Method and device for predicting corrosion behavior of iron-chromium alloy in closed environment based on improved LSTM (Long Short Term Memory)

Abstract

The invention discloses an improved LSTM (Long Short Term Memory)-based method and device for predicting a corrosion behavior of an iron-chromium alloy in a closed environment. The method comprises the following steps: acquiring corrosion dynamics experimental data of the iron-chromium alloy; based on the improved LSTM model, constructing an iron-chromium alloy corrosion model; the corrosion kinetics experimental data are input into the ferrochromium alloy corrosion model, and corrosion kinetics prediction data are obtained; based on the improved LSTM model, combining physical constraint conditions to construct an iron-chromium alloy oxidation weight gain prediction model; obtaining a weight change prediction value and an oxidation rate prediction value of the iron-chromium alloy through the iron-chromium alloy oxidation weight increment prediction model; and obtaining the corrosion behavior of the ferrochromium alloy according to the corrosion kinetics prediction data, the weight change prediction value and the oxidation rate prediction value. The method can improve the accuracy of iron-chromium alloy corrosion behavior prediction, and can be widely applied to the technical field of machine learning.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of machine learning, and particularly relates to a closed environment iron-chromium alloy corrosion behavior prediction method and device based on an improved LSTM. BACKGROUND

[0002] Iron-chromium alloy is widely used in nuclear energy, aerospace and other fields due to its excellent corrosion resistance, good mechanical properties, rich raw material sources and low cost, and plays an irreplaceable role. However, during long-term service, iron-chromium alloy is prone to corrosion reaction with environmental media such as hydrogen, oxygen and water vapor, which leads to material performance degradation, service life shortening, and even causes serious problems such as structural failure. In addition, the traditional method for studying the corrosion behavior of iron-chromium alloy highly depends on theoretical modeling and expert experience, and has certain limitations in data processing efficiency and model universality. SUMMARY

[0003] Therefore, the embodiments of the present application mainly aim to provide a closed environment iron-chromium alloy corrosion behavior prediction method and device based on an improved LSTM, so as to solve at least one of the problems in the prior art, and the present application can improve the accuracy of iron-chromium alloy corrosion behavior prediction.

[0004] To achieve the above-mentioned purpose, one aspect of the embodiments of the present application provides a closed environment iron-chromium alloy corrosion behavior prediction method based on an improved LSTM, which comprises the following steps:

[0005] Obtaining corrosion kinetics experimental data of iron-chromium alloy;

[0006] Based on the improved LSTM model, an iron-chromium alloy corrosion model is constructed;

[0007] The corrosion kinetics experimental data is input into the iron-chromium alloy corrosion model to obtain corrosion kinetics prediction data;

[0008] Based on the improved LSTM model, an iron-chromium alloy oxidation weight gain prediction model is constructed in combination with physical constraint conditions;

[0009] The weight change prediction value and the oxidation rate prediction value of the iron-chromium alloy are obtained through the iron-chromium alloy oxidation weight gain prediction model;

[0010] According to the corrosion kinetics prediction data, the weight change prediction value and the oxidation rate prediction value, the corrosion behavior of the iron-chromium alloy is obtained.

[0011] In some embodiments, the step of obtaining corrosion kinetics experimental data of iron-chromium alloy comprises the following steps:

[0012] Preparation of the iron-chromium alloy;

[0013] The iron-chromium alloy is subjected to a closed environment corrosion simulation experiment under a single corrosion gas condition to obtain the corrosion kinetics experimental data.

[0014] In some embodiments, the improved LSTM model is used to construct an iron-chromium alloy corrosion model, including the following steps:

[0015] The LSTM model is improved through a double convolution structure, a pooling layer, an LSTM layer, a Dropout layer, and a fully connected layer to obtain the iron-chromium alloy corrosion model.

[0016] In some embodiments, the corrosion kinetics experimental data is input into the iron-chromium alloy corrosion model to obtain corrosion kinetics prediction data, including the following steps:

[0017] The corrosion kinetics experimental data is input into the iron-chromium alloy corrosion model.

[0018] The local time sequence features of the corrosion kinetics experimental data are extracted through the double convolution structure to obtain first features.

[0019] The first features are down-sampled through the pooling layer to obtain second features.

[0020] The potential long-term dependency relationships in the second features are captured through the LSTM layer to obtain third features.

[0021] The third features are mapped into the corrosion kinetics prediction data through the fully connected layer.

[0022] In some embodiments, the improved LSTM model is used to construct an iron-chromium alloy oxidation weight gain prediction model in combination with physical constraint conditions, including the following steps:

[0023] A first constraint condition is constructed based on gas pressure, gas volume, gas amount of substance, absolute temperature of gas, and gas constant.

[0024] A second constraint condition is constructed based on weight change of the iron-chromium alloy, exposed area of the iron-chromium alloy, and corrosion time of the iron-chromium alloy.

[0025] The improved LSTM model is used to obtain the iron-chromium alloy oxidation weight gain prediction model in combination with the first constraint condition and the second constraint condition.

[0026] In some embodiments, the improved LSTM model is used to construct an iron-chromium alloy oxidation weight gain prediction model in combination with physical constraint conditions, and the formula used includes:

[0027] PV = nRT

[0028]

[0029] In the formula, P represents the gas pressure; V represents the gas volume; n represents the amount of gas substance; R represents the gas constant; T represents the absolute temperature of the gas; r represents the oxidation rate of the iron-chromium alloy; Δm represents the weight change of the iron-chromium alloy; A represents the exposed area of the iron-chromium alloy; and Δt represents the corrosion time of the iron-chromium alloy.

[0030] To achieve the above object, another aspect of the embodiment of the present application provides a closed environment iron-chromium alloy corrosion behavior prediction device based on an improved LSTM, which comprises:

[0031] A first module is configured to acquire corrosion kinetics experimental data of an iron-chromium alloy.

[0032] A second module is configured to construct an iron-chromium alloy corrosion model based on an improved LSTM model.

[0033] A third module is configured to input the corrosion kinetics experimental data into the iron-chromium alloy corrosion model to obtain corrosion kinetics prediction data.

[0034] A fourth module is configured to construct an iron-chromium alloy oxidation weight gain prediction model based on the improved LSTM model in combination with physical constraint conditions.

[0035] A fifth module is configured to acquire a weight change prediction value and an oxidation rate prediction value of the iron-chromium alloy through the iron-chromium alloy oxidation weight gain prediction model.

[0036] A sixth module is configured to obtain the corrosion behavior of the iron-chromium alloy according to the corrosion kinetics prediction data, the weight change prediction value, and the oxidation rate prediction value.

[0037] To achieve the above object, another aspect of the embodiment of the present application provides an electronic device, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the above method when executing the computer program.

