Method, system and equipment for predicting corrosion rate of water wall tube of boiler and storage medium

Abstract

The embodiment of the invention provides a method, a system and equipment for predicting the corrosion rate of a water wall tube of a boiler and a storage medium. The method comprises the following steps: acquiring corrosion related parameters of the water wall tube, including electrochemical noise, surface strain, medium components and temperature and humidity; performing space-time alignment and fusion processing on the corrosion related parameters to obtain a corrosion state feature matrix; inputting the corrosion state characteristic matrix into a pre-established physical characteristic extraction model, and dynamically predicting the corrosion rate of the water wall tube by adopting a space-time dual-channel network; wherein the physical feature extraction model is pre-established through a physical information neural network, and the space-time dual-channel network comprises a long short-term memory network, a graph convolutional network and a sliding window. According to the embodiment of the invention, various factors influencing corrosion are comprehensively considered, the physical information neural network is introduced to carry out deep autonomous learning, finally, a reliable corrosion rate prediction condition is provided, and dynamic, lasting, intelligent and flexible corrosion state monitoring and prediction of the boiler water wall tube are realized.

Description

Technical Field

[0001] Embodiments of the present disclosure belong to the technical field of boiler water wall tube corrosion rate prediction, and particularly relate to a method, system, device, and storage medium for predicting the corrosion rate of boiler water wall tubes. Background Art

[0002] Currently, thermal power generation still occupies a dominant position in industrial and civil electricity consumption. As an important component of the thermal power generation system, the boiler plays a significant role in providing power for power generation. The furnace water wall is the main heat-absorbing part of the boiler, generally a three-dimensional rectangular space composed of multiple rows of steel pipes. It undertakes important functions such as absorbing heat, cooling the furnace, sealing, and improving combustion efficiency, and is the core component of the boiler.

[0003] Since the furnace water wall is directly in contact with high-temperature flue gas and flames, it faces problems such as high temperature, fly ash erosion, sootblowing erosion, slagging, and ash accumulation. At the same time, the tubes near the burners are prone to high-temperature corrosion, and there may also be steam-water corrosion on the inner wall of the tubes. The working environment is extremely harsh. Due to the concealment, local non-uniformity, and progressive harm degree of corrosion thinning, the harm is extremely great. During shutdown maintenance, due to various coal slags and ash deposits attached to the water wall tubes, the corrosion pits are hidden in them, with concealment, and there is a high risk of undetected leakage. At the same time, the number of water wall tubes is numerous, and there is a risk of corrosion not only near the burners but also at various structural mutation positions, weld positions, and near sootblowers on the water wall. Even in the clamping block area of the platen superheater, corrosion thinning has been detected many times.

[0004] Traditional corrosion detection mainly relies on the detection personnel to manually clean the surface foreign matters and conduct visual inspection. However, the detection efficiency is extremely low, highly dependent on the experience of the detection personnel, it is difficult to comprehensively grasp the corrosion situation of the tubes in the entire furnace, there is a high risk of undetected leakage, and it is extremely difficult to detect the corrosion starting from the inner wall, posing a safety hazard to the safe operation of the boiler. Summary of the Invention

[0005] Embodiments of the present disclosure aim to at least solve one of the technical problems existing in the prior art, and provide a method, system, device, and storage medium for predicting the corrosion rate of boiler water wall tubes.

[0006] One aspect of the present disclosure provides a method for predicting the corrosion rate of boiler water wall tubes, the method comprising:

[0007] Obtaining corrosion-related parameters of the water wall tubes, including electrochemical noise, surface strain, medium composition, and temperature and humidity;

[0008] Performing spatio-temporal alignment and fusion processing on the corrosion-related parameters to obtain a corrosion state feature matrix;

[0009] Input the corrosion state feature matrix into a pre-established physical feature extraction model, and use a spatio-temporal dual-channel network to dynamically predict the corrosion rate of the water wall tube; among them,

[0010] The physical feature extraction model is pre-established through a physics-informed neural network. The spatio-temporal dual-channel network includes a long short-term memory network, a graph convolutional network, and a sliding window.

