Molten iron temperature drop prediction method and prediction system based on BP neural network
By screening and structuring the historical data of the iron and water transportation process, the BP neural network model is constructed, which solves the problem of low accuracy in the prediction of temperature drop in the existing technology, and accurately predicts the temperature drop value of each stage of the iron and water transportation process, and improves production stability.
Patent Information
- Application Number
- CN202411869310.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-18
- Publication Date
- 2025-05-06
AI Technical Summary
The prior art has low accuracy in predicting temperature drop in the iron and water transport process, mainly relying on workers' experience or mathematical mechanism formulas, and historical data is incompletely preserved, with errors and missing data, making it difficult to directly use in neural network training.
The iron temperature drop prediction method based on BP neural network is adopted. By screening and structuring historical data, abnormal data are eliminated, the BP neural network model is constructed, and the model training and verification data set is trained and verified, the iron temperature drop prediction model in the iron transport process is obtained.
Accurate prediction of the temperature drop value of each stage of the iron transport process is achieved, the prediction accuracy is improved, the workers' labor volume is reduced, and the production stability is ensured.
Smart Images
Figure CN119939150A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a molten iron temperature drop prediction method based on BP neural network and a prediction method. Background Art
[0002] At present, with the implementation of the double reduction policy, energy conservation and emission reduction requirements have been put forward for all links of molten iron transportation. In recent years, most large steel mills in China have chosen torpedo tank cars for molten iron transportation. In the actual converter steelmaking process, in order to make the molten iron temperature reach the required temperature when it arrives at the steel plant, it is necessary to accurately predict the temperature drop of the molten iron during the transportation of the molten iron by the torpedo tank car. At present, there are two main ways to predict the temperature drop during the transportation of molten iron: one is to rely on the experience of on-site workers to estimate how much the temperature of the molten iron will drop when it arrives at the steel plant through the torpedo tank. This method has low accuracy and large manual errors; the other is to use mathematical mechanism formulas to calculate, but the on-site environment and production process are relatively complex, and the calculation accuracy of the mechanism formula is low. Therefore, in the field of molten iron transportation, there is still a lack of a method that can accurately predict the temperature drop of molten iron during molten iron transportation.
[0003] BP neural network is one of the most widely used neural network models. It has very high requirements for the quality of sample data. The quality of sample data will directly affect the training effect of the neural network. If the original data is not processed and directly used to train the neural network model, the training effect will be poor. Low-quality sample data can only achieve poor training results. Due to the particularity of torpedo tanker transportation and the complexity of the on-site application environment, it is difficult to preserve the historical sample data of the molten iron transportation process. Even if the data is preserved, there are some erroneous data and missing data, which makes it difficult to use it directly for neural network training. Summary of the invention
[0004] In response to the technical problems existing in the background technology, the present invention provides a molten iron temperature drop prediction method and prediction method based on BP neural network. On the basis of improving the original sample data, it performs structured processing on it, utilizes the overall distribution law of the data, eliminates some sample data with obvious errors, ensures the high quality of the sample data, thereby overcoming the shortcomings of traditional neural networks, and can predict the temperature drop values of each stage of molten iron transportation.
[0005] According to one aspect of the present application, the present invention provides a method for predicting molten iron temperature drop based on a BP neural network, which comprises:
[0006] Collect historical molten iron temperature drop data during the molten iron transportation process of the torpedo tanker, each item of the historical molten iron temperature drop data during the molten iron transportation process of the torpedo tanker includes: torpedo tanker model, molten iron loading, initial molten iron temperature, and molten iron temperature drop values or working layer temperature drop values corresponding to each stage of the molten iron transportation process;
[0007] The collected historical molten iron temperature drop data of the molten iron transportation process of the tank car are screened and classified, abnormal data in the data are removed, and valid data are structured and stored to obtain structured data sets corresponding to each stage of the molten iron transportation process;
[0008] The structured data set is divided into a training data set and a verification data set in proportion, a BP neural network model is constructed, and the constructed BP neural network model is trained and verified by the training data set and the verification data set respectively, so as to obtain a prediction model of molten iron temperature drop in the molten iron transportation process;
[0009] The temperature drop value at each stage of the molten iron transportation process is predicted by the molten iron temperature drop prediction model during the molten iron transportation process.