[0038] To achieve the above object, another aspect of the embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program implements the above method when executed by a processor.

[0039] To achieve the above object, another aspect of the embodiment of the present application provides a computer program product or computer program, which comprises computer instructions stored in a computer readable storage medium. A processor of a computer device can read the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to make the computer device execute the aforementioned method.

[0040] The embodiment of the present application at least has the following beneficial effects: the present application provides an improved LSTM-based closed environment iron-chromium alloy corrosion behavior prediction method and device, which obtains corrosion kinetics experimental data of the iron-chromium alloy; constructs an iron-chromium alloy corrosion model based on an improved LSTM model; inputs the corrosion kinetics experimental data into the iron-chromium alloy corrosion model to obtain corrosion kinetics prediction data; constructs an iron-chromium alloy weight gain prediction model based on the improved LSTM model combined with physical constraint conditions; obtains weight change prediction values and oxidation rate prediction values of the iron-chromium alloy through the iron-chromium alloy weight gain prediction model; and obtains the corrosion behavior of the iron-chromium alloy according to the corrosion kinetics prediction data, the weight change prediction values and the oxidation rate prediction values, which can improve the accuracy of the iron-chromium alloy corrosion behavior prediction. BRIEF DESCRIPTION OF DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0042] Figure 1 is a flowchart of the improved LSTM-based closed environment iron-chromium alloy corrosion behavior prediction method provided by the embodiment of the present application;

[0043] Figure 2 is a schematic diagram of the network model structure of the improved LSTM provided by the embodiment of the present application;

[0044] Figure 3 is a schematic diagram of the change of the hydrogen content of Fe-12.5Cr in a closed environment with time predicted by the physical experimental values and the improved LSTM model provided by the embodiment of the present application;

[0045] Figure 4 is a schematic diagram of the change of the oxygen content of Fe-12.5Cr in a closed environment with time predicted by the physical experimental values and the improved LSTM model provided by the embodiment of the present application;

[0046] Figure 5is a schematic diagram of the change of relative humidity of Fe-12.5Cr with time in a closed environment predicted by the physical experimental value and the improved LSTM model provided by the embodiment of the present application;

[0047] Figure 6 is a schematic diagram of the change of weight of Fe-12.5Cr alloy reacting with oxygen with time predicted by the improved LSTM model provided by the embodiment of the present application;

[0048] Figure 7 is a schematic diagram of the change of weight of Fe-12.5Cr alloy reacting with water vapor with time predicted by the improved LSTM model provided by the embodiment of the present application;

[0049] Figure 8 is a schematic diagram of the change of oxidation rate of Fe-12.5Cr alloy in an oxygen corrosion environment with time predicted by the improved LSTM model provided by the embodiment of the present application;

[0050] Figure 9 is a schematic diagram of the change of oxidation rate of Fe-12.5Cr alloy in a water vapor corrosion environment with time predicted by the improved LSTM model provided by the embodiment of the present application;

[0051] Figure 10 is a schematic diagram of the processing steps of the corrosion behavior prediction of iron-chromium alloy based on the improved LSTM provided by the embodiment of the present application;

[0052] Figure 11 is a schematic diagram of the digital evaluation and prediction platform of the corrosion behavior of iron-chromium alloy provided by the embodiment of the present application;

[0053] Figure 12 is a schematic diagram of the hardware structure of the electronic device provided by the embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application, but are only examples of devices and methods consistent with some aspects of the embodiments of the present application as detailed in the appended claims.

[0055] It should be noted that although the functional modules are divided in the system schematic diagram, and the logical order is shown in the flowchart, in some cases, the steps shown or described can be performed in a manner different from the module division in the system or the order in the flowchart. The terms "first / S100", "second / S200" in the description and claims and the above-mentioned drawings can be used herein to describe various concepts, but unless specifically stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of the present application, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as the first information. Depending on the context, the word "if" as used herein can be interpreted as "when" or "upon" or "in response to determining".

[0056] The terms "at least one", "multiple", "each", "any" and the like used herein include one, two or more than two, multiple includes two or more than two, each refers to each of the corresponding multiple, and any refers to any one of the multiple.

[0057] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0058] As Figure 1 shown, the embodiments of the present application provide a closed environment iron-chromium alloy corrosion behavior prediction based on improved LSTM, which can include but is not limited to steps S100 to S600:

[0059] Step S100, acquiring corrosion kinetics experimental data of iron-chromium alloy;

[0060] Step S200, constructing an iron-chromium alloy corrosion model based on an improved LSTM model;

[0061] Step S300, inputting the corrosion kinetics experimental data into the iron-chromium alloy corrosion model to obtain corrosion kinetics prediction data;

[0062] Step S400, constructing an iron-chromium alloy oxidation weight gain prediction model based on an improved LSTM model combined with physical constraint conditions;

[0063] Step S500, acquiring the weight change prediction value and the oxidation rate prediction value of the iron-chromium alloy through the iron-chromium alloy oxidation weight gain prediction model;

[0064] Step S600, according to the corrosion kinetics prediction data, the weight change prediction value and the oxidation rate prediction value, the corrosion behavior of the iron-chromium alloy is obtained.

[0065] In some embodiments, step S100 can include but is not limited to steps S110 to S120:

[0066] Step S110, the iron-chromium alloy is prepared;

[0067] Step S120, the iron-chromium alloy is subjected to a closed environment corrosion simulation experiment under a single corrosion gas condition to obtain the corrosion kinetics experimental data.

[0068] In step S110 of some embodiments, the iron-chromium alloy for experiment is prepared. Illustratively, Fe-12.5Cr alloy is prepared by a vacuum arc melting furnace, and the obtained Fe-12.5Cr alloy is cast into long strip ingots. Subsequently, the ingots are placed in an environment of 1100°C for 2 hours of heat preservation treatment, and then rapidly quenched to room temperature. The long strip ingots are cut into round pieces with a diameter of 10 mm and a thickness of 3 mm by wire cutting technology. For the cut samples, the surface attached impurities and dust are carefully cleaned with ethanol, and after drying treatment, they are properly stored for subsequent use.

[0069] In step S120 of some embodiments, the prepared iron-chromium alloy is subjected to a corrosion simulation experiment in a closed environment under a single corrosion gas condition to obtain the corrosion kinetics experimental data. Illustratively, taking the processing of the corrosion kinetics data of Fe-12.5Cr alloy in single hydrogen, oxygen and water vapor as an example, the designed gas corrosion simulation device is used to perform the corrosion simulation experiment of the iron-chromium alloy in a closed environment. The gas corrosion simulation device includes thin film gauge gas total pressure probe, hydrogen sensor, oxygen sensor, water vapor sensor and closed reaction container and other key components. Among them, the hydrogen sensor, oxygen sensor and water vapor sensor are used to monitor the changes of hydrogen content, oxygen content and relative humidity in the closed environment in real time, and then the real-time experimental data of the corrosion kinetics of the iron-chromium alloy in the corrosion process are obtained.