[0011] Further, the spatio-temporal alignment of the corrosion-related parameters includes:

[0012] Unify the clock sources of the corrosion-related parameters using the Precision Time Protocol, and establish a spatial coordinate mapping relationship for the corrosion-related parameters using laser marking;

[0013] Synchronize the electrochemical noise and the surface strain through time-domain interpolation;

[0014] Perform spatial grid matching between the medium composition and the surface strain.

[0015] Further, the pre-establishment of the physical feature extraction model includes:

[0016] Construct the input layer of the physics-informed neural network with spatial coordinates and time, and construct the hidden layer of the physics-informed neural network using a multi-layer fully connected network activated by the tanh function;

[0017] Forward propagation is used to calculate the predicted value of the concentration field in the data path and the residual of the corrosion kinetics equation in the physical path respectively;

[0018] Define a composite loss function that includes the mean square error of the measured corrosion depth and the residual of the corrosion kinetics equation; among them, the corrosion kinetics equation is constructed based on Faraday's law and the activation polarization equation.

[0019] Further, the dynamic prediction of the corrosion rate of the water wall tube using the spatio-temporal dual-channel network includes:

[0020] Set a 7-day sliding time window and activate model retraining every 6 hours;

[0021] Model the corrosion diffusion process as graph-structured data, and capture the corrosion correlation features of adjacent monitoring points in the physical constraint features through the graph convolutional network;

[0022] Use a long short-term memory network to extract the time-dependent features of the physical constraint features, and fuse them with the spatial features output by the graph convolutional network to obtain the corrosion rate of the water wall tube.

[0023] Further, after dynamically predicting the corrosion rate of the water wall tube using the spatio-temporal dual-channel network, the method further includes:

[0024] Based on the prediction results of the corrosion rate, a multi-objective optimization model is established, and Monte Carlo tree search is used to generate dynamic maintenance decisions.

[0025] Further, the method of establishing a multi-objective optimization model based on the prediction results of the corrosion rate and using Monte Carlo tree search to generate dynamic maintenance decisions includes:

[0026] Establish a multi-objective optimization model with constraint equations for the predicted corrosion rate value, the cost of protective materials, and the economic loss of shutdown as optimization objectives;

[0027] Use Monte Carlo tree search to simulate the long-term benefits of different maintenance strategies, and update the node value function through backpropagation;

[0028] Output dynamic maintenance decisions including the optimal maintenance time window, the selection of protective materials, and the ranking of construction priorities.

[0029] Another aspect of the present disclosure provides a boiler water wall tube corrosion rate prediction system, characterized in that the system includes:

[0030] A parameter acquisition module for acquiring corrosion-related parameters of the water wall tube, including electrochemical noise, surface strain, medium composition, and temperature and humidity;

[0031] An alignment and fusion module for performing spatio-temporal alignment and fusion processing on the corrosion-related parameters to obtain a corrosion state feature matrix;

[0032] A dynamic prediction module for inputting the corrosion state feature matrix into a pre-established physical feature extraction model and dynamically predicting the corrosion rate of the water wall tube using a spatio-temporal dual-channel network; wherein,

[0033] The physical feature extraction model is pre-established through a physical information neural network, and the spatio-temporal dual-channel network includes a long short-term memory network, a graph convolutional network, and a sliding window.

[0034] Further, the system further includes a decision generation module for establishing a multi-objective optimization model based on the prediction results of the corrosion rate and using Monte Carlo tree search to generate dynamic maintenance decisions.

[0035] Another aspect of the present disclosure provides an electronic device, including:

[0036] At least one processor; and,

[0037] A memory communicatively connected to the at least one processor for storing one or more programs, which when executed by the at least one processor, enable the at least one processor to implement the above-mentioned boiler water wall tube corrosion rate prediction method.

[0038] Another aspect of the present disclosure provides a computer-readable storage medium storing a computer program, which when executed by a processor implements the method for predicting the corrosion rate of the boiler water wall tubes described above.