[0010] Furthermore, the molten iron transportation process includes: a heavy tank waiting stage, a heavy tank transportation stage, a tank emptying waiting stage, an empty tank waiting stage, an empty tank return stage and a waiting stage for receiving iron; each item of historical molten iron temperature drop data of the torpedo tank truck molten iron transportation process also includes: molten iron heavy tank waiting time, molten iron heavy tank transportation time, tank emptying waiting time, empty tank waiting time, empty tank return time and iron receiving waiting time.
[0011] Furthermore, the historical molten iron temperature drop data collected during the molten iron transportation process by torpedo tank cars are divided into open torpedo tank car molten iron transportation data and covered torpedo tank car molten iron transportation data, and the data of the two types of data under the same working conditions are sorted out respectively, and 95% is used as the confidence interval to observe the normal distribution of each data item, and the molten iron temperature drop data with abnormal data are removed. Finally, the effective and reasonable temperature drop data are structured and stored to form multiple structured data sets corresponding to each stage of the molten iron transportation process; each data item in the structured data set contains 3 columns, and the first to second columns are feature columns: the first column is the molten iron temperature or the working layer temperature, and the second column is the required time; the third column is the label column: the molten iron temperature drop value or the working layer temperature drop value.
[0012] Furthermore, the structured data set is divided into a training data set and a validation data set in a ratio of 8:2.
[0013] Furthermore, the BP neural network model has three layers, namely, an input layer, a hidden layer and an output layer from top to bottom. The input layer has 20 neurons, the hidden layer has 10 neurons, and the output layer has 1 neuron. The optimization algorithm used in training the BP neural network model is the Levenberg-Marquardt algorithm.
[0014] Further, when the temperature drop value of each stage of the molten iron transportation process is predicted by the molten iron temperature drop prediction model during the molten iron transportation process, the torpedo tank car model, the molten iron loading amount, the time value of each stage of the molten iron transportation process, and the molten iron temperature value or the working layer temperature value at the beginning of each stage are used as input parameters of the molten iron temperature drop prediction model during the molten iron transportation process to predict the temperature drop value of the corresponding stage, specifically including:
[0015] The initial temperature of molten iron and the waiting time of the molten iron reloading tank are taken as input parameters to predict the temperature drop of molten iron during the waiting process of the molten iron reloading tank.
[0016] The molten iron transportation temperature and the molten iron heavy tank transportation time are taken as input parameters to predict the molten iron temperature drop during the molten iron heavy tank transportation process.
[0017] The molten iron temperature and the waiting time for molten iron pouring are taken as input parameters to predict the molten iron temperature drop during the waiting process of molten iron pouring.
[0018] The empty tank waiting time is taken as an input parameter to predict the working layer temperature drop value during the empty tank waiting process.
[0019] The empty tank return time is used as an input parameter to predict the working layer temperature drop during the empty tank return process.
[0020] The iron-receiving waiting time is taken as an input parameter to predict the temperature drop of the working layer during the iron-receiving waiting process.
[0021] According to another aspect of the present application, a molten iron temperature drop prediction system based on a BP neural network is provided, comprising:
[0022] The module for collecting historical molten iron temperature drop data during molten iron transportation is used to collect historical molten iron temperature drop data during molten iron transportation by torpedo tank trucks. Each item of historical molten iron temperature drop data during molten iron transportation by torpedo tank trucks includes: torpedo tank truck model, molten iron loading capacity, initial molten iron temperature, and molten iron temperature drop values or working layer temperature drop values corresponding to each stage of the molten iron transportation process.