[0070] In some embodiments, the Long Short-Term Memory (LSTM) network is a variant structure of a recurrent neural network (RNN), and its core is to introduce three gating mechanisms, namely the forgetting gate, the input gate and the output gate. These mechanisms enable the network to learn and remember sequence information over a long time span. Among them, the forgetting gate mechanism determines the retention degree of historical information through the Sigmoid function, and its calculation formula is as follows:

[0071] f t =σ(Wf [Y t-1 ,U t ]+b f ) (1)

[0072] where f t is the output of the forget gate; W f is the forget gate weight matrix; b f is the bias term; Y t-1 is the hidden state at the previous time step; U t is the input at the current time step; and σ denotes the Sigmoid activation function.

[0073] The input gate mechanism is responsible for determining whether to incorporate the current data into the cell state, thereby regulating the flow of information. The formula used includes:

[0074] x t = σ(W i [Y t-1 ,U t ]+b i ) (2)

[0075] C t =tanh(W c [Y t-1 ,U t ]+b c ) (3)

[0076] where x t is the output of the input gate; C t is the input state of the current cell; W i is the input gate weight matrix; W c is the candidate state weight matrix; and b i and b c are bias terms.

[0077] The output gate mechanism is responsible for controlling the information output of the current time state. The formula used includes:

[0078] O t =f t O t-1 +x t C t (4)

[0079] where O t is the current cell state; and O t-1 is the cell state at the previous time step.

[0080] The output gate mechanism not only determines which information will be passed to the next moment, but also multiplies the result after the activation function processing with the output of the tanh function through a branch, and finally passes to the subsequent network layer in the form of hidden state. The expression is as follows:

[0081] i t = sigma(W o [O t-1 , U t ]+b o ) (5)

[0082] Y t = i t tanh(O t ) (6)

[0083] In the formula, i t is the activation value of the output gate; W o is the output gate weight matrix; b o is the bias term; Y t is the current moment hidden state.

[0084] In step S200 of some embodiments, the long short-term memory network is improved to form an improved LSTM model combined with a double convolution structure, a pooling layer, an LSTM layer, a Dropout layer and a fully connected layer, and an iron-chromium alloy corrosion model is obtained. Exemplarily, the improved LSTM model network structure is as shown in Figure 2 The double convolution structure is composed of two convolution layers to enhance the extraction ability of local time sequence features: the first layer has an input channel of 1, an output channel of 16, a convolution kernel length of 3, and a ReLU activation function; the second layer keeps the channel number as 16, the convolution kernel parameters are unchanged with a length of 3, and also uses the ReLU activation function. Then a maximum pooling layer with a step of 2 is connected. The subsequently connected LSTM layer is used to capture the potential long-term dependence in the time series, and the number of hidden units of the LSTM layer is 28, followed by a Dropout layer (dropout rate is 0.35) to realize model regularization and effectively alleviate the risk of overfitting.

[0085] The improved LSTM network model of the embodiment of the application combines the extraction ability of the double convolution structure for local time sequence features and the modeling ability of the LSTM network for long-term sequence dependence, thereby constructing a corrosion behavior prediction model of iron-chromium alloy under different gas conditions.

[0086] Optionally, in terms of model training, the Adam optimizer is adopted for parameter updating, the learning rate is set to 0.0005, the upper limit of the training period is 200, and the early stopping mechanism (patience value is set to 10) is introduced to prevent overfitting and improve training efficiency. The batch size of the model is set to 32, and the window length of the time series is fixed to 20 time steps.

[0087] In some embodiments, step S300 can include but is not limited to steps S310-S350:

[0088] Step S310, inputting the corrosion kinetics experimental data into the iron-chromium alloy corrosion model;

[0089] Step S320, extracting local time sequence features of the corrosion kinetics experimental data through the double convolution structure to obtain first features;

[0090] Step S330, down-sampling the first features through the pooling layer to obtain second features;

[0091] Step S340, capturing potential long-term dependencies in the second features through the LSTM layer to obtain third features;

[0092] Step S350, mapping the third features to the corrosion kinetics prediction data through the fully connected layer.

[0093] In steps S310-S350 of some embodiments, the corrosion kinetics experimental data is input into the iron-chromium alloy corrosion model, the local time sequence features of the corrosion kinetics experimental data are extracted through the double convolution structure, the features are down-sampled through the pooling layer, the potential long-term dependencies in the time sequence features are captured through the LSTM layer, and finally the corrosion kinetics prediction data is mapped through the fully connected layer. Exemplarily, as shown in FIG. 3, the corrosion kinetics experimental data is input into the iron-chromium alloy corrosion model, the local time sequence features of the corrosion kinetics experimental data are extracted through the double convolution structure, the features are down-sampled through the pooling layer, the potential long-term dependencies in the time sequence features are captured through the LSTM layer, and finally the corrosion kinetics prediction data is mapped through the fully connected layer. Figure 2As shown, the corrosion kinetics experimental data of the Fe-12.5Cr alloy under different corrosion gas conditions are taken as the input of the iron-chromium alloy corrosion model. Optionally, the corrosion kinetics experimental data include the change of hydrogen content in the hydrogen corrosion process, the change of oxygen content in the oxygen corrosion process, and the change of relative humidity in the water vapor corrosion process. In the iron-chromium alloy corrosion model, first, the time series data input is subjected to local feature extraction through two one-dimensional convolutional neural network (CNN) modules, and feature down-sampling is performed through a maximum pooling layer with a step of 2. The features extracted by the convolutional neural network module are then input to the LSTM layer for capturing potential long-term dependencies in the time series. Finally, the 28-dimensional features output by the LSTM layer are mapped to a single value prediction result of the target corrosion parameter through a fully connected layer, i.e., the corrosion kinetics prediction data, which can include the prediction result of the dynamic change of hydrogen content in the closed cavity of the Fe-12.5Cr alloy in the hydrogen corrosion process, the prediction result of the dynamic change of oxygen content in the closed cavity of the Fe-12.5Cr alloy in the corrosion process, and the prediction result of the dynamic change of relative humidity in the closed cavity of the Fe-12.5Cr alloy in the water vapor corrosion process.