[0039] A method, system, device, and storage medium for predicting the corrosion rate of boiler water wall tubes according to an embodiment of the present disclosure overcome the disadvantages of traditional methods that rely solely on empirical formulas, single-point static monitoring, ignore environmental impacts, and lack guidance on maintenance strategies. By comprehensively considering various factors affecting corrosion and introducing a physics-informed neural network for in-depth autonomous learning, a reliable prediction of the corrosion rate is finally provided, realizing dynamic, persistent, intelligent, and flexible corrosion state monitoring and prediction of boiler water wall tubes. Its advantages are as follows: on the one hand, it is convenient for staff to grasp the corrosion situation of the water wall tubes at any time and improve the safety service coefficient of the boiler; on the other hand, it can promptly alert the staff to take corresponding measures according to the predicted corrosion rate to avoid major economic losses; finally, it can refer to the historical data of corrosion monitoring to provide guidance and reference for maintenance personnel on key detection areas during maintenance, greatly reducing the labor intensity of maintenance personnel and improving maintenance efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 is a schematic flowchart of a method for predicting the corrosion rate of boiler water wall tubes according to an embodiment of the present disclosure;

[0041] Figure 2 is a schematic structural diagram of a federated filter according to another embodiment of the present disclosure;

[0042] Figure 3 is a schematic structural diagram of a system for predicting the corrosion rate of boiler water wall tubes according to another embodiment of the present disclosure;

[0043] Figure 4 is a schematic structural diagram of an electronic device according to another embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] The technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present disclosure. Apparently, the described embodiments are only a part of the embodiments of the present disclosure, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present disclosure without creative efforts shall fall within the scope of protection of the present disclosure.

[0045] In addition, the described features, structures, or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, numerous specific details are provided to give a thorough understanding of the embodiments of the present disclosure. However, those skilled in the art will realize that the technical solutions of the present disclosure may be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, well-known methods, devices, implementations, or operations are not shown or described in detail to avoid obscuring aspects of the present disclosure.

[0046] The flowcharts shown in the drawings are merely illustrative and not necessarily include all the contents and operations / steps, nor are they necessarily executed in the described order. For example, some operations / steps may be decomposed, while some operations / steps may be combined or partially combined, so the actual execution order may change according to the actual situation.

[0047] It should be understood that although terms such as first, second, and third may be used in the present disclosure to describe various components, these components should not be limited by these terms. These terms are used to distinguish one component from another. Therefore, the first component discussed below may be referred to as the second component without departing from the teachings of the concept of the present disclosure. As used in the present disclosure, the term "and / or" includes any one and all combinations of one or more of the associated listed items.

[0048] Those skilled in the art can understand that the drawings are only schematic diagrams of exemplary embodiments, and the modules or processes in the drawings are not necessarily essential for implementing the present disclosure, so they cannot be used to limit the protection scope of the present disclosure.

[0049] The present disclosure will collect corrosion-related parameters (such as temperature, humidity, medium composition, stress and strain, etc.) through multi-sensor fusion, and combine physics-informed neural networks (PINNs) with time series analysis to construct a dynamic prediction of the corrosion rate (unit: mm / a) of the boiler water wall tubes.

[0050] As Figure 1 shown, an embodiment of the present disclosure provides a method for predicting the corrosion rate of boiler water wall tubes, and the method includes:

[0051] Step S1: Obtain corrosion-related parameters of the water wall tubes, including electrochemical noise, surface strain, medium composition, and temperature and humidity.

[0052] Specifically, according to actual requirements, corrosion-related parameters are obtained by deploying a micro-sensor array in key monitoring areas in the furnace, specifically including an electrochemical noise sensor (which can monitor the corrosion current and the fluctuation of the corrosion potential), an optical fiber strain sensor (which can capture the crack propagation on the tube surface based on fiber Bragg gratings), a laser-induced breakdown spectroscopy (LIBS) composition analysis sensor (integrated with a micro-laser and a spectrometer, which can monitor the concentration of corrosion ions such as chlorine and sulfur in the environment), and a MEMS temperature and humidity sensor (which can monitor the local temperature gradient and humidity distribution). The above sensors can be encapsulated with high-temperature-resistant ceramics and are preferably arranged at welds, locations where the medium flow direction changes (such as tees, elbows, etc.), and high-temperature and high-pressure areas. The optical fiber strain sensor can be arranged along the tube wall in a spiral winding manner, and the other sensors can be fixed on the tube wall through magnetic suction bases. The sensors can consider making comprehensive use of the advantages of multiple communication protocols to overcome the signal obstruction of the furnace wall metal and perform wireless transmission, so as to realize multi-modal, real-time, and long-term acquisition and recording of data.