[0023] The module for screening and classifying historical data on the temperature drop of molten iron during the molten iron transportation process is used to screen and classify the historical data on the temperature drop of molten iron collected during the molten iron transportation process by the tank car, remove abnormal data in the data, and perform structured storage on the valid data to obtain structured data sets corresponding to each stage of the molten iron transportation process.
[0024] A data training and verification module is used to divide the structured data set into a training data set and a verification data set in proportion, construct a BP neural network model, and train and verify the constructed BP neural network model through the training data set and the verification data set respectively, to obtain a molten iron temperature drop prediction model during the molten iron transportation process.
[0025] The temperature drop prediction module is used to predict the temperature drop value at each stage of the molten iron transportation process through the molten iron temperature drop prediction model of the molten iron transportation process.
[0026] Furthermore, the temperature drop prediction module uses the torpedo tank car model, the molten iron loading amount, the time value of each stage of the molten iron transportation process, and the molten iron temperature value or the working layer temperature value at the beginning of each stage as the input parameters of the molten iron temperature drop prediction model of the molten iron transportation process, and predicts the temperature drop value of the corresponding stage of the molten iron transportation process through submodules, and the submodules include:
[0027] The re-tank waiting stage submodule is used to obtain data for predicting the molten iron re-tank waiting stage, including: the initial temperature of the molten iron and the waiting time of the molten iron re-tank as input parameters, and predict the temperature drop of the molten iron during the molten iron re-tank waiting process.
[0028] The heavy tank transportation stage submodule is used to obtain data for predicting the heavy tank transportation stage, including: molten iron transportation temperature and molten iron heavy tank transportation time as input parameters, and predict the molten iron temperature drop during the molten iron heavy tank transportation process.
[0029] The tank pouring waiting stage submodule is used to obtain data of the tank pouring waiting stage, including: molten iron temperature and tank pouring waiting time as input parameters, and predict the molten iron temperature drop during the molten iron pouring waiting process.
[0030] The empty tank waiting stage submodule is used to obtain the data of the empty tank waiting stage, including: the empty tank waiting time is used as an input parameter, and the temperature drop value of the working layer in the empty tank waiting stage is predicted.
[0031] The empty tank return stage submodule is used to obtain the data of the empty tank return stage, including: the empty tank return time is used as an input parameter, and the temperature drop value of the working layer in the return stage is predicted.
[0032] The iron receiving waiting stage submodule is used to obtain data of the iron receiving waiting stage, including: the iron receiving waiting time is used as an input parameter to predict the temperature drop value of the working layer in the iron receiving waiting stage.
[0033] According to another aspect of the present application, an electronic device is provided, comprising a memory and a processor, wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method for predicting the temperature drop of molten iron described in the above technical solution.
[0034] According to another aspect of the present application, a readable storage medium stores computer instructions thereon, wherein the computer instructions, when executed by a processor, implement the method for predicting the temperature drop of molten iron described in the above technical solution.
[0035] Compared with the prior art, the present invention provides a molten iron temperature drop prediction method and prediction method based on BP neural network, and the beneficial effects are as follows:
[0036] (1) The present invention uses a BP neural network model to train and verify the actual historical data of molten iron temperature drop on site, and obtains a prediction model for molten iron temperature drop during molten iron transportation, thereby accurately predicting the temperature drop of high-temperature molten iron at each stage of torpedo tanker transportation. This prediction method solves the problem of poor accuracy of relying on workers' experience and mathematical mechanism formulas, reduces the workload of workers, ensures the stability of production, and helps to carry out subsequent processes.
[0037] (2) The present invention can effectively filter and format the historical data of the temperature drop of molten iron during the original molten iron transportation process by filtering the historical data of the temperature drop of molten iron and classifying whether it is covered, so that the data meets the requirements of BP neural network training.