[0094] In some embodiments, the applicability and prediction accuracy of the constructed iron-chromium alloy corrosion model under different gas environment conditions are verified. To ensure the rigor of model training and verification, the same hardware and software configuration is uniformly used to train the iron-chromium alloy corrosion behavior prediction model in the embodiments of the present application. In addition, the experimental data set is collected by hydrogen, oxygen, and water vapor sensors in the gas corrosion simulation device, with a sampling frequency of once every 5 seconds. The data set is divided into a training set and a test set according to the time sequence, with a ratio of 8:2. The first 80% of the time series samples are used as the training set to support model learning, while the last 20% of the time series samples are used as the test set to evaluate the generalization ability of the model on unseen data. The test set is strictly independent of the training phase and only contains gas environment change information at future time points, avoiding data leakage problems.

[0095] Exemplarily, the Fe-12.5Cr alloy is taken as the research object, and the applicability and prediction accuracy of the constructed iron-chromium alloy corrosion model under different gas environment conditions are verified.

[0096] (1) Under hydrogen environment conditions, the applicability and prediction accuracy of the iron-chromium alloy corrosion model are verified.

[0097] The initial experimental conditions of the gas corrosion simulation device are set as follows: 0.1 MPa of hydrogen and nitrogen mixed gas, wherein the initial content of hydrogen is 2.66%, the initial temperature is 200 DEG C, and the corrosion experiment duration is 330 hours. The dynamic change of the hydrogen content in the closed cavity during the hydrogen corrosion process of the Fe-12.5Cr alloy is modeled and predicted by the iron-chromium alloy corrosion model. In the embodiment of the present application, the training set determination coefficient (Train R 2 ) and the training set mean absolute error (Train MAE) are selected as evaluation indexes for evaluating the performance of the model on the training data of the Fe-12.5Cr alloy hydrogen corrosion physical experiment. Table 1 shows the prediction performance of the iron-chromium alloy corrosion model for the change of the hydrogen content of Fe-12.5Cr in a closed environment over time during the corrosion process from 0 to 265 hours.

[0098] Table 1

[0099] Model R-squared on training set Mean absolute error on training set LSTM 0.9889 0.0011 Improved LSTM 0.9924 0.0010

[0100] For the improved LSTM model, i.e., the iron-chromium alloy corrosion model, the Train R 2 value reaches 0.9924, and the Train MAE value reaches 0.0010. After introducing the CNN, the Train R 2 value increases by 0.35%, and the Train MAE value decreases by 9.09%. This indicates that the introduction of the CNN can effectively extract the nonlinear features of the change of the hydrogen content of the Fe-12.5Cr alloy in a closed hydrogen corrosion environment, make up for the limitations of the single LSTM model in modeling ability, and improve the prediction accuracy of the hydrogen corrosion model of the Fe-12.5Cr alloy. In order to further evaluate and verify the prediction performance of the hydrogen corrosion model constructed by the LSTM and the improved model, the test set determination coefficient (Test R 2 ) and the test set mean absolute error (Test MAE) are selected as evaluation indexes for measuring the prediction effect of the model on the hydrogen content in the future time period. Table 2 shows the prediction ability of the iron-chromium alloy corrosion model for the change of the hydrogen content of the Fe-12.5Cr alloy in a closed environment over time during the corrosion process from 265 hours to 330 hours.

[0101] Table 2

[0102] Model R-squared on training set Mean absolute error on training set LSTM 0.9069 0.0012 Improved LSTM 0.9229 0.0011

[0103] As shown in Table 2, for the improved LSTM model, the Test R 2 value reaches 0.9229, and the Test MAE value reaches 0.0011. After introducing the CNN structure, the Test R 2The value increased by 1.76%, and the Test MAE value decreased by 8.33%. This result shows that the introduction of CNN effectively extracts the nonlinear characteristics and dynamic evolution of hydrogen content changes in the future time period during the hydrogen corrosion process of Fe-12.5Cr alloy. The improved LSTM model is better than the LSTM model in predicting future hydrogen content changes based on historical experimental data. Overall, the R 2 The value reaches 0.9913 and the MAE value reaches 0.0010. Compared with the LSTM model, R 2 The value increased by 0.85% and the MAE value decreased by 9.09%. Figure 3 The change of hydrogen content in Fe-12.5Cr alloy in a closed environment with reaction time is predicted based on the improved LSTM model, as well as the experimental measurement value monitored by a gas corrosion simulation device. Figure 3 The improved LSTM model accurately reflects the overall trend of hydrogen content changes during hydrogen corrosion of Fe-12.5Cr alloy, demonstrating a certain degree of reliability in predicting the hydrogen corrosion kinetics of Fe-12.5Cr alloy. Based on physical experimental results and model predictions, it was found that the hydrogen content did not remain constant over a 330-hour corrosion process at 200°C in a closed corrosion environment. Instead, it decreased parabolically, with hydrogen consumption reaching only approximately 0.078%. This phenomenon indicates that Fe-12.5Cr alloy does not react vigorously with hydrogen in the hydrogen environment. The decrease in hydrogen content is primarily attributed to hydrogen adsorption and dissociation on the alloy surface, followed by its subsequent diffusion into the alloy.

[0104] (2) Verify the applicability and prediction accuracy of the iron-chromium alloy corrosion model under oxygen environment conditions.

[0105] The initial experimental conditions of the gas corrosion simulator were set to: a 0.1 MPa mixture of oxygen and nitrogen, with an initial oxygen content of 16.54%, an initial temperature of 200°C, and a corrosion test duration of 333 hours. Using the iron-chromium alloy corrosion model, the dynamic changes in oxygen content within a closed cavity during the corrosion process of the Fe-12.5Cr alloy were modeled and predicted.

[0106] Table 3

[0107] Model R-squared on training set Mean absolute error on training set LSTM 0.9913 0.0181 Improved LSTM 0.9969 0.0093

[0108] The embodiment of the present invention selects Train R 2 The Train MAE and Train MAE were used as evaluation indicators to assess the performance of the model on the training data of the Fe-12.5Cr alloy oxygen corrosion physical experiment. Table 3 shows the prediction performance of the LSTM model and its improved model for the time-varying oxygen content of Fe-12.5Cr during the corrosion process from 0 to 268 hours in a closed environment.

[0109] For the improved LSTM model, the Train R 2 value reached 0.9969, and the Train MAE value reached 0.0093. After the introduction of CNN, the Train R 2 value increased by 0.56%, and the Train MAE value decreased by 48.62%. This indicates that the introduction of CNN can effectively extract the nonlinear characteristics of the change in oxygen content of Fe-12.5Cr alloy in a closed oxygen corrosion environment, making up for the limitations of the single LSTM model in modeling ability and improving the prediction accuracy of the oxygen corrosion model of Fe-12.5Cr alloy. In order to further evaluate and verify the prediction performance of the oxygen corrosion model constructed by the LSTM model and its improved model, the present application selects Test R 2 and TestMAE as evaluation indexes to measure the prediction effect of the model on the oxygen content in the future time period. Table 4 shows the prediction ability of LSTM and its improved model for the change of oxygen content of Fe-12.5Cr alloy in a closed environment over time during the corrosion process from 268 hours to 333 hours.

