Model training method, temperature prediction method, electronic device, storage medium, and computer program product

Abstract

This application provides a model training method, a temperature prediction method, an electronic device, a storage medium, and a computer program product; the model training method includes: acquiring sample data, the sample data including first data, second data, and third data, the first data including process parameters and corresponding time steps related to the temperature of the molten material in the melting device, the second data including the temperature of the molten material in the melting device predicted by the first model and the corresponding time step, and the third data including the actual temperature of the molten material in the melting device and the corresponding time step; training a second model using at least the acquired sample data, the second model being used to correct the temperature prediction of the molten material predicted by the first model.

Description

Technical Field

[0001] This application relates to the field of model training technology, specifically to a model training method, a temperature prediction method, an electronic device, a storage medium, and a computer program product. Background Technology

[0002] In fields such as chemical synthesis, bio-fermentation, materials preparation, and environmental engineering, the temperature measurement of high-temperature solutions is a critical step that directly affects process safety, reaction efficiency, and product consistency. Most of the methods for predicting solution temperature in related technologies rely on fixed parameters, which are not robust enough to environmental fluctuations and systematic errors caused by sensors. This leads to a significant decrease in the accuracy of temperature prediction results, making it difficult to meet the measurement requirements under complex working conditions. Summary of the Invention

[0003] This application provides a model training method, a temperature prediction method, an electronic device, a storage medium, and a computer program product.

[0004] The model training method provided in this application includes: Acquire sample data, which includes first data, second data, and third data. The first data includes process parameters and corresponding time steps related to the temperature of the melt in the melting device. The second data includes the temperature of the melt in the melting device predicted by the first model and the corresponding time step. The third data includes the actual temperature of the melt in the melting device and the corresponding time step. At least the acquired sample data is used to train a second model, which is used to correct the temperature prediction of the melt predicted by the first model.

[0005] The temperature prediction method provided in this application includes: Obtain the fourth and fifth data corresponding to the melting device at the first moment. The fourth data includes process parameters related to the temperature of the melt in the melting device and the corresponding time step. The fifth data includes the temperature of the melt in the melting device predicted by the first model at the historical moment corresponding to the first moment and the corresponding time step. Based on the fourth data, the temperature of the melt in the melting device at the first moment is predicted using the first model, and based on the fourth and fifth data, the corresponding correction amount is predicted using the second model, which is obtained using the model training method provided in any embodiment of this application. The temperature of the molten material in the melting device at the first moment is obtained by using the predicted temperature and the predicted correction amount.

[0006] The electronic device provided in this application includes a processor and a memory. The memory is used to store computer programs, and the processor is used to call and run the computer programs stored in the memory to execute the model training method provided in any embodiment of this application.

[0007] The storage medium provided in this application embodiment is used to store a computer program, which causes a computer to execute the model training method provided in any embodiment of this application.

[0008] The model training method, temperature prediction method, electronic device, storage medium, and computer program product provided in this application embodiment train a second model by using temperature-related process parameters, time steps, temperature prediction values ​​of the first model, and actual temperatures in the melting device. Since the actual temperature data used for training and the temperature prediction values ​​of the first model naturally include the influence of environmental fluctuations on the melting process, the second model can automatically identify and offset the interference caused by environmental changes by correcting only the errors. Moreover, without adding environmental sensors or modifying the first model, it can achieve accurate compensation for temperature prediction in a low-coupling manner, significantly improving the anti-interference ability and prediction stability under complex working conditions. Attached Figure Description

[0009] Figure 1 A schematic diagram of a cold crucible provided in an embodiment of this application; Figure 2 This is a schematic diagram of the temperature measuring rod structure provided in an embodiment of this application; Figure 3 A schematic diagram illustrating the implementation process of the model training method provided in this application embodiment; Figure 4 This is a schematic diagram showing the comparison of current data before and after the implementation of this application. Figure 5 A schematic diagram showing the comparison of the mixing position and rotating shaft speed data before and after the present application embodiment; Figure 6 A schematic diagram illustrating the implementation process of the temperature prediction method provided in this application embodiment; Figure 7 This is a schematic diagram of the structure of the temperature prediction model provided in the embodiments of this application; Figure 8 The verification effect comparison chart provided for the embodiments of this application; Figure 9 This is a schematic diagram of the structure of the model training device provided in the embodiments of this application; Figure 10 This is a schematic diagram of the structure of the temperature prediction device provided in the embodiments of this application; Figure 11 A schematic structural diagram of an electronic device provided in the embodiments of this application; Figure 12 This is a schematic structural diagram of the chip provided in an embodiment of this application. Detailed Implementation

[0010] The technical solutions of the embodiments of this application will now be described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0011] It should be noted that, in the embodiments of this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, in the embodiments of this application, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0012] In the description of the embodiments of this application, the term "correspondence" may indicate that there is a direct or indirect correspondence between two things, or that there is an association between two things, or that there is a relationship of instruction and being instructed, configuration and being configured, etc.

