A method, system, equipment and medium for predicting kiln temperature

Abstract

This invention provides a method, system, device, and medium for predicting kiln temperature. The method includes: obtaining a historical dataset using monitored kiln data; acquiring a first training set and a first test set based on the historical dataset; initially training and testing a model using the first training set and the first test set; further training and testing the model using a pre-acquired second training set and a second test set; and selecting the optimal temperature prediction model based on the output error of the second test set. Based on kiln data within a preset prediction period and the optimal temperature prediction model, the predicted temperature of the kiln temperature zone for the prediction period is obtained. This invention enables the prediction of future kiln temperatures. The kiln temperature prediction method provided by this invention, through two training and testing stages, can not only quickly and accurately predict future kiln temperatures but also predict the temperature change trends of each temperature zone in advance, exhibiting high reliability and low cost.

Description

Technical Field

[0001] This invention relates to the field of temperature prediction technology, and in particular to a method, system, equipment and medium for predicting kiln temperature. Background Technology

[0002] Industrial kilns are equipment constructed from refractory materials for calcining materials or firing products. They are widely used in industries such as machinery, metallurgy, petroleum, and gas, and their creation and development have played a vital role in human progress.

[0003] In existing technologies, the main components of traditional industrial kilns include: kiln lining, exhaust system, preheater, and combustion unit. To accelerate firing and shorten the firing cycle, traditional industrial kilns typically rely on operators to adjust fuel levels or power output to control the kiln's internal temperature. However, due to the instability of manual control, excessive or insufficient power adjustments are inevitable, leading to either excessively high or low kiln temperatures, thus affecting product sintering.

[0004] Currently, after the cathode material raw materials are mixed, they are loaded into a kiln to begin sintering. The raw materials pass through the heating zone, the heat preservation zone, and the cooling zone at a fixed pushing speed in the kiln. Each temperature zone needs to maintain a corresponding temperature for sintering. Under normal working conditions, the temperature of each temperature zone will fluctuate within a stable range. However, there are cases where the temperature fluctuation is large, resulting in a large deviation between the actual sintering temperature of the kiln and the set temperature, which affects the material properties and increases the company's costs. Summary of the Invention

[0005] The purpose of this invention is to provide a method for predicting kiln temperature, so as to predict the future temperature of the kiln.

[0006] To achieve the above objectives, in a first aspect, embodiments of the present invention provide a kiln temperature prediction method, the method comprising:

[0007] Acquire kiln data for any kiln temperature zone within a set time range corresponding to that temperature zone, and construct a historical dataset; the kiln data includes temperature data and power data;

[0008] The first training set and the first test set are obtained based on the historical dataset. The first training set is used to train several pre-selected initial temperature prediction models to obtain several first temperature prediction models.

[0009] Using the first test set, several of the first temperature prediction models are tested respectively, the output error of the first test set is output, and a second temperature prediction model is selected from the first temperature prediction models based on the output error of the first test set and the preset range of the first test error.

[0010] The third temperature prediction model is obtained by training several second temperature prediction models using a pre-acquired second training set.

[0011] Based on the pre-acquired second test set, several third temperature prediction models are tested respectively, the output error of the second test set is output, and the third temperature prediction model corresponding to the minimum value of the output error of the second test set is taken as the optimal temperature prediction model for the corresponding kiln temperature zone.

[0012] The kiln temperature data to be predicted, collected within a preset early prediction period, is input into the optimal temperature prediction model to obtain the predicted temperature of the kiln temperature zone during the prediction period.

[0013] Furthermore, the step of obtaining the first training set and the first test set based on the historical dataset includes:

[0014] Based on the historical dataset, select the kiln data within the corresponding temperature anomaly set time range of the kiln temperature zone, and use it as the historical dataset to be processed.

[0015] The historical dataset to be processed is preprocessed to obtain a preprocessed historical dataset; wherein, the preprocessing includes data cleaning.

[0016] The first training set and the first test set are obtained based on the preprocessed historical dataset.

[0017] Furthermore, the data cleaning process specifically includes:

[0018] When there are abnormally fluctuating historical kiln data in the historical dataset to be processed, the abnormally fluctuating historical kiln data is replaced with the median value of all the historical kiln data until there are no abnormally fluctuating temperature data in the kiln data; the abnormally fluctuating historical kiln data is the temperature data whose difference from the nearby historical kiln data exceeds a preset difference.

[0019] Furthermore, both the second training set and the second test set include historical kiln data for the kiln temperature zone within a preset prediction period; wherein, the preset prediction period includes a normal prediction period and a fault prediction period.