[0110] Table 4

[0111] Model R-squared on training set Mean absolute error on training set LSTM 0.9018 0.0243 Improved LSTM 0.9283 0.0197

[0112] From the data in Table 4, for the improved LSTM model, the Test R 2 value reached 0.9283, and the Test MAE value reached 0.0197. After the introduction of CNN, the Test R 2 value increased by 2.94%, and the Test MAE value decreased by 18.93%. This result reflects that the introduction of CNN effectively extracts the nonlinear characteristics and dynamic evolution law of the change in oxygen content in the future time period during the oxygen corrosion process of Fe-12.5Cr alloy, and the improved LSTM model has better ability than the LSTM model in predicting the future change in oxygen content based on historical experimental data. In summary, the R 2 value of the improved LSTM model reached 0.9961, and the MAE value reached 0.0123. Compared with the LSTM model, the R 2 value increased by 0.56%, and the MAE value decreased by 33.51%. Figure 4 shows the change of oxygen content of Fe-12.5Cr alloy in a closed environment with reaction time predicted based on the improved LSTM model, and the experimental measured value monitored by the gas corrosion simulation device. From Figure 4It can be seen that the improved LSTM model accurately reflects the overall trend of the change of oxygen content in the Fe-12.5Cr oxygen corrosion process, and shows a certain reliability in the prediction of the Fe-12.5Cr oxygen corrosion kinetics. Based on the physical experiment results and model prediction data, it is concluded that in a closed corrosion environment at 200℃, Fe-12.5Cr undergoes a reaction process of 333 hours, and the oxygen content experiences several slow parabolic stages and gentle transition stages, with an oxygen consumption of about 0.90%. During this period, the densification of the oxide film inhibits the further reaction between the alloy and oxygen.

[0113] (3) In the water vapor environment, the applicability and prediction accuracy of the iron-chromium alloy corrosion model are verified.

[0114] The initial experimental conditions of the gas corrosion simulation device are set as follows: 0.1 MPa of water vapor and nitrogen mixed gas, with an initial water vapor content of 56.25%, an initial temperature of 200℃, and a corrosion experiment duration of 338 hours. By constructing LSTM and its improved model, the dynamic change of relative humidity in the closed cavity of Fe-12.5Cr alloy in the water vapor corrosion process is modeled and predicted. In the present embodiment, Train R 2 and Train MAE are selected as evaluation indexes to evaluate the performance of the model on the training data of Fe-12.5Cr alloy water vapor corrosion physical experiment. Table 5 shows the prediction performance of the LSTM model and its improved model for the change of relative humidity of Fe-12.5Cr in the closed environment from 0 to 272 hours in the corrosion process.

[0115] Table 5

[0116] Model R-squared on training set Mean absolute error on training set LSTM 0.9951 0.0695 Improved LSTM 0.9962 0.0567

[0117] For the improved LSTM model, the Train R 2 value reaches 0.9962, and the Train MAE value reaches 0.0567. After the introduction of CNN, the Train R 2 value increases by 0.11%, and the Train MAE value decreases by 18.42%. This indicates that the introduction of CNN can effectively extract the nonlinear features of the change of relative humidity of Fe-12.5Cr alloy in the closed water vapor corrosion environment, make up for the limitations of the single LSTM model in modeling ability, and improve the prediction accuracy of the Fe-12.5Cr alloy water vapor corrosion model. In order to further evaluate and verify the performance of the water vapor corrosion prediction model constructed by LSTM and its improved model, the present invention selects Test R 2and Test MAE as evaluation indexes, which are used to measure the prediction effect of the model on the relative humidity in the future time period. Table 6 shows the prediction ability of the LSTM model and its improved model on the relative humidity of Fe-12.5Cr alloy in a closed environment during the corrosion process from 272 hours to 338 hours.

[0118] Table 6

[0119] Model R-squared on training set Mean absolute error on training set LSTM 0.9921 0.0853 Improved LSTM 0.9928 0.0784

[0120] From the data in Table 6, for the improved LSTM model, the Test R 2 value reaches 0.9928, and the Test MAE value reaches 0.0784. After introducing the CNN structure, the Test R 2 value increases by 0.07%, and the Test MAE value decreases by 8.09%. This result reflects that the introduction of CNN effectively extracts the nonlinear characteristics and dynamic evolution law of the relative humidity change in the future time period during the water vapor corrosion process of Fe-12.5Cr alloy, and the prediction ability of the improved LSTM model based on historical experimental data on the future relative humidity change is better than that of the LSTM model. In summary, the R 2 value of the improved LSTM model reaches 0.9953, and the MAE value reaches 0.0673. Compared with the LSTM model, the R 2 value increases by 0.21%, and the MAE value decreases by 8.93%. Figure 5 shows the change of the relative humidity of Fe-12.5Cr alloy in a closed environment with the reaction time predicted based on the improved LSTM model, and the experimental measured value monitored by the gas corrosion simulation device. From Figure 5 it can be known that the improved LSTM model accurately reflects the overall trend of the change of the relative humidity during the water vapor corrosion process of Fe-12.5Cr, and shows a certain reliability in the prediction research of the water vapor corrosion kinetics of Fe-12.5Cr. Based on the physical experimental results and the model prediction data, it is concluded that in a closed corrosion environment at 200°C, after a reaction process of 338 hours, the change of the relative humidity shows a linear rapid decrease trend in the initial stage, and then shows a parabolic slow decrease trend, and the water vapor consumption is about 16.59%. During this period, with the gradual formation and densification of the oxide film, its barrier effect on water vapor is enhanced, thereby inhibiting the progress of the oxidation corrosion reaction.

[0121] In some embodiments, step S400 can include but is not limited to steps S410 to S430:

[0122] Step S410, constructing a first constraint condition according to the gas pressure, the gas volume, the amount of gas substance, the absolute temperature of the gas, and the gas constant;

[0123] Step S420, constructing a second constraint condition according to the weight change of the iron-chromium alloy, the exposed area of the iron-chromium alloy, and the corrosion time of the iron-chromium alloy;

[0124] Step S430, obtaining the iron-chromium alloy oxidation weight gain prediction model based on the improved LSTM model, combining the first constraint condition and the second constraint condition.