[0053] In some specific embodiments, a three-dimensional CAD model of the boiler furnace can be imported, and the deployment positions of the above sensors can be marked as a visualization means. The electrochemical noise sensor returns corrosion current information, the LIBS composition analysis sensor returns the concentrations of chlorine and sulfur elements in the medium, the optical fiber strain sensor returns the thermal map on the surface of the water wall tube, and the MEMS temperature and humidity sensor returns the current temperature and humidity.

[0054] Step S2: Perform spatio-temporal alignment and fusion processing on the corrosion-related parameters to obtain a corrosion state feature matrix.

[0055] Specifically, due to the differences in time scales and spatial resolutions among multi-modal sensors, such as electrochemical signals (millisecond level) and stress data (second level), the point measurement of the LIBS sensor and the continuous distribution of the optical fiber strain sensor, the embodiments of the present disclosure introduce a Figure 2 federated filter as shown. The precise time protocol is used to unify the clock sources of the corrosion-related parameters, control the time difference, and establish the spatial coordinate mapping relationship of the corrosion-related parameters by embedding laser markers in the sensor array to control the spatial resolution difference. Then, the collected signals are fused and the drift deviation of the sensors is eliminated, that is, the electrochemical noise data and the surface strain data are synchronized by time-domain interpolation, and the medium composition data and the surface strain data are matched by spatial grid, forming a corrosion state feature matrix (including corrosion current, corrosion potential, corrosion ion concentration, etc.).

[0056] It should be noted that for each local filter in the federated filter, the potential-corrosion current with an obvious linear relationship can be processed by Kalman filtering (KF), the non-linear response of the composition and corrosion rate can be processed by particle filtering (PF), the non-stationary strain data can be processed by adaptive sliding window filtering, and the low-frequency useful signals such as temperature and humidity can be processed by low-pass filtering to effectively suppress high-frequency noise.

[0057] Step S3: Input the corrosion state feature matrix into a pre-established physical feature extraction model, and use a spatio-temporal dual-channel network to dynamically predict the corrosion rate of the water wall tube.

[0058] Specifically, the physical feature extraction model is pre-established through a physics-informed neural network (PINNs). Based on Faraday's law, the activation polarization equation, and corrosion kinetics, a partial differential equation constraint layer based on the corrosion electrochemistry principle is constructed, and the physical equation is embedded into the loss function as a regularization term to make the output term of the physical feature extraction model conform to physical laws, and continuous model training is carried out. Specifically, it includes:

[0059] (1) PINNs network architecture: an input layer constructed with spatial coordinates x and time t, a hidden layer constructed with a multi-layer fully connected network activated by the tanh function, and an output layer composed of physical constraint features (concentration field C(x,t) and corrosion current density i(x,t)).

[0060] (2) Forward propagation: data path and physical path. Among them, the data path process is: input (x,t), neural network, output predicted concentration field C pred (x,t); the physical path process is: calculate C pred The derivatives with respect to time and space are calculated and substituted into the partial differential equation to calculate the corrosion kinetics equation residual R(x,t).

[0061] (3) Loss function calculation: total loss = data loss (mean square error of measured corrosion depth) + physical loss (corrosion kinetics equation residual).