[0013] To facilitate understanding of the technical solutions of the embodiments of this application, the relevant technologies of the embodiments of this application are described below. The following relevant technologies are optional solutions and can be combined with the technical solutions of the embodiments of this application in any way, and they all fall within the protection scope of the embodiments of this application.

[0014] In fields such as chemical synthesis, bio-fermentation, materials preparation, and environmental engineering, temperature measurement of high-temperature solutions is a critical step directly affecting process safety, reaction efficiency, and product consistency. For example, in cold crucible glass curing, high-radioactive waste liquid (also known as high-level radioactive waste liquid) and a glass-forming agent are melted at 1100-1300°C using high-frequency induction heating to form a stable glass. The glass-forming agent and the high-level radioactive waste liquid melt at high temperatures to form molten glass, which is then cooled and solidified into glass. In this process, the temperature of the molten glass is a core parameter affecting the vitrification effect, equipment safety, and the quality of the final product. Temperature monitoring during cold crucible glass curing presents unique challenges: 1. Extreme environmental limitations: The temperature of molten glass is as high as 1100-1300°C, and it is highly corrosive and radioactive, resulting in a very short lifespan for traditional contact temperature sensors.

[0015] 2. Measurement blind zone: It is difficult to install physical sensors inside the crucible and in the area where the molten glass flows.

[0016] 3. Process complexity: Multiple parameters such as high-frequency power supply parameters, mechanical stirring status, and feed and discharge rates are strongly coupled, and a single parameter is difficult to accurately reflect the true temperature.

[0017] 4. High safety requirements: Improper temperature control may lead to equipment damage or leakage of radioactive materials.

[0018] refer to Figure 1 , Figure 1 This is a schematic diagram of the cold crucible provided in an embodiment of this application, as shown below. Figure 1 As shown, the upper diagram is a schematic diagram of the principle, and the lower diagram is a schematic diagram of the actual object. Mechanical stirring is used to mix the materials evenly; cooling return water and cooling water are used to cool relevant parts of the equipment; molten glass beads are the form of glass in its molten state; convection induction heating is the method of heating the materials; induction coil is the key component for realizing induction heating; discharge outlet is used for material discharge; temperature measuring device is used to measure temperature; molten glass is the form of the processed glass; glass cold shell is the part formed by cooling the glass surface; crucible bottom is the bottom structure of the crucible; hot cell remote manipulator can remotely operate the relevant equipment in a hot chamber environment; and the cold crucible is the container that holds the materials for processing. To withstand high temperatures, the temperature measuring rod is protected by an external circulating water-cooling jacket. The outer shell of the temperature measuring head, which directly contacts the molten metal, is made of platinum. Thanks to its excellent corrosion resistance, it can maintain its physical and chemical integrity and perform temperature measurement without active cooling. However, platinum inherently exhibits a significant decrease in tensile strength under prolonged high temperatures. In historical operations, process fluctuations caused molten glass to solidify and adhere to localized areas of the temperature measuring rod. This solidified glass was subjected to strong shear forces during subsequent mechanical stirring, generating a mechanical load on the temperature measuring rod far exceeding expectations. This stress ultimately exceeded the load-bearing limit of the high-temperature platinum probe, leading to brittle fracture and subsequent plastic bending of the probe body. (Reference) Figure 2 , Figure 2 This is a schematic diagram of the temperature measuring rod structure provided in the embodiments of this application, as shown below. Figure 2As shown, 2-a shows the overall structure of the temperature measuring rod; 2-b is a normal temperature measuring head; 2-c and 2-d show the state after the temperature measuring head is broken. In 2-d, the broken thermocouple wire can be observed; in 2-e, it can be seen that the temperature measuring rod has color layering. The lower part is black because a thin glass shell is attached, and the temperature measuring rod is bent at the layering point.