[0125] In step S410 of some embodiments, a physical constraint condition, i.e., a first constraint condition, is constructed by gas pressure, gas volume, gas amount of substance, absolute temperature of gas, and gas constant. Illustratively, based on the iron-chromium alloy corrosion model constructed by the improved LSTM model, real-time prediction data of key atmosphere parameters in the closed cavity are obtained, and the weight prediction data of the iron-chromium alloy at each time under the closed atmosphere are further obtained by combining the first constraint condition, so as to realize the prediction of the weight change of the iron-chromium alloy in the corrosion process, and thus the regularity of the weight change of the iron-chromium alloy in the corrosion process is obtained. The expression of the first constraint condition is as follows:

[0126] PV = nRT

[0127] In the formula, P represents the gas pressure; V represents the gas volume; n represents the amount of gas substance; R represents the gas constant; and T represents the absolute temperature of the gas.

[0128] Reference Figure 6 , Figure 6 The prediction results of the weight change of the Fe-12.5Cr alloy with time in the oxygen corrosion process based on the improved LSTM model are shown. Figure 6 It can be seen that in the closed oxygen corrosion environment at 200°C, after 333 hours of reaction process, the weight change of the Fe-12.5Cr alloy shows a nonlinear rising trend, corresponding to two stages: in 0-52 hours, the weight of the Fe-12.5Cr alloy increases rapidly in a parabolic manner; in 52-333 hours, the weight increases slowly in a linear manner. At the 333rd hour, the model predicts the weight of the Fe-12.5Cr alloy to be slightly higher than the experimental weighing, with an absolute error of 0.000018 g, further verifying the feasibility and reliability of the Fe-12.5Cr alloy oxygen corrosion model constructed based on the improved LSTM model.

[0129] Reference Figure 7 , Figure 7 The prediction results of the weight change of the Fe-12.5Cr alloy with time in the water vapor corrosion process based on the improved LSTM model are shown. Figure 7It can be seen that in the closed water vapor corrosion environment at 200℃, after 338 hours of reaction process, the weight change of Fe-12.5Cr alloy shows a nonlinear upward trend, corresponding to two stages: in 0-79 hours, the weight of Fe-12.5Cr alloy increases linearly and rapidly; in 79-338 hours, it increases linearly and gently. At the 338th hour, the model predicts that the weight of Fe-12.5Cr alloy is slightly higher than the experimental weighing, with an absolute error of 0.000014g, further verifying the feasibility and reliability of the water vapor corrosion model of Fe-12.5Cr alloy based on the improved LSTM model.

[0130] In step S420 of some embodiments, a physical constraint condition, i.e., a second constraint condition, is constructed by the weight change of the iron-chromium alloy, the exposed area of the iron-chromium alloy, and the corrosion time of the iron-chromium alloy. Illustratively, based on the real-time weight data output by the improved LSTM model, combined with the second constraint condition, the oxidation rate of Fe-12.5Cr alloy at different time periods is further derived, realizing the prediction of the oxidation rate of the iron-chromium alloy in the corrosion process, thereby quantitatively describing the oxidation process of the iron-chromium alloy, and providing a quantitative analysis tool for the study of iron-chromium alloy oxidation kinetics. The expression of the second constraint condition is as follows:

[0131]

[0132] In the formula, r represents the oxidation rate of the iron-chromium alloy; Δm represents the weight change of the iron-chromium alloy; A represents the exposed area of the iron-chromium alloy; and Δt represents the corrosion time of the iron-chromium alloy.

[0133] Reference Figure 8 , Figure 8 The change of the oxidation rate of Fe-12.5Cr alloy with time in the oxygen corrosion environment predicted based on the improved LSTM model is shown. Figure 8 It can be seen that in the closed oxygen corrosion environment at 200℃, after 333 hours of reaction process, the oxidation rate of Fe-12.5Cr alloy presents two obvious stage characteristics: in 0-4 hours, the oxidation rate of Fe-12.5Cr increases linearly and rapidly; in 4-333 hours, it decreases slowly in a parabolic manner. At the 333rd hour, the model predicts that the oxidation rate of Fe-12.5Cr alloy is slightly higher than the experimental result, with an absolute error of 0.0231μg·cm -2 ·h -1 .

[0134] Reference Figure 9 , Figure 9 The change of the oxidation rate of Fe-12.5Cr alloy with time in the water vapor corrosion environment predicted based on the improved LSTM model is shown. Figure 9It can be seen that in the closed water vapor corrosion environment at 200℃, after 338 hours of reaction process, the oxidation rate of Fe-12.5Cr alloy presents two obvious stage characteristics: in 0-67 hours, the oxidation rate of Fe-12.5Cr presents a parabolic rapid increase; in 67-338 hours, the oxidation rate presents a parabolic slow decrease. In the 338th hour, the model predicts the oxidation rate of Fe-12.5Cr alloy to be slightly higher than the experimental result, and the absolute error is 0.0155 μg·cm -2 ·h -1 .

[0135] In step S430 of some embodiments, on the basis of the improved LSTM model built iron-chromium alloy corrosion model, combined with physical constraint conditions, an iron-chromium alloy oxidation weight gain prediction model is further built. Among them, the physical constraint conditions include the first constraint condition and the second constraint condition.

[0136] In step S500 of some embodiments, through the built iron-chromium alloy oxidation weight gain prediction model, the effective prediction and verification of iron-chromium alloy weight gain and oxidation rate are realized, and the stage characteristics of weight change rule and oxidation rate in the corrosion process of iron-chromium alloy can be obtained.

[0137] In step S600 of some embodiments, corrosion kinetics prediction data is obtained through the built iron-chromium alloy corrosion model, and weight change prediction value and oxidation rate prediction value are obtained through the built iron-chromium alloy oxidation weight gain prediction model, so as to obtain the corrosion behavior of iron-chromium alloy.

[0138] As Figure 10 shown, taking the processing of Fe-12.5Cr alloy corrosion kinetics experimental data as an example, the iron-chromium alloy corrosion behavior prediction processing process based on improved LSTM of the embodiment of the application is: first, Fe-12.5Cr alloy is prepared, corrosion simulation experiments under single hydrogen, oxygen and water vapor conditions are carried out, and iron-chromium alloy key gas parameter corrosion kinetics experimental data are obtained. Secondly, the LSTM model is built and optimized by the embodiment of the application, the iron-chromium alloy corrosion model is built, including the iron-chromium alloy hydrogen corrosion model, the iron-chromium alloy oxygen corrosion model and the iron-chromium alloy water vapor corrosion model, and Train R 2 , Train MAE, Test R 2And Test MAE, multiple evaluation indexes, comprehensive evaluation of the performance of iron chromium alloy corrosion model, ultimately selected improved LSTM model as the machine learning modeling scheme of iron chromium alloy corrosion behavior research. By inputting the gas parameter experimental data of iron chromium alloy under different conditions, the effective prediction and verification of the hydrogen content, oxygen content and relative humidity in the future time period are realized, and the dynamic evolution law of the key gas parameters in the corrosion process of iron chromium alloy is obtained. On the basis of the improved LSTM model, an iron chromium alloy oxidation weight gain prediction model is further constructed by combining physical constraint conditions, the effective prediction and verification of the weight gain and oxidation rate of iron chromium alloy are realized, and the stage characteristics of weight change and oxidation rate in the corrosion process of iron chromium alloy are obtained.