[0062] The establishment process of the physical feature extraction model can be specifically represented by the following code:

[0063]

[0064]

[0065] The spatio-temporal dual-channel network includes a Long Short-Term Memory (LSTM) network, a Graph Convolutional Network (GCN), and a sliding window. In the corrosion prediction scenario involved in the embodiments of the present disclosure, the dynamic prediction model needs to combine time series analysis, multi-modal data fusion, physical law constraints, and adaptive learning to continuously optimize the accuracy of corrosion rate prediction to achieve high-precision and high-robustness prediction. Its workflow can be divided into three stages: data integration, model inference, and dynamic optimization. The core is to provide basic support for the dynamic prediction of corrosion rate by designing a spatio-temporal dual-channel network. Specifically, the LSTM network is used to process the time data of sensors, the GCN is used to establish the corrosion diffusion topological relationship on the tube surface, and finally, through the sliding window, the data of the most recent 7 days are retained, and the retraining of the above model is activated at a frequency of 6 hours per time. The dynamic prediction of the corrosion rate of the water-cooled wall tube performed in this way will have an error of less than 0.02 mm / a. In some embodiments, the coordinates of the corrosion area location can be marked in the 3D CAD model of the boiler furnace to facilitate the positioning of the staff.

[0066] The establishment process of the spatio-temporal dual-channel network can be specifically represented by the following code:

[0067]

[0068]

[0069]

[0070] Step S4: Establish a multi-objective optimization model based on the prediction results of the corrosion rate, and use Monte Carlo tree search to generate dynamic maintenance decisions.

[0071] Specifically, first, establish a multi-objective optimization model with optimization objective constraint equations for the predicted value of the corrosion rate, the cost of the protective material, and the economic loss of shutdown; then use Monte Carlo Tree Search (MCTS) to simulate the long-term benefits of different maintenance strategies, and update the node value function through backpropagation; finally, output dynamic maintenance decisions including the optimal maintenance time window, the selection of protective materials, and the construction priority ranking. When the predicted corrosion rate exceeds the safety threshold, trigger an alarm and start Monte Carlo tree search to give work order suggestions based on the current corrosion rate; when the predicted corrosion rate drops below the safety threshold, cancel the alarm and repeat the above steps S1-S4 to monitor and predict the corrosion situation of the boiler water-cooled wall tube for a long time, comprehensively, intelligently, flexibly, and accurately.

[0072] A method for predicting the corrosion rate of boiler water wall tubes according to an embodiment of the present disclosure overcomes the disadvantages of traditional methods that rely solely on empirical formulas, single-point static monitoring, ignore environmental impacts, and lack guidance on maintenance strategies. Through dynamic monitoring with a multi-sensor array, comprehensively considering various factors affecting corrosion, introducing a physics-informed neural network for in-depth autonomous learning, and finally providing a reliable prediction of the corrosion rate and giving reasonable work order suggestions in combination with Monte Carlo tree search, it realizes dynamic, persistent, intelligent, and flexible corrosion state monitoring and prediction of boiler water wall tubes. Its advantages are as follows: on the one hand, it is convenient for staff to keep track of the corrosion situation in key areas of the water wall tubes at any time, improving the safety service factor of the boiler; on the other hand, it can timely warn the staff to take corresponding measures according to the predicted corrosion rate to avoid causing significant economic losses; finally, referring to the historical data of corrosion monitoring, it can provide guidance and reference for the inspection personnel on key inspection areas during the maintenance period, greatly reducing the labor intensity of the inspection personnel and improving the inspection efficiency.

[0073] As Figure 3 shown, another embodiment of the present disclosure provides a system for predicting the corrosion rate of boiler water wall tubes, the system comprising:

[0074] A parameter acquisition module 310, configured to acquire corrosion-related parameters of the water wall tubes, including electrochemical noise, surface strain, medium composition, and temperature and humidity;

[0075] An alignment and fusion module 320, configured to perform spatio-temporal alignment and fusion processing on the corrosion-related parameters to obtain a corrosion state feature matrix;

[0076] A dynamic prediction module 330, configured to input the corrosion state feature matrix into a pre-established physical feature extraction model, and perform dynamic prediction on the corrosion rate of the water wall tubes by using a spatio-temporal dual-channel network; wherein,

[0077] The physical feature extraction model is pre-established through a physics-informed neural network, and the spatio-temporal dual-channel network includes a long short-term memory network, a graph convolutional network, and a sliding window.