[0019] While virtual temperature measurement technology exists in related technologies, it has significant shortcomings in cold crucible glass curing applications: it lacks a dedicated model for the temperature characteristics of high-temperature molten glass, fails to consider the special characteristics of data acquisition in radioactive environments, its training strategy is not adapted to the scenario of gradual sensor failure in actual operation, and it responds slowly to dynamic changes in molten glass temperature.

[0020] refer to Figure 3 , Figure 3 This is a schematic diagram illustrating the implementation process of the model training method provided in the embodiments of this application, as shown below. Figure 3 As shown, the model training method provided in this application embodiment includes the following steps: Step 101: Obtain sample data, which includes first data, second data, and third data. The first data includes process parameters related to the temperature of the melt in the melting device and the corresponding time step. The second data includes the temperature of the melt in the melting device predicted by the first model and the corresponding time step. The third data includes the actual temperature of the melt in the melting device and the corresponding time step.

[0021] In this embodiment, the melting device can be a cold crucible or other melting equipment, and this embodiment does not limit it.

[0022] In this embodiment, the molten material can be molten glass or other molten materials, and this embodiment does not limit the molten material.

[0023] In this embodiment, the cold crucible engineering prototype can be continuously operated on the cold platform for more than 90 days, and data is collected at a frequency of once every 5 seconds, accumulating 1,486,080 original historical data entries. Each data entry includes nearly 100 process parameters for the operation of equipment such as the cold crucible and calcining furnace. Nine sets of process parameters directly related to the temperature measurement data of the cold crucible are selected: high-frequency power supply voltage, current, stirring paddle position, lifting shaft torque, lifting shaft speed, rotating shaft torque, rotating shaft speed, discharge rate, and feeding rate. The voltage and current of the high-frequency power supply determine the input power of the cold crucible, thus affecting the temperature field of the cold crucible. The movement of the stirring paddle is controlled by two rotary motors, which control the up-and-down movement and rotation of the stirring paddle. The position of the stirring paddle, the speed of the lifting shaft, and the rotation speed of the rotating shaft determine the operating state of the stirring paddle, which directly affects the glass melt flow field and thus the temperature field. The torque of the lifting shaft and the torque of the rotating shaft reflect the viscosity of the glass melt. The viscosity of the glass melt is related to the temperature, so it can be used to predict the temperature of the glass melt. During the discharge process, the bottom medium-frequency induction heating discharge pipe is activated, affecting the temperature field distribution of the glass melt. During the batch feeding of glass beads, the cold glass beads enter from the top of the cold crucible, directly affecting the temperature field distribution. Therefore, the above nine sets of process parameters are considered as sample data and used as input samples for the second model learning sample.

[0024] Based on this, in the embodiments of this application, the process parameters include: Voltage, current, agitator position, lifting shaft torque, lifting shaft speed, rotating shaft torque, rotating shaft speed, discharge rate, and feed rate.

[0025] In this embodiment, the original dataset contains abnormal data segments generated during the experiment due to equipment failure and subsequent pauses, garbled data caused by abnormal recording functions of the control system, and unavoidable measurement noise, making it unsuitable for direct model training. Therefore, this embodiment establishes strict data cleaning rules: first, invalid data segments during and before / after equipment failures are removed; second, garbled records that clearly exceed the physical range are identified and deleted; finally, obvious impulse noise is smoothed using sliding window filtering. After the above cleaning process, the effective data volume is 1,293,370 records, and the data quality is significantly improved. To eliminate the impact of differences in the dimensions of different process parameters on model training, all data are normalized to the [0,1] interval.

[0026] Using high-frequency power supply current, stirring position, and rotation speed data as examples, the comparison of data before and after cleaning and normalization is shown. Figure 4 This is a schematic diagram comparing the current data before and after the implementation of this application. Figure 5 This is a schematic diagram showing the comparison of the position of the stirring paddle and the rotational speed of the rotating shaft before and after the embodiments of this application. Figure 4 and Figure 5From top to bottom, the data consists of raw data, cleaned data, and normalized data. The data marked in the diagram in the raw data are the data from the high-frequency power supply current data during the test pause and the abnormal garbled data in the agitator position and rotor speed data. After data cleaning, the above data has been completely cleaned up. The raw data fluctuated greatly and contained many abnormally high values. Data cleaning effectively reduced outliers and made the data smoother. Data normalization uniformly scaled the data to the range of 0-1, eliminating the influence of units and preserving the relative distribution trend. Such a data sequence is smooth and has uniform units, laying a reliable data foundation for subsequent machine learning training.