[0139] As shown in Figure 11 The improved LSTM-based iron chromium alloy corrosion behavior prediction method in a closed atmosphere is based on iron chromium alloy corrosion data to build an iron chromium alloy corrosion behavior digital evaluation and prediction platform, realize the digital evaluation of the corrosion process of iron chromium alloy and the digital prediction of the corrosion behavior, and provide an important technical means for analyzing the corrosion characteristics and rules of iron chromium alloy. In the iron chromium alloy corrosion behavior digital evaluation and prediction platform, the database is constructed, the model is constructed, the model is predicted, and the experiment is verified, so that the improved LSTM-based iron chromium alloy corrosion behavior prediction method in a closed environment can be realized.

[0140] The embodiment of the application also provides an improved LSTM-based iron chromium alloy corrosion behavior prediction device in a closed environment, which can realize the improved LSTM-based iron chromium alloy corrosion behavior prediction method in a closed environment.

[0141] The first module is used for acquiring corrosion kinetics experimental data of iron chromium alloy.

[0142] The second module is used for constructing an iron chromium alloy corrosion model based on the improved LSTM model.

[0143] The third module is used for inputting the corrosion kinetics experimental data into the iron chromium alloy corrosion model to obtain corrosion kinetics prediction data.

[0144] The fourth module is used for constructing an iron chromium alloy oxidation weight gain prediction model based on the improved LSTM model and combining physical constraint conditions.

[0145] The fifth module is used for acquiring weight change prediction values and oxidation rate prediction values of the iron chromium alloy through the iron chromium alloy oxidation weight gain prediction model.

[0146] The sixth module is used for obtaining the corrosion behavior of the iron chromium alloy according to the corrosion kinetics prediction data, the weight change prediction values and the oxidation rate prediction values.

[0147] It can be understood that the contents in the above method embodiments are all applicable to the device embodiments, the device embodiments specifically implement the functions same as the above method embodiments, and achieve the same beneficial effects as the above method embodiments.

[0148] The embodiment of the application further provides an electronic device, which comprises a processor and a memory, the memory stores a computer program, and the processor implements the above method when executing the computer program. The electronic device can be any intelligent terminal including a tablet computer, a vehicle-mounted computer, etc.

[0149] It can be understood that the contents in the above method embodiments are all applicable to the device embodiments, the device embodiments specifically implement the functions same as the above method embodiments, and achieve the same beneficial effects as the above method embodiments.

[0150] Reference Figure 12 , Figure 12 The hardware structure of the electronic device of another embodiment is illustrated, and the electronic device comprises:

[0151] The processor 701 can be implemented in the form of a general-purpose CPU (Central Processing Unit), a microprocessor, an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits, etc., and is used to execute related programs to implement the technical solutions provided by the embodiments of the application;

[0152] The memory 702 can be implemented in the form of a ROM (Read Only Memory), a static storage device, a dynamic storage device, or a RAM (Random Access Memory), etc. The memory 702 can store an operating system and other application programs, and when the technical solutions provided by the embodiments of the application are implemented by software or firmware, the related program codes are saved in the memory 702 and are called and executed by the processor 701 to implement the method of the embodiments of the application;

[0153] The input / output interface 703 is used to realize information input and output;

[0154] The communication interface 704 is used to realize the communication interaction between the device and other devices, and can realize communication through a wired manner (for example, a USB, a network cable, etc.) or a wireless manner (for example, a mobile network, WIFI, Bluetooth, etc.);

[0155] A bus 705 is used to transmit information between the various components (e.g., the processor 701, the memory 702, the input / output interface 703, and the communication interface 704) in the device.

[0156] The processor 701, the memory 702, the input / output interface 703, and the communication interface 704 are communicatively connected to each other within the device through the bus 705.

[0157] The embodiment of the present application also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the method.

[0158] It can be understood that the content in the above method embodiments is applicable to the present storage medium embodiment, the present storage medium embodiment specifically implements the functions of the above method embodiments, and achieves the same beneficial effects as the above method embodiments.

[0159] The embodiment of the present application also provides a computer program product or a computer program, which comprises computer instructions stored in a computer readable storage medium. A processor of a computer device can read the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to enable the computer device to perform the above method.

[0160] To sum up, the method and device for predicting corrosion behavior of iron-chromium alloy in a sealed environment based on an improved LSTM have the following advantages:

[0161] 1. The embodiment of the present application uses a machine learning method to develop a more accurate prediction method for the corrosion behavior of iron-chromium alloy, improves the adaptability of the model under different atmosphere conditions, enhances the adaptability and prediction accuracy of the corrosion model under different gas conditions, and significantly improves the efficiency of data processing of the corrosion of iron-chromium alloy.

[0162] 2. The embodiment of the present application can effectively predict and verify the hydrogen content, oxygen content, and relative humidity in the future time period, and can obtain the dynamic evolution law of the key atmosphere parameters in the corrosion process of iron-chromium alloy. It also effectively predicts and verifies the weight gain and oxidation rate of iron-chromium alloy, and can obtain the stage characteristics of weight change and oxidation rate in the corrosion process of iron-chromium alloy.

[0163] In some alternative embodiments, the function / operations mentioned in the block diagrams can not occur in the order mentioned in the operational illustrations. For example, depending on the involved function / operation, two blocks shown in succession can in fact be executed substantially concurrently or the blocks can sometimes be executed in reverse order, depending upon the functionality / operations involved. Furthermore, embodiments presented and described in the flowcharts are only examples of implementing the present application. Alternative embodiments can be implemented where various operations are changed, omitted, and / or added. For example, the order of the operations can be changed, and sub-operations described as part of a larger operation can be implemented independently.

[0164] Furthermore, although the present application is described in the context of functional modules, it is to be understood that one or more of the functions and / or features described can be integrated in a single physical device and / or software module, or one or more functions and / or features can be implemented in separate physical devices or software modules. It is also to be understood that detailed discussion of the actual implementation of each module is unnecessary to an understanding of the present application. Rather, the actual implementation is within the routine skill of engineers familiar with the attributes, functions, and internal relationships of the various functional modules disclosed herein. Accordingly, the present application is not limited to the details of the implementations described herein, but rather is limited only by the scope of the claims and the equivalents thereof.

[0165] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application or the part of the technical solutions that make essential contributions to the prior art or the part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various program code storage media.