[0038] (3) The present invention utilizes the training module of the BP neural network model to establish a three-layer neural network structure, and adjusts the training parameters based on the training effect of the verification model using a verification data set, thereby obtaining a high-precision molten iron temperature drop prediction model. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] Figure 1 is a flow chart of the method described in the present invention; DETAILED DESCRIPTION
[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Unless otherwise defined, the technical terms or scientific terms used herein should be the usual meanings understood by people with general skills in the field to which the present disclosure belongs. What is not described in detail in the present invention is the known technology of technicians in this technical field.
[0041] like Figure 1 As shown, the present invention provides a method for predicting molten iron temperature drop based on BP neural network, comprising:
[0042] The historical molten iron temperature drop data of the molten iron transportation process of the torpedo tank truck is collected. Each item of the historical molten iron temperature drop data of the molten iron transportation process of the torpedo tank truck includes: the type of torpedo tank truck, the molten iron loading capacity, the initial molten iron temperature, and the molten iron temperature drop value or the working layer temperature drop value of each stage of the corresponding molten iron transportation process. Specifically, the molten iron transportation process includes the heavy tank waiting stage, the heavy tank transportation stage, the tank emptying waiting stage, the empty tank transportation stage, and the waiting stage for iron. The historical molten iron temperature drop data of each torpedo tank truck molten iron transportation process also includes the molten iron heavy tank waiting time, the molten iron heavy tank transportation time, the tank emptying waiting time, the empty tank transportation time, the empty tank transportation time, and the iron receiving waiting time.
[0043] The historical molten iron temperature drop data collected during the molten iron transportation process in the tank car are screened and classified, the abnormal data in the data are removed, and the valid data is structured and stored to obtain structured data sets corresponding to each stage of the molten iron transportation process.
[0044] The structured data set is divided into a training data set and a validation data set in proportion, a BP neural network model is constructed, and the constructed BP neural network model is trained and verified by the training data set and the validation data set respectively, to obtain a prediction model for the molten iron temperature drop in the molten iron transportation process.
[0045] The temperature drop value at each stage of the molten iron transportation process is predicted through the molten iron temperature drop prediction model.
[0046] In this embodiment, the historical molten iron temperature drop data collected during the molten iron transportation process of torpedo tank cars are first divided into open torpedo tank car molten iron transportation data and covered torpedo tank car molten iron transportation data, and then the data of the two types of data under the same working conditions are sorted out respectively, with 95% as the confidence interval, and the molten iron temperature drop data with abnormal data are removed by observing the normal distribution of each data item. Finally, the effective and reasonable temperature drop data are structured and stored to form multiple structured data sets corresponding to each stage of the molten iron transportation process, that is, 6 structured data sets, corresponding to the heavy tank waiting stage, heavy tank transportation stage, empty tank waiting stage, empty tank transportation back stage and waiting for iron receiving stage; each data item in the structured data set contains 3 columns, and the first to second columns are feature columns: the first column is the molten iron temperature or the working layer temperature, and the second column is the required time; the third column is the label column: the molten iron temperature drop value or the working layer temperature drop value.
[0047] In this embodiment, the structured data set is divided into a training data set and a verification data set in a ratio of 8:2. The training data set is used to train the model, and the verification data set is used to verify the effect of the training. When the expected effect is achieved, the training is stopped to obtain a prediction model for the temperature drop of molten iron in the molten iron transportation process.
[0048] The BP neural network model has three layers, namely the input layer, hidden layer and output layer from top to bottom. The input layer dense_input has 20 neurons, the hidden layer dense_1 has 10 neurons, and the output layer dense_output has 1 neuron. The optimization algorithm used in the training of the BP neural network model is the Levenberg-Marquardt algorithm.
[0049] In this embodiment, when the temperature drop value of each stage of the molten iron transportation process is predicted by the molten iron temperature drop prediction model during the molten iron transportation process, the torpedo tank car model, the molten iron loading amount, the time value of each stage of the molten iron transportation process, and the molten iron temperature value or the working layer temperature value at the beginning of each stage are used as input parameters of the molten iron temperature drop prediction model during the molten iron transportation process to predict the predicted temperature drop value of the corresponding stage, specifically including:
[0050] The initial temperature of molten iron and the waiting time of the molten iron reloading tank are taken as input parameters to predict the temperature drop of molten iron during the waiting process of the molten iron reloading tank.