[0027] In this embodiment, both the first and second models can employ a Long Short-Term Memory (LSTM) network. LSTM is a special type of recurrent neural network capable of learning long-term dependencies, making it particularly suitable for processing time-series data. LSTM uses gating mechanisms (input gate, forget gate, and output gate) to control the flow of information, avoiding the vanishing gradient problem in traditional neural networks. It is understood that other neural networks can also be used for the first and second models, and this embodiment does not limit their use.

[0028] Step 102: At least using the acquired sample data, train the second model, which is used to correct the temperature prediction of the melt predicted by the first model.

[0029] In this embodiment, the training process of the second model can employ planned sampling, a training strategy that gradually transitions from using real data (third data) to using model-predicted data (second data) during training. This is because in actual model deployment, temperatures are predicted by the first model, which is an autoregressive prediction scenario. Autoregressive prediction refers to using the model's own prediction output as input for subsequent predictions. In a virtual temperature measurement scenario, this means that when the model predicts the temperature at the next time step, it uses the predicted temperature value from the previous time step instead of the actual measured value. The planned sampling strategy helps the model adapt to the autoregressive prediction scenario in actual deployment, avoiding performance degradation caused by mismatch between training and deployment environments. Early training phase: Primarily uses real temperature data as input to help the model quickly learn basic temperature change patterns; Mid-training phase: Gradually introduces the first model's predicted temperature as part of the input to train the model's ability to handle prediction errors; Late training phase: Uses the first model's predicted temperature entirely as input to simulate the actual deployment scenario.

[0030] In this embodiment, the input to the second model is the first data and the temperature sequence. When training the Nth batch, N is an integer greater than or equal to 1. In the temperature sequence that is the input to the second model, the real measured value is selected according to the first probability Pn, that is, the third data, and the predicted value of the upstream network is selected according to the probability 1-Pn, that is, the second data. The first probability decreases as the number of iterations increases, so that during the training process, the model gradually transitions from using the real data (third data) to using the model predicted data (second data).

[0031] Based on this, in the embodiments of this application, during the training of the second model, the proportion of the third data in the sample data used in the Nth iteration is determined based on the first probability, where N is an integer greater than or equal to 1.

[0032] In this embodiment of the application, the first probability decreases as the number of iterations increases.

[0033] In this embodiment, the first model is trained based at least on the first data, and can be a long-term modeling of 1200 time steps (6000 seconds) to fully capture the large inertial characteristics of the glass melt temperature. Of course, in practical applications, other numbers of time steps of the first data can also be used, such as 1000 time steps (5000 seconds) of the first data. This embodiment does not limit this.

[0034] Therefore, in this embodiment of the application, the first model is trained at least based on the first data.

[0035] In this embodiment, the model can also be subjected to actual tests, including: prediction accuracy testing under normal operating conditions; robustness testing under sensor aging conditions; and stability testing under process parameter fluctuations. Process parameter fluctuations refer to prediction accuracy testing (stability) under process fluctuations caused by factors such as operator error leading to abnormally high input current, impeller malfunction leading to a significant reduction in mixing capacity, and abnormal feeding of glass beads or high-level radioactive waste product leading to changes in the conductivity of the molten glass. Sensor failure refers to the loss of input process parameter data due to sensor failure (e.g., torque data cannot be obtained after impeller motor failure). The system will automatically supplement the pre-set parameters to ensure that the model can continue to predict the molten glass temperature smoothly, and then test the prediction accuracy (robustness) under this condition.

[0036] refer to Figure 6 , Figure 6 This is a schematic diagram illustrating the implementation process of the temperature prediction method provided in the embodiments of this application, as shown below. Figure 6 As shown, the temperature prediction method provided in this application embodiment includes the following steps: Step 201: Obtain the fourth and fifth data corresponding to the melting device at the first moment. The fourth data includes process parameters related to the temperature of the melt in the melting device and the corresponding time step. The fifth data includes the temperature of the melt in the melting device predicted by the first model at the historical moment corresponding to the first moment and the corresponding time step.