[0078] Exemplarily, as Figure 3 shown, the system for predicting the corrosion rate of boiler water wall tubes further includes a decision generation module 340, configured to establish a multi-objective optimization model based on the prediction result of the corrosion rate, and generate a dynamic maintenance decision by using Monte Carlo tree search.

[0079] Specifically, a system for predicting the corrosion rate of boiler water wall tubes according to an embodiment of the present disclosure is used to implement the method for predicting the corrosion rate of boiler water wall tubes described in the above embodiment, and the specific implementation process has been described in detail in the above embodiment and will not be elaborated herein.

[0080] A boiler water-cooled wall tube corrosion rate prediction system according to an embodiment of the present disclosure overcomes the shortcomings of traditional methods that rely solely on empirical formulas, single-point static monitoring, ignore environmental impacts, and lack guidance on maintenance strategies. Through dynamic monitoring of a multi-sensor array, comprehensively considering various factors affecting corrosion, introducing a physics-informed neural network for deep autonomous learning, and finally providing a reliable corrosion rate prediction and giving reasonable work order suggestions in combination with Monte Carlo tree, it realizes dynamic, persistent, intelligent, and flexible corrosion state monitoring and prediction of boiler water-cooled wall tubes. Its advantages are as follows. On the one hand, it is convenient for staff to grasp the corrosion situation of key areas of water-cooled wall tubes at any time and improve the safe service coefficient of the boiler. On the other hand, it can timely warn staff to take corresponding measures according to the predicted corrosion rate to avoid causing significant economic losses. Finally, it can refer to the historical data of corrosion monitoring to provide guidance and reference for maintenance personnel on key detection areas during maintenance, greatly reducing the labor intensity of maintenance personnel and improving the maintenance efficiency.

[0081] As Figure 4 shown, another embodiment of the present disclosure provides an electronic device, including:

[0082] At least one processor 401; and a memory 402 communicatively connected to the at least one processor 401 for storing one or more programs, which, when executed by the at least one processor 401, enable the at least one processor 401 to implement the boiler water-cooled wall tube corrosion rate prediction method described above.

[0083] Among them, the memory 402 and the processor 401 are connected in a bus manner. The bus can include any number of interconnected buses and bridges, and the bus connects various circuits of one or more processors 401 and the memory 402 together. The bus can also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art, so they will not be further described herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be an element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices on the transmission medium. The data processed by the processor 401 is transmitted on the wireless medium through the antenna. Further, the antenna also receives data and transmits the data to the processor 401.

[0084] The processor 401 is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interface, voltage regulation, power management, and other control functions. The memory 402 can be used to store data used by the processor 401 when executing operations.

[0085] Another embodiment of the present disclosure provides a computer-readable storage medium storing a computer program, which when executed by a processor implements the boiler water-cooled wall tube corrosion rate prediction method described above.

[0086] Among them, the computer-readable storage medium may be included in the system or electronic device of the present disclosure, or may exist alone.

[0087] The computer-readable storage medium can be any tangible medium that contains or stores a program. It can be an electrical, magnetic, optical, electromagnetic, infrared, semiconductor system, device, or equipment. More specific examples include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, an optical fiber, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0088] The computer-readable storage medium may also include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Specific examples include, but are not limited to, electromagnetic signals, optical signals, or any suitable combination thereof.

[0089] It can be understood that the above embodiments are merely exemplary embodiments adopted to illustrate the principle of the present disclosure. However, the present disclosure is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present disclosure, and these modifications and improvements are also considered within the protection scope of the present disclosure.

Claims

1. A method for predicting the corrosion rate of boiler water-cooled wall tubes, characterized in that, The method includes: Obtaining corrosion-related parameters of the water wall tube, including electrochemical noise, surface strain, medium composition, and temperature and humidity; Performing spatio-temporal alignment and fusion processing on the corrosion-related parameters to obtain a corrosion state feature matrix; Inputting the corrosion state feature matrix into a pre-established physical feature extraction model, and using a spatio-temporal dual-channel network to dynamically predict the corrosion rate of the water wall tube; wherein, The physical feature extraction model is pre-established through a physics-informed neural network, and the spatio-temporal dual-channel network includes a long short-term memory network, a graph convolutional network, and a sliding window.