[0051] The molten iron transportation temperature and the molten iron heavy tank transportation time are taken as input parameters to predict the molten iron temperature drop during the molten iron heavy tank transportation process.
[0052] The molten iron temperature and the waiting time for molten iron pouring are taken as input parameters to predict the molten iron temperature drop during the waiting process of molten iron pouring.
[0053] The empty tank waiting time is taken as an input parameter to predict the working layer temperature drop value during the empty tank waiting process.
[0054] The empty tank return time is used as an input parameter to predict the working layer temperature drop during the empty tank return process.
[0055] The iron-receiving waiting time is taken as an input parameter to predict the temperature drop of the working layer during the iron-receiving waiting process.
[0056] Embodiment 2:
[0057] The present application provides a molten iron temperature drop prediction system based on BP neural network, comprising:
[0058] The module for collecting historical molten iron temperature drop data during molten iron transportation is used to collect historical molten iron temperature drop data during molten iron transportation by torpedo tank cars. Each item of historical molten iron temperature drop data during molten iron transportation by torpedo tank cars includes: torpedo tank car model, molten iron loading capacity, initial molten iron temperature, and molten iron temperature drop values or working layer temperature drop values at each stage of the molten iron transportation process.
[0059] The module for screening and classifying historical data on molten iron temperature drop during molten iron transportation is used to screen and classify the historical data on molten iron temperature drop during molten iron transportation by tank cars, remove abnormal data from the data, and perform structured storage on valid data to obtain structured data sets for each stage of the molten iron transportation process.
[0060] The data training and verification module is used to divide the structured data set into a training data set and a verification data set in proportion, construct a BP neural network model, and train and verify the constructed BP neural network model through the training data set and the verification data set respectively, so as to obtain a molten iron temperature drop prediction model during the molten iron transportation process.
[0061] The temperature drop prediction module is used to predict the temperature drop value at each stage of the molten iron transportation process through the molten iron temperature drop prediction model of the molten iron transportation process.
[0062] In this embodiment, the temperature drop prediction module uses the type of torpedo tank car, the molten iron loading amount, the time value of each stage of the molten iron transportation process, and the molten iron temperature value or the working layer temperature value at the beginning of each stage as the input parameters of the molten iron temperature drop prediction model during the molten iron transportation process, and predicts the temperature drop value of the corresponding stage of the molten iron transportation process through multiple sub-modules, and the sub-modules include:
[0063] The re-tank waiting stage submodule is used to obtain data for predicting the molten iron re-tank waiting stage, including: the initial temperature of the molten iron and the waiting time of the molten iron re-tank as input parameters, and predict the temperature drop of the molten iron during the molten iron re-tank waiting process.
[0064] The heavy tank transportation stage submodule is used to obtain data for predicting the heavy tank transportation stage, including: molten iron transportation temperature and molten iron heavy tank transportation time as input parameters, and predict the temperature drop of molten iron during the molten iron heavy tank transportation process;
[0065] The tank pouring waiting stage submodule is used to obtain data of the tank pouring waiting stage, including: molten iron temperature and tank pouring waiting time as input parameters, and predict the molten iron temperature drop during the molten iron pouring waiting process.
[0066] The empty tank waiting stage submodule is used to obtain the data of the empty tank waiting stage, including: the empty tank waiting time is used as an input parameter, and the temperature drop value of the working layer in the empty tank waiting stage is predicted.
[0067] The empty tank return stage submodule is used to obtain the data of the empty tank return stage, including: the empty tank return time is used as an input parameter, and the temperature drop value of the working layer in the return stage is predicted.
[0068] The iron receiving waiting stage submodule is used to obtain data of the iron receiving waiting stage, including: the iron receiving waiting time is used as an input parameter to predict the temperature drop value of the working layer in the iron receiving waiting stage.