[0020] Furthermore, the initial temperature prediction model includes: LSTM model, RNN model, GRU model, CNN model, and GRN model.

[0021] Furthermore, the method also includes:

[0022] Based on the kiln data of the kiln temperature zone, several statistical moduli are obtained, and based on the statistical moduli, the temperature distribution characteristics of the kiln temperature zone are obtained.

[0023] Based on the temperature distribution characteristics, the temperature change trend of the kiln temperature zone is obtained.

[0024] Furthermore, the statistical modulus includes mean, variance, skewness, and kurtosis. In a second aspect, embodiments of the present invention also provide a kiln temperature prediction system, the system comprising:

[0025] The data acquisition module is used to acquire kiln data for any kiln temperature zone within a set time range in its corresponding temperature zone, and to form a historical dataset; the kiln data includes temperature data and power data;

[0026] The first screening module is used to obtain a first training set and a first test set based on the historical dataset, and to train several pre-selected initial temperature prediction models using the first training set to obtain several first temperature prediction models; it is also used to test several first temperature prediction models using the first test set, output the first test set output error, and select a second temperature prediction model from the first temperature prediction models based on the first test set output error and the preset range of the first test error.

[0027] The second screening module is used to train several second temperature prediction models using a pre-acquired second training set to obtain a third temperature prediction model; it is also used to test several third temperature prediction models according to a pre-acquired second test set, output the output error of the second test set, and take the third temperature prediction model corresponding to the minimum value of the output error of the second test set as the optimal temperature prediction model for the corresponding kiln temperature zone.

[0028] The temperature prediction module is used to input the kiln temperature data to be predicted of the kiln temperature zone collected during the preset prediction period into the optimal temperature prediction model to obtain the predicted temperature of the kiln temperature zone during the prediction period.

[0029] Thirdly, embodiments of the present invention also provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0030] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0031] This invention provides a method, system, device, and medium for predicting kiln temperature. The method includes: monitoring the temperature zones of the kiln to obtain kiln data and constructing a historical dataset, wherein the kiln data includes temperature data and power data; obtaining a first training set and a first test set based on the historical dataset; training and testing an initial temperature prediction model using the first training set and the first test set respectively to preliminarily screen the initial temperature prediction model and obtain a second temperature prediction model; and training and testing the initial temperature prediction model using the second training set and the first test set respectively to screen the second temperature prediction model and obtain the optimal temperature prediction model for the corresponding kiln temperature zone, thereby achieving accurate prediction of the temperature of each temperature zone of the kiln. This invention, by accurately predicting the future temperature of the kiln, can anticipate the temperature change trend of each temperature zone in advance, thereby reducing the impact of large temperature fluctuations on materials. Attached Figure Description

[0032] Figure 1 This is a schematic flowchart of a kiln temperature prediction method according to an embodiment of the present invention;

[0033] Figure 2 This is a time-series prediction schematic diagram of a kiln temperature prediction method according to an embodiment of the present invention;

[0034] Figure 3 This is a system block diagram of a kiln temperature prediction system according to an embodiment of the present invention;

[0035] Figure 4 This is an internal structural diagram of the computer device in an embodiment of the present invention. Detailed Implementation

[0036] To make the objectives, technical solutions, and beneficial effects of this application clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Obviously, the embodiments described below are only part of the embodiments of the present invention and are used to illustrate the present invention, but are not intended to limit the scope of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0037] In one embodiment, such as Figure 1 As shown, an embodiment of the present invention provides a method for predicting kiln temperature, the method comprising:

[0038] S11. Obtain kiln data for any kiln temperature zone within its corresponding temperature zone set time range, and construct a historical dataset; the kiln data includes temperature data and power data; the temperature zone set time range includes a normal temperature set time range and a temperature abnormal set time range. It should be noted that those skilled in the art can set the time length of the normal temperature set time range according to specific circumstances, and the embodiments of the present invention do not impose specific limitations.

[0039] In this embodiment, Postman is used to acquire the measurement point number data of the temperature zone. Temperature and power data are acquired at 1-minute intervals within both the normal and abnormal temperature setting time ranges. In other embodiments, other existing acquisition methods and intervals can also be used to acquire temperature and power data of the temperature zone, which will not be elaborated upon here.

[0040] An example of this embodiment is as follows:

[0041] For different temperature zones with different set temperatures, this embodiment selects temperature data and power data of the same temperature zone of different furnaces (such as the 4u temperature zone of each kiln) within the temperature abnormality set time range. In this embodiment, the temperature abnormality set time range is preferably set to 6 days before and after the temperature abnormality, as shown in Table 1.