[0037] In this embodiment, the fourth data can be the fourth data from the 1200th time step (6000 seconds) closest to the first moment, fully capturing the large inertial characteristics of the glass melt temperature. Of course, the first data from other time steps can also be used, such as the first data from 1000 time steps (5000 seconds), and this embodiment does not limit this.

[0038] In this embodiment of the application, the historical moment can be the 500th moment closest to the first moment, that is, the 500th time step closest to the first moment, and the fifth data can be the temperature of the melt in the melting device and the corresponding time step predicted by the first model in the 500th time step closest to the first moment.

[0039] Step 202: Based on the fourth data, the temperature of the molten material in the melting device at the first moment is predicted using the first model, and based on the fourth and fifth data, the corresponding correction amount is predicted using the second model, which is obtained using the model training method provided in any embodiment of this application.

[0040] By predicting temperature using the first model, the long-term trend of temperature changes can be effectively captured. By correcting the temperature predicted by the first model using the second model, short-term dynamic fluctuations caused by high-frequency process disturbances can be resisted, while taking into account the large inertia characteristics and instantaneous dynamic changes of the process.

[0041] Step 203: Using the predicted temperature and the predicted correction amount, obtain the temperature of the molten material in the melting device at the first moment.

[0042] In this embodiment, the temperature of the melt in the melting device at the first moment is the sum of the temperature predicted by the first model and the correction amount predicted by the second model.

[0043] In this embodiment of the application, after the first model predicts the temperature of the melt in the melting device at the first moment, the temperature of the melt in the melting device at the first moment predicted by the first model is saved as a historical temperature for subsequent correction by the second model.

[0044] Based on this, in the embodiments of this application, the method further includes: Save the temperature of the molten material in the melting device corresponding to the first moment predicted by the first model.

[0045] refer to Figure 7 , Figure 7 This is a schematic diagram of the structure of the temperature prediction model provided in the embodiments of this application, as shown below. Figure 7 As shown, the temperature prediction model comprises a first model and a second model, which are cascaded together. The upstream long-period prediction network (first model) takes a long-term process parameter sequence (e.g., 5000 seconds) as input and outputs a preliminary temperature trend prediction. The downstream short-period correction network (second model) is responsible for dynamic fine-tuning. It receives a recent (e.g., 500 seconds) process parameter sequence Xtshort and the corresponding historical temperature data sequence as input and outputs a correction amount ΔT. Finally, the complete prediction output of the temperature prediction model is composed of the superposition of the outputs of the two networks.

[0046] The entire training process of the temperature prediction model is carried out in stages: first, the first model is pre-trained independently; then, the second model is trained, and planned sampling is introduced during the training of the second model. Specifically, during the training of the nth batch, for each time step in the temperature sequence, the temperature value is selected using the actual measured value with a first probability Pn, and the predicted value of the first model is selected with a probability 1-Pn. The probability Pn decays with the increase of training batches. This method allows the correction network to converge quickly with the actual value in the early stage of training, and gradually adapt to the predicted value input in the later stage, thereby enhancing the robustness of the model against the accumulation of prediction errors when deployed in actual situations.

[0047] refer to Figure 8 , Figure 8 The verification effect comparison chart provided in the embodiments of this application is as follows: Figure 8 As shown, Figure 8 The mean squared error (MSE) and prediction accuracy (J) on the test set are shown. As can be seen from the figure, the temperature prediction model provided by this application has a significantly lower mean squared error than decision trees, feedforward neural networks, and temporal neural networks on all test sets. Its prediction accuracy is the highest, reaching 95.231%, and the average temperature error is less than 50°C, which fully meets the conditions for replacing physical temperature measurement devices.

[0048] The model training and temperature prediction methods provided in this application construct a loosely coupled soft-sensor architecture of basic prediction + correction. Supervised training is performed using process parameters, time steps, and temperature sequences, implicitly encoding the nonlinear compensation laws of cold shell thermal resistance evolution and environmental disturbances under extreme conditions. This achieves high-precision prediction without ideal boundary assumptions. During operation, it relies solely on highly reliable process variables for independent deduction, seamlessly maintaining monitoring continuity even when physical sensors fail. Maintenance-free process parameters (such as discharge rate, feed rate, agitator speed, and time steps) completely eliminate real-time dependence on contact-type temperature sensing elements. Simultaneously, the fusion of temporal features and operating condition inputs endows the model with forward-looking thermal inertia perception capabilities, transforming passive lag tracking into dynamic trend compensation, significantly improving response speed. The prediction mechanism provided in this application significantly reduces the deployment dependence and replacement frequency of temperature probes in radioactive environments, effectively avoiding the cumulative radiation risks associated with remote operation and maintenance.

