the model building module, a boiler digital twin model is established based on actual design and operation data, including: S1.1: Monitor the boiler operation data collected by each group of sensors, fill in the missing data caused by sensor failure or incomplete data collection by average filling, and then build a set of sliding windows according to the preset size, and repeatedly move the window until it moves to the end of the data, and perform sliding average on the data in the window each time it moves; S1.2: After S1.1 is processed, the abnormal data points of the operating information are identified by the 3-times standard deviation method, and the detected abnormal data points are processed by the interpolation method. Based on the physical model and operating experience of the boiler, the boiler features are extracted from each group of processed data, and the dimension of the input features is reduced by dimensionality reduction technology; S1.3: Obtain the geometric design data of the boiler, and perform a structured analysis of the boiler based on the extracted boiler features, decompose the boiler into different subsystems, construct a three-dimensional geometric model of the boiler, and then obtain the physical parameters of each component in each subsystem of the boiler according to the boiler design manual; S1.4: Based on the combustion information of the boiler collected by the sensor and the physical parameters of each component of the boiler, determine the heating surfaces in the boiler, calculate the heat generation and loss in the combustion process based on the law of conservation of energy, and build a thermodynamic model based on the calculation results. According to the three heat transfer modes of the boiler, namely convection, conduction and radiation, calculate the heat transfer in different areas respectively to establish a digital twin model of the boiler.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the boiler fault diagnosis method based on a digital twin model according to any one of claims 1 to 7.

Images