[0042] Table 1. Furnace Area Value Collection Schedule

[0043]

[0044] It should be noted that the detection of kiln temperature data is not limited to temperature and power data. Other kiln data can be added as needed to make the prediction results more accurate.

[0045] S12. Obtain a first training set and a first test set based on the historical dataset, and use the first training set to train several pre-selected initial temperature prediction models to obtain several first temperature prediction models.

[0046] In this embodiment, the step of obtaining the first training set and the first test set based on the historical dataset includes:

[0047] Based on the historical dataset, kiln data within the corresponding temperature anomaly set time range of the kiln temperature zone are selected and used as the historical dataset to be processed. In this embodiment, the temperature anomaly set time range is preferably set to 6 days before and after the temperature anomaly. Those skilled in the art can set it according to the actual situation, which is not limited to the present invention.

[0048] The historical dataset to be processed is preprocessed to obtain a preprocessed historical dataset; wherein, the preprocessing includes data cleaning.

[0049] The first training set and the first test set are obtained based on the preprocessed historical dataset.

[0050] The data cleaning process specifically includes: when there are abnormally fluctuating historical kiln data in the historical dataset to be processed, the abnormally fluctuating historical kiln data is replaced with the median value of all the historical kiln data, until there are no abnormally fluctuating temperature data in the kiln data; the abnormally fluctuating historical kiln data is temperature data whose difference from the nearby historical kiln data exceeds a preset difference.

[0051] In this embodiment, the acquired historical data to be processed is cleaned before training the prediction model. For example, if data suddenly changes to 0, the value at the moment the data suddenly changes to 0 is replaced with the median of the entire sample population to obtain a cleaned historical dataset. This avoids error interference caused by sudden data changes. For example, when there is temperature data with a value of 0 in the kiln data, the temperature data with a value of 0 is replaced with the median value of all temperature data until there is no temperature data with a value of 0 in the kiln data. Similarly, when there is power data with a value of 0 in the kiln data, the power data with a value of 0 is replaced with the median value of all power data until there is no power data with a value of 0 in the kiln data.

[0052] S13. Using the first test set, test several of the first temperature prediction models respectively, output the first test set output error, and select the second temperature prediction model from the first temperature prediction models according to the first test set output error and the preset range of the first test error.

[0053] S14. Using the pre-acquired second training set, train several second temperature prediction models to obtain a third temperature prediction model.

[0054] S15. Test several of the third temperature prediction models according to the pre-acquired second test set, output the output error of the second test set, and take the third temperature prediction model corresponding to the minimum value of the output error of the second test set as the optimal temperature prediction model for the corresponding kiln temperature zone.

[0055] S16. Input the kiln temperature data to be predicted collected in the kiln temperature zone during the preset prediction period into the optimal temperature prediction model to obtain the predicted temperature of the kiln temperature zone during the prediction period; wherein, the kiln temperature data to be predicted includes the predicted temperature data and the predicted power data.

[0056] In this embodiment, the historical dataset includes a first training set and a first test set. The historical dataset generally contains data for 6 days before and after the time of temperature anomaly. The second training set and the second test set both include historical kiln data of the kiln temperature zone within a preset prediction period. The preset prediction period includes the normal period and the fault period before the time when the temperature needs to be predicted. The historical kiln data includes historical temperature data and historical power data.

[0057] This embodiment uses the above two training and testing processes to select the third temperature prediction model corresponding to the minimum output error of the second test set as the optimal temperature prediction model for the corresponding kiln temperature zone. It should be noted that, with a suitable dataset selected, a third or more model training and testing selection process can also be arranged.

[0058] In this embodiment, the prediction model includes: LSTM model, RNN model, GRU model, CNN model, and GRN model. Preferably, this embodiment can use an LSTM (long short term memory) based method to predict temperature data for future times based on historical data.

[0059] For example, a 12-hour advance warning for heating rods is needed to provide sufficient preparation time. Therefore, temperature prediction for the next 12 hours is performed. Considering that heating power data is also available, the heating power data for the current period and the temperature zone data are input together into the optimal temperature prediction model as features to predict the kiln temperature data for the next 12 hours. That is, the temperature (t) and power (p) from time period 1 to n are used to predict the temperature value from time n+1 to time k. Figure 2 As shown.

[0060] In actual modeling, the time step is first determined. For example, a 24-hour time step is used to predict the temperature data for the next 12 hours. Then, a time period is randomly selected in the temperature range where the temperature needs to be predicted for prediction and the prediction is verified.