[0049] This application also provides a model training apparatus, for reference... Figure 9 , Figure 9 This is a schematic diagram of the structure of the model training device provided in an embodiment of this application. The model training device in this embodiment includes: Model training unit: used to acquire sample data, the sample data includes first data, second data and third data, the first data includes process parameters and corresponding time steps related to the temperature of the melt in the melting device, the second data includes the temperature of the melt in the melting device predicted by the first model and the corresponding time step, and the third data includes the actual temperature of the melt in the melting device and the corresponding time step. The model training unit is used to train a second model using at least the acquired sample data. The second model is used to correct the temperature prediction of the melt predicted by the first model.

[0050] In this embodiment of the application, during the training of the second model, the proportion of the third data in the sample data used in the Nth iteration is determined based on the first probability, where N is an integer greater than or equal to 1.

[0051] In this embodiment of the application, the first probability decreases as the number of iterations increases.

[0052] In this embodiment of the application, the first model is trained based at least on the first data.

[0053] In this embodiment of the application, the process parameters include: Voltage, current, agitator position, lifting shaft torque, lifting shaft speed, rotating shaft torque, rotating shaft speed, discharge rate, and feed rate.

[0054] Those skilled in the art should understand that Figure 9 The functions of each unit in the model training device shown can be understood by referring to the relevant descriptions of the aforementioned methods. Figure 9 The functions of each unit in the model training device shown can be implemented by a program running on a processor or by specific logic circuits.

[0055] This application also provides a temperature prediction device, referring to... Figure 10 , Figure 10 This is a schematic diagram of the temperature prediction device provided in an embodiment of this application. The temperature prediction device in this embodiment includes: Prediction unit: used to acquire the fourth and fifth data corresponding to the melting device at the first moment. The fourth data includes process parameters related to the temperature of the melt in the melting device and the corresponding time step. The fifth data includes the temperature of the melt in the melting device and the corresponding time step predicted by the first model at the historical moment corresponding to the first moment. The prediction unit is used to predict the temperature of the melt in the melting device at the first moment based on the fourth data and the first model, and to predict the corresponding correction amount based on the fourth data and the fifth data using the second model. The second model is obtained using the model training method provided in any embodiment of this application. The prediction unit is used to obtain the temperature of the molten material in the melting device at the first moment by using the predicted temperature and the predicted correction amount.

[0056] In this embodiment of the application, the prediction unit is used to store the temperature of the molten material in the melting device corresponding to the first moment predicted by the first model.

[0057] Those skilled in the art should understand that Figure 10 The functions of each unit in the temperature prediction device shown can be understood by referring to the relevant description of the aforementioned method. Figure 10 The functions of each unit in the temperature prediction device shown can be implemented by a program running on a processor or by specific logic circuits.

[0058] Figure 11 This is a schematic structural diagram of an electronic device provided in an embodiment of this application. Figure 11 The electronic device shown includes a processor 1110, which can call and run computer programs from memory to implement the model training method provided in the embodiments of this application.

[0059] Optionally, such as Figure 11 As shown, the electronic device may also include a memory 1120. The processor 1110 can retrieve and run computer programs from the memory 1120 to implement the model training method provided in the embodiments of this application.

[0060] The memory 1120 can be a separate device independent of the processor 1110, or it can be integrated into the processor 1110.

[0061] Optionally, such as Figure 11 As shown, the electronic device may also include a transceiver 1130, and the processor 1110 may control the transceiver 1130 to communicate with other devices. Specifically, it may send information or data to other devices or receive information or data sent by other devices.

[0062] The transceiver 1130 may include a transmitter and a receiver. The transceiver 1130 may further include an antenna, and the number of antennas may be one or more.

[0063] This electronic device can implement the corresponding processes implemented by the model training device / temperature prediction device in the various methods of the embodiments of this application, which will not be described in detail here for the sake of brevity.

[0064] For example, embodiments of this application also provide a computer program product, including a computer program that can be executed by a processor 1110 of an electronic device to perform the steps described in any of the foregoing methods.