[0069] Embodiment 3:
[0070] The present application provides an electronic device, including a memory and a processor, wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the molten iron temperature drop prediction method of Example 1.
[0071] Embodiment 4:
[0072] The present application provides a readable storage medium having computer instructions stored thereon, wherein the computer instructions, when executed by a processor, implement the method for predicting the temperature drop of molten iron of Example 1.
Claims
1. A method for predicting molten iron temperature drop based on BP neural network, characterized in that: include: Collect historical molten iron temperature drop data during the molten iron transportation process of the torpedo tanker, each item of the historical molten iron temperature drop data during the molten iron transportation process of the torpedo tanker includes: torpedo tanker model, molten iron loading, initial molten iron temperature, and molten iron temperature drop values or working layer temperature drop values corresponding to each stage of the molten iron transportation process; The collected historical molten iron temperature drop data of the molten iron transportation process of the tank car are screened and classified, abnormal data in the data are removed, and valid data are structured and stored to obtain structured data sets corresponding to each stage of the molten iron transportation process; The structured data set is divided into a training data set and a verification data set in proportion, a BP neural network model is constructed, and the constructed BP neural network model is trained and verified by the training data set and the verification data set respectively, so as to obtain a prediction model of molten iron temperature drop in the molten iron transportation process; The temperature drop value at each stage of the molten iron transportation process is predicted by the molten iron temperature drop prediction model during the molten iron transportation process.
2. The method for predicting molten iron temperature drop according to claim 1, characterized in that: The molten iron transportation process includes: the heavy tank waiting stage, the heavy tank transportation stage, the empty tank waiting stage, the empty tank transportation stage and the waiting stage for receiving iron; each item of the historical molten iron temperature drop data of the torpedo tank truck molten iron transportation process also includes: the molten iron heavy tank waiting time, the molten iron heavy tank transportation time, the empty tank waiting time, the empty tank waiting time, the empty tank transportation time and the iron receiving waiting time.
3. The method for predicting molten iron temperature drop according to claim 1, characterized in that: The historical molten iron temperature drop data collected during the molten iron transportation process by torpedo tank cars are divided into open torpedo tank car molten iron transportation data and covered torpedo tank car molten iron transportation data. The data of the two types of data under the same working conditions are sorted out respectively, and 95% is used as the confidence interval to observe the normal distribution of each data item, and the molten iron temperature drop data with abnormal data are removed. Finally, the effective and reasonable temperature drop data are structured and stored to form multiple structured data sets corresponding to each stage of the molten iron transportation process; each data item in the structured data set contains 3 columns, and the first to second columns are feature columns: the first column is the molten iron temperature or the working layer temperature, and the second column is the required time; the third column is the label column: the molten iron temperature drop value or the working layer temperature drop value.
4. The method for predicting molten iron temperature drop according to claim 1, characterized in that: The structured dataset is divided into a training dataset and a validation dataset in a ratio of 8:
2.
5. The method for predicting molten iron temperature drop according to claim 1, characterized in that: The BP neural network model has three layers, namely, an input layer, a hidden layer and an output layer from top to bottom. The input layer has 20 neurons, the hidden layer has 10 neurons, and the output layer has 1 neuron. The optimization algorithm used in training the BP neural network model is the Levenberg-Marquardt algorithm.
6. The method for predicting molten iron temperature drop according to claim 1, characterized in that: When the temperature drop value of each stage of the molten iron transportation process is predicted by the molten iron temperature drop prediction model during the molten iron transportation process, the torpedo tank car model, the molten iron loading amount, the time value of each stage of the molten iron transportation process, and the molten iron temperature value or the working layer temperature value at the beginning of each stage are used as input parameters of the molten iron temperature drop prediction model during the molten iron transportation process to predict the temperature drop value of the corresponding stage, specifically including: Taking the initial temperature of molten iron and the waiting time of the molten iron reloading tank as input parameters, the temperature drop of molten iron during the waiting process of the molten iron reloading tank is predicted; Taking the molten iron transportation temperature and the molten iron heavy tank transportation time as input parameters, the molten iron temperature drop during the molten iron heavy tank transportation process is predicted; Taking the molten iron temperature and the waiting time for molten iron pouring as input parameters, the temperature drop of molten iron during the waiting process of molten iron pouring is predicted; Taking the empty tank waiting time as an input parameter, the temperature drop value of the working layer during the empty tank waiting process is predicted; The empty tank return time is used as an input parameter to predict the working layer temperature drop during the empty tank return process; The iron-receiving waiting time is taken as an input parameter to predict the temperature drop of the working layer during the iron-receiving waiting process.