[0061] In other embodiments of this application, more kiln temperature zones can be selected for prediction tests according to actual conditions. For example, based on fault repair records, data from 30 days prior to the repair time can be selected. Excluding the data used for prediction, the data used to train the initial temperature prediction model can be used.

[0062] Optionally, in this application, for different temperature zones of different kilns, corresponding trained models are set for prediction. Different optimal temperature prediction models are used for different temperature zones. When making temperature prediction for a certain temperature zone, the prediction model is retrained using the dataset before the prediction time and the parameters of the optimal temperature prediction model are adjusted.

[0063] When the heating equipment of the kiln is replaced, the model training and screening process needs to be repeated to select a suitable prediction model.

[0064] After obtaining the predicted temperature of the kiln temperature zone, this embodiment also includes:

[0065] When the predicted temperature of the kiln temperature zone exceeds the preset temperature threshold, the communication device address of the person in charge of the corresponding kiln is obtained.

[0066] A notification message is sent to the equipment address of the person in charge of the corresponding kiln.

[0067] If the predicted temperature of the kiln temperature zone exceeds the first preset value, the communication device address of the person in charge of the corresponding kiln will be called and a prompt voice message will be played as a prompt message.

[0068] If the predicted temperature of the kiln temperature zone exceeds the second preset value, a text message will be sent to the communication device address of the person in charge of the corresponding kiln as a reminder.

[0069] The first preset value is greater than the second preset value.

[0070] When the predicted temperature fluctuates significantly, a notification will be sent to the relevant personnel to take preventative measures against such fluctuations.

[0071] The relevant responsible persons can set different alarm levels and use different notification methods such as calls or text messages for different alarm levels in order to implement more targeted response measures.

[0072] In one embodiment, the kiln temperature prediction method provided in this embodiment further includes:

[0073] Based on the kiln data of the kiln temperature zone, several statistical moduli are obtained, and the temperature distribution characteristics of the kiln temperature zone are obtained based on the statistical moduli; wherein, the statistical moduli include mean, variance, skewness and kurtosis;

[0074] Based on the temperature distribution characteristics, the temperature change trend of the kiln temperature zone is obtained.

[0075] This embodiment uses multiple statistical moduli to analyze kiln data (temperature zone data, power data) to obtain the temperature distribution characteristics of the kiln temperature zone. These characteristics are then used to analyze the temperature of the kiln temperature zone, for example, to obtain the temperature change trend. When an abnormal temperature occurs in a kiln temperature zone, it indicates that the heating element needs to be replaced. This embodiment uses different statistical moduli to analyze and calculate the kiln data before and after the temperature zone abnormality, obtaining the corresponding temperature distribution characteristics. Based on these characteristics, the temperature change trend before and after the abnormality is obtained, and the temperature change trends of the same temperature zone in different kilns before and after the abnormality are compared and analyzed. This allows for the detection of the correctness of the heating element's operating condition or maintenance time based on the temperature change trend before and after the abnormality. For example, if a significant temperature change is detected based on the temperature change trend before and after the abnormality, it indicates a change in the stability of the heating element in the temperature zone. Those skilled in the art can also detect the quality and other characteristics of the heating element based on the temperature change trend before and after the abnormality. Let the kiln data X = [x1, x2, ..., x...]. n Let i represent the data of the i-th kiln, and n represent the number of elements contained in the kiln data. Then the statistical value of X can be expressed as:

[0076] Mean:

[0077] variance:

[0078] Skewness:

[0079] Kuroshi:

[0080] Here, the mean represents the central location of the data. If X approximately follows a Gaussian distribution, then the data are basically distributed around the mean. The variance represents the degree to which the data variable deviates from the mean. The variance is also the second central moment of the variable X.

[0081] Skewness characterizes the degree of asymmetry between the probability distribution density function curve and the mean, and measures the asymmetry of the probability distribution of a random variable.

[0082] Kurtosis characterizes the state of the probability distribution density function curve at the mean. If the kurtosis is less than 3, the distribution is flat; if the kurtosis is greater than 3, the distribution is steep.

[0083] For example, this embodiment can generate a histogram of temperature and power frequency distribution for the 4u temperature zone of furnace number 7 (7#4u) within a predetermined time. The analysis shows that the temperature is concentrated around 700℃, and the power data of the 4u temperature zone of furnace number 7 can be regarded as approximately following a Gaussian distribution. Therefore, this embodiment can perform discriminant analysis on temperature anomalies based on the characteristics of temperature and power data.

[0084] In another embodiment, this embodiment can also use the ARIMA prediction algorithm for prediction. The ARIMA prediction algorithm mainly focuses on the changes of the data at a future point in time, while ignoring the process data. Specifically, it focuses on the data node position 12 hours later, rather than how the data changes within 12 hours.