[0065] Figure 12 This is a schematic structural diagram of the chip according to an embodiment of this application. Figure 12 The chip shown includes a processor 1210, which can call and run computer programs from memory to implement the methods in the embodiments of this application.

[0066] Optionally, such as Figure 12 As shown, the chip may also include a memory 1220. The processor 1210 can retrieve and run computer programs from the memory 1220 to implement the methods described in this embodiment.

[0067] The memory 1220 can be a separate device independent of the processor 1210, or it can be integrated into the processor 1210.

[0068] Optionally, the chip may also include an input interface 1230. The processor 1210 can control the input interface 1230 to communicate with other devices or chips; specifically, it can acquire information or data sent by other devices or chips.

[0069] Optionally, the chip may also include an output interface 1240. The processor 1210 can control the output interface 1240 to communicate with other devices or chips, specifically, to output information or data to other devices or chips.

[0070] This chip can be applied to the electronic devices in the embodiments of this application, and the chip can implement the corresponding processes implemented by the electronic devices in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0071] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0072] It should be understood that the processor in the embodiments of this application may be an integrated circuit chip with signal processing capabilities. In implementation, the steps of the above method embodiments can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor described above can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software units in the decoding processor. The software units can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0073] It is understood that the memory in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DR RAM). It should be noted that the memory used in the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0074] It should be understood that the above-described memory is exemplary and not a limiting description. For example, the memory in the embodiments of this application may also be static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DR RAM), etc. That is to say, the memory in the embodiments of this application is intended to include, but is not limited to, these and any other suitable types of memory.

[0075] This application also provides a storage medium for storing a computer program. This storage medium can be applied to the electronic device in this application embodiment, and the computer program causes the computer to execute the corresponding processes implemented by the electronic device in the various methods of this application embodiment; for brevity, further details are omitted here.

[0076] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0077] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0078] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0079] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0080] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0081] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or electronic device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0082] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A model training method, characterized in that, include: Acquire sample data, which includes first data, second data, and third data. The first data includes process parameters and corresponding time steps related to the temperature of the melt in the melting device. The second data includes the temperature of the melt in the melting device predicted by the first model and the corresponding time step. The third data includes the actual temperature of the melt in the melting device and the corresponding time step. At least the acquired sample data is used to train a second model, which is used to correct the temperature prediction of the melt predicted by the first model.

2. The method of claim 1, wherein, During the training of the second model, the proportion of the third data in the sample data used in the Nth iteration is determined based on the first probability, where N is an integer greater than or equal to 1.

3. The method according to claim 2, characterized in that, The first probability decreases as the number of iterations increases.

4. The method according to claim 1, characterized in that, The first model is trained based on at least the first data.

5. The method according to any one of claims 1 to 4, characterized in that, The process parameters include: Voltage, current, agitator position, lifting shaft torque, lifting shaft speed, rotating shaft torque, rotating shaft speed, discharge rate, and feed rate.

6. A temperature prediction method, characterized in that, include: Obtain the fourth and fifth data corresponding to the melting device at the first moment. The fourth data includes process parameters related to the temperature of the melt in the melting device and the corresponding time step. The fifth data includes the temperature of the melt in the melting device predicted by the first model at the historical moment corresponding to the first moment and the corresponding time step. Based on the fourth data, the temperature of the melt in the melting device at the first moment is predicted using the first model, and based on the fourth and fifth data, the corresponding correction amount is predicted using the second model, wherein the second model is obtained using the model training method of any one of claims 1 to 5. The temperature of the molten material in the melting device at the first moment is obtained by using the predicted temperature and the predicted correction amount.

7. The method according to claim 6, characterized in that, The method further includes: Save the temperature of the molten material in the melting device corresponding to the first moment predicted by the first model.

8. An electronic device, characterized in that, include: A processor and a memory, the memory for storing a computer program, the processor for calling and running the computer program stored in the memory to perform the model training method as described in any one of claims 1 to 5, or the temperature prediction method as described in claims 6 to 7.

9. A storage medium, characterized in that, Used to store computer programs that cause a computer to perform the model training method as described in any one of claims 1 to 5, or the temperature prediction method as described in claims 6 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the model training method as described in any one of claims 1 to 5, or the temperature prediction method as described in claims 6 to 7.

Images