2. The method according to claim 1, wherein The spatio-temporal alignment of the corrosion-related parameters includes: Unifying the clock sources of the corrosion-related parameters using the Precision Time Protocol, and establishing a spatial coordinate mapping relationship of the corrosion-related parameters using laser marking; Performing time-domain interpolation synchronization on the electrochemical noise and the surface strain; Performing spatial grid matching on the medium composition and the surface strain.

3. The method according to claim 1, characterized in that The pre-establishment of the physical feature extraction model includes: Constructing an input layer of the physics-informed neural network with spatial coordinates and time, and constructing a hidden layer of the physics-informed neural network using a multi-layer fully connected network activated by the tanh function; Performing forward propagation to calculate the predicted value of the concentration field of the data path and the residual of the corrosion kinetics equation of the physical path respectively; Defining a composite loss function including the mean square error of the measured corrosion depth and the residual of the corrosion kinetics equation; wherein, the corrosion kinetics equation is constructed based on Faraday's law and the activation polarization equation.

4. The method according to claim 1, characterized in that The dynamic prediction of the corrosion rate of the water wall tube using the spatio-temporal dual-channel network includes: Setting a 7-day sliding time window, and activating model retraining every 6 hours; Modeling the corrosion diffusion process as graph-structured data, and capturing the corrosion correlation features of adjacent monitoring points in the physical constraint features through a graph convolutional network; Using a long short-term memory network to extract the time-dependent features of the physical constraint features, and fusing them with the spatial features output by the graph convolutional network to obtain the corrosion rate of the water wall tube.

5. The method according to any one of claims 1 to 4, characterized in that After dynamically predicting the corrosion rate of the water wall tube using the spatio-temporal dual-channel network, the method further includes: Establishing a multi-objective optimization model based on the prediction results of the corrosion rate, and using Monte Carlo tree search to generate dynamic maintenance decisions.

6. The method according to claim 5, characterized in that, The establishment of the multi-objective optimization model based on the prediction results of the corrosion rate and the generation of dynamic maintenance decisions using Monte Carlo tree search includes: Establishing a multi-objective optimization model with the predicted value of the corrosion rate, the cost of the protective material, and the economic loss of shutdown as the optimization objective constraint equations; Using Monte Carlo tree search to simulate the long-term benefits of different maintenance strategies, and updating the node value function through backpropagation; Outputting dynamic maintenance decisions including the optimal maintenance time window, the selection of protective materials, and the construction priority ranking.

7. A corrosion rate prediction system for boiler water wall tubes, characterized in that, The system includes: A parameter acquisition module for acquiring corrosion-related parameters of the water wall tube, including electrochemical noise, surface strain, medium composition, and temperature and humidity; An alignment and fusion module for performing spatio-temporal alignment and fusion processing on the corrosion-related parameters to obtain a corrosion state feature matrix; A dynamic prediction module, configured to input the corrosion state feature matrix into a pre-established physical feature extraction model, and dynamically predict the corrosion rate of the water wall tubes by using a spatio-temporal dual-channel network; wherein, the physical feature extraction model is pre-established through a physics-informed neural network, and the spatio-temporal dual-channel network includes a long short-term memory network, a graph convolutional network, and a sliding window.

8. The system according to claim 7, characterized in that, The system further includes a decision generation module, configured to establish a multi-objective optimization model based on the prediction result of the corrosion rate, and generate a dynamic maintenance decision by using Monte Carlo tree search.

9. An electronic device, characterized in that, Comprising: at least one processor; and a memory communicatively connected to the at least one processor, for storing one or more programs, which, when executed by the at least one processor, enable the at least one processor to implement the method for predicting the corrosion rate of the boiler water wall tubes according to any one of claims 1 to 6.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for predicting the corrosion rate of the boiler water wall tubes according to any one of claims 1 to 6.

Images