[0166] The logic and / or steps represented in the flowcharts and / or described herein, for example, can be embodied in non-transitory computer-readable media, executed by one or more computing devices, and / or in any other way. The logic and / or steps represented in the flowcharts and / or described herein, for example, can be considered a list of executable instructions for implementing logic functions, and can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor- containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For purposes of this specification, a "computer-readable medium" can be any apparatus that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. The computer-readable medium can be, for example but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection (electronic) having one or more wires, a portable computer diskette (magnetic), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber (optical), and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example via an optical scanner, then compiled, interpreted, or otherwise processed, and stored in a computer memory in a form that can be later executed by a computer. In this context, a "computer-readable medium" can be any means that can store the program for use by or in connection with the instruction execution system, apparatus, or device.

[0167] The foregoing description, for purposes of clarity, describes the present application in terms of specific embodiments thereof. However, it is to be understood that the terminology used, and the contents of the description, are not to be considered in any way limiting of the present application. The present application is to be considered as encompassing all embodiments falling within the scope of the appended claims, and their equivalents.

[0168] It should be understood that aspects of the present application can be implemented in hardware, software, firmware or combinations thereof. In the above embodiments, various steps or methods can be implemented, for example, by software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and in another embodiment, any of the following can be used: discrete logic circuitry having logic gates for implementing logic functions upon an application of data signals; application specific integrated circuits having appropriate combinational logic gates; programmable gate arrays (PGA), field programmable gate arrays (FPGA), and / or other implementations known to those skilled in the art.

[0169] In the description of the present application, the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" are not necessarily referring to the same embodiment or example. Furthermore, the described specific features, structures, materials or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0170] While the embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary and are not to be construed as limiting the scope of the application. The scope of the application is defined by the appended claims and their equivalents.

[0171] The above is a specific description of the preferred embodiment of the present application, but the present application is not limited to the described embodiment, and those skilled in the art can make various equivalent modifications or replacements without departing from the spirit of the present application, and these equivalent modifications or replacements are all included in the scope defined by the claims of the present application.

Claims

1. A method for predicting the corrosion behavior of iron-chromium alloy in a closed environment based on improved LSTM, characterized in that: The following steps are involved: Obtain experimental data on corrosion kinetics of iron-chromium alloys; Based on the improved LSTM model, an iron-chromium alloy corrosion model was constructed; Inputting the corrosion kinetics experimental data into the iron-chromium alloy corrosion model to obtain corrosion kinetics prediction data; Based on the improved LSTM model and combined with physical constraints, a prediction model for the oxidation weight gain of iron-chromium alloy was constructed; Obtaining a weight change prediction value and an oxidation rate prediction value of the ferrochromium alloy by using the ferrochromium alloy oxidation weight gain prediction model; The corrosion behavior of the iron-chromium alloy is obtained according to the corrosion kinetics prediction data, the weight change prediction value, and the oxidation rate prediction value.

2. The method for predicting corrosion behavior of iron-chromium alloy in a closed environment based on improved LSTM according to claim 1, characterized in that: The method of obtaining the corrosion kinetics experimental data of the iron-chromium alloy comprises the following steps: preparing the ferrochromium alloy; A closed environment corrosion simulation experiment under a single corrosive gas condition is performed on the iron-chromium alloy to obtain the corrosion kinetics experimental data.

3. The method for predicting corrosion behavior of iron-chromium alloy in a closed environment based on improved LSTM according to claim 1, characterized in that: The improved LSTM model is used to construct an iron-chromium alloy corrosion model, which includes the following steps: The LSTM model is improved through a double convolution structure, a pooling layer, an LSTM layer, a Dropout layer, and a fully connected layer to obtain the iron-chromium alloy corrosion model.

4. The method for predicting corrosion behavior of iron-chromium alloy in a closed environment based on improved LSTM according to claim 3 is characterized in that: The step of inputting the corrosion kinetics experimental data into the iron-chromium alloy corrosion model to obtain corrosion kinetics prediction data comprises the following steps: Inputting the corrosion kinetics experimental data into the iron-chromium alloy corrosion model; Extracting the local time series features of the corrosion dynamics experimental data through the double convolution structure to obtain a first feature; Downsampling the first feature through the pooling layer to obtain a second feature; The potential long-term dependency in the second feature is captured by the LSTM layer to obtain a third feature; The third feature is mapped into the corrosion dynamics prediction data through the fully connected layer.

5. The method for predicting corrosion behavior of iron-chromium alloy in a closed environment based on improved LSTM according to claim 1, characterized in that: Based on the improved LSTM model and combined with physical constraints, a prediction model for the oxidation weight gain of iron-chromium alloy is constructed. The following steps are involved: Construct the first constraint condition based on gas pressure, gas volume, amount of gas substance, absolute temperature of gas and gas constant; constructing a second constraint condition according to the weight change of the ferrochromium alloy, the exposed area of ​​the ferrochromium alloy, and the corrosion time of the ferrochromium alloy; Based on the improved LSTM model, combined with the first constraint condition and the second constraint condition, the iron-chromium alloy oxidation weight gain prediction model is obtained.

6. The method for predicting corrosion behavior of iron-chromium alloy in a closed environment based on improved LSTM according to claim 1, characterized in that: Based on the improved LSTM model and combined with physical constraints, a prediction model for the oxidation weight gain of ferrochromium alloy is constructed. The formula used includes: PV=nRT Wherein, P represents gas pressure; V represents gas volume; n represents the amount of gas substance; R represents gas constant; T represents absolute temperature of gas; r represents oxidation rate of FeCr alloy; Δm represents weight change of FeCr alloy; A represents exposed area of ​​FeCr alloy; Δt represents corrosion time of FeCr alloy.

7. A device for predicting the corrosion behavior of iron-chromium alloy in a closed environment based on improved LSTM, characterized in that: include: The first module is used to obtain experimental data on corrosion kinetics of iron-chromium alloys; The second module is used to build an iron-chromium alloy corrosion model based on the improved LSTM model; The third module is used to input the corrosion kinetics experimental data into the iron-chromium alloy corrosion model to obtain corrosion kinetics prediction data; The fourth module is used to build a prediction model for the oxidation weight gain of iron-chromium alloy based on the improved LSTM model and combined with physical constraints; A fifth module is configured to obtain a weight change prediction value and an oxidation rate prediction value of the ferrochromium alloy by using the ferrochromium alloy oxidation weight gain prediction model; The sixth module is used to obtain the corrosion behavior of the iron-chromium alloy based on the corrosion kinetics prediction data, the weight change prediction value and the oxidation rate prediction value.

8. An electronic device, characterized in that: including a processor and a memory; The memory is used to store programs; The processor executes the program to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The storage medium stores a program, and the program is executed by a processor to implement the method according to any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images