7. A molten iron temperature drop prediction system based on BP neural network, characterized in that: include: A module for collecting historical molten iron temperature drop data during molten iron transportation, which is used to collect historical molten iron temperature drop data during molten iron transportation by torpedo tankers, wherein each item of historical molten iron temperature drop data during molten iron transportation by torpedo tankers includes: torpedo tanker model, molten iron loading, initial molten iron temperature, and molten iron temperature drop values or working layer temperature drop values corresponding to each stage of the molten iron transportation process; A module for screening and classifying historical molten iron temperature drop data during molten iron transportation, which is used to screen and classify the collected historical molten iron temperature drop data during the molten iron transportation process by the tank car, remove abnormal data in the data, and perform structured storage on the valid data to obtain structured data sets corresponding to each stage of the molten iron transportation process; A data training and verification module, which is used to divide the structured data set into a training data set and a verification data set in proportion, construct a BP neural network model, and respectively train and verify the constructed BP neural network model through the training data set and the verification data set to obtain a molten iron temperature drop prediction model during molten iron transportation; The temperature drop prediction module is used to predict the temperature drop value at each stage of the molten iron transportation process through the molten iron temperature drop prediction model of the molten iron transportation process.
8. The molten iron temperature drop prediction system based on BP neural network according to claim 7, characterized in that: The temperature drop prediction module uses the torpedo tank car model, the molten iron loading amount, the time value of each stage of the molten iron transportation process, and the molten iron temperature value or the working layer temperature value at the beginning of each stage as the input parameters of the molten iron temperature drop prediction model of the molten iron transportation process, and predicts the temperature drop value of the corresponding stage of the molten iron transportation process through submodules, and the submodules include: The re-tank waiting stage submodule is used to obtain data for predicting the molten iron re-tank waiting stage, including: the molten iron initial temperature and the molten iron re-tank waiting time as input parameters, and predict the molten iron temperature drop during the molten iron re-tank waiting process; The heavy tank transportation stage submodule is used to obtain data for predicting the heavy tank transportation stage, including: molten iron transportation temperature and molten iron heavy tank transportation time as input parameters, and predict the temperature drop of molten iron during the molten iron heavy tank transportation process; The tank pouring waiting stage submodule is used to obtain the data of the tank pouring waiting stage, including: molten iron temperature and tank pouring waiting time as input parameters, and predict the temperature drop of molten iron during the molten iron pouring waiting process; The empty tank waiting stage submodule is used to obtain the data of the empty tank waiting stage, including: the empty tank waiting time is used as an input parameter to predict the temperature drop value of the working layer during the empty tank waiting stage; The empty tank return stage submodule is used to obtain the data of the empty tank return stage, including: the empty tank return time is used as an input parameter to predict the temperature drop value of the working layer in the return stage; The iron receiving waiting stage submodule is used to obtain data of the iron receiving waiting stage, including: the iron receiving waiting time is used as an input parameter to predict the temperature drop value of the working layer in the iron receiving waiting stage.
9. An electronic device comprising a memory and a processor, wherein: The memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the molten iron temperature drop prediction method according to any one of claims 1 to 6.
10. A readable storage medium having computer instructions stored thereon, wherein: When the computer instruction is executed by a processor, the method for predicting the temperature drop of molten iron as described in any one of claims 1 to 6 is implemented.