[0085] This invention provides a method for predicting kiln temperature, enabling the prediction of future kiln temperature, the advance knowledge of temperature change trends in each temperature zone, and assisting technicians in detecting kiln temperature anomalies in advance, making timely judgments and taking corresponding measures to reduce the impact of large temperature fluctuations on materials.

[0086] Based on the above-described kiln temperature prediction method, this invention also provides a kiln temperature prediction system, such as... Figure 3 As shown, the system includes:

[0087] Data acquisition module 1 is used to acquire kiln data for any kiln temperature zone within a set time range of its corresponding temperature zone, and to form a historical dataset; the kiln data includes temperature data and power data;

[0088] The first screening module 2 is used to obtain a first training set and a first test set based on the historical dataset, and to train several pre-selected initial temperature prediction models using the first training set to obtain several first temperature prediction models; it is also used to test several first temperature prediction models using the first test set, output the first test set output error, and select a second temperature prediction model from the first temperature prediction models based on the first test set output error and the preset range of the first test error.

[0089] The second screening module 3 is used to train several second temperature prediction models using a pre-acquired second training set to obtain a third temperature prediction model; it is also used to test several third temperature prediction models according to a pre-acquired second test set, output the output error of the second test set, and take the third temperature prediction model corresponding to the minimum value of the output error of the second test set as the optimal temperature prediction model for the corresponding kiln temperature zone.

[0090] Temperature prediction module 4 is used to input the kiln temperature data to be predicted of the kiln temperature zone collected during the preset prediction period into the optimal temperature prediction model to obtain the predicted temperature of the kiln temperature zone during the prediction period.

[0091] For specific limitations regarding a kiln temperature prediction system, please refer to the limitations of a kiln temperature prediction method described above, which will not be repeated here. Each module in the above system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0092] Figure 4 An internal structural diagram of a computer device is shown in one embodiment. This computer device may specifically be a terminal or a server. Figure 4 As shown, the computer device includes a processor, memory, network interface, display, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. The display screen can be an LCD screen or an e-ink display screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0093] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computing devices may include more or fewer components than shown in the diagram, or combine certain components, or have the same component arrangement.

[0094] In summary, this invention provides a method, system, device, and medium for predicting kiln temperature. The method includes: obtaining a historical dataset using monitored kiln data; acquiring a first training set and a first test set based on the historical dataset; training and testing a prediction model using the first training set and the first test set respectively; and initially selecting a prediction model based on the output error of the first test set; further training and testing the prediction model using a pre-acquired second training set and a second test set respectively; and selecting the optimal temperature prediction model based on the output error of the second test set; and obtaining the predicted temperature of the kiln temperature zone for the prediction period based on kiln data within a preset pre-prediction time period and the optimal temperature prediction model. This invention enables the prediction of future temperatures in kiln temperature zones, not only providing more accurate predictions and anticipating temperature change trends in each zone, but also reducing the computational load in the future temperature prediction process.

[0095] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. Each embodiment focuses on its differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0096] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the technical principles of this invention, and these improvements and substitutions should also be considered within the scope of protection of this application. Therefore, the scope of protection of this patent application should be determined by the scope of the claims.

Claims

1. A method of predicting the temperature of a kiln, characterized by, The method comprises: acquiring kiln data of any kiln temperature zone within its corresponding temperature zone setting time range and constituting a historical data set; the kiln data comprises temperature data and power data; acquiring a first training set and a first test set from the historical data set, training a plurality of initially selected temperature prediction models respectively using the first training set to obtain a plurality of first temperature prediction models; testing the plurality of first temperature prediction models respectively using the first test set, outputting a first test set output error, and selecting a second temperature prediction model from the first temperature prediction models according to the first test set output error and a first test error preset range; training a plurality of second temperature prediction models respectively using a pre-acquired second training set to obtain third temperature prediction models; testing a plurality of third temperature prediction models respectively according to a pre-acquired second test set, outputting a second test set output error, and taking the third temperature prediction model corresponding to the minimum second test set output error as the optimal temperature prediction model of the corresponding kiln temperature zone; inputting the predicted kiln data of the kiln temperature zone collected within a preset prediction early stage time period into the optimal temperature prediction model to obtain the predicted temperature of the kiln temperature zone in the prediction time period; wherein the step of acquiring a first training set and a first test set from the historical data set comprises: selecting kiln data of the kiln temperature zone within its corresponding temperature abnormality setting time range from the historical data set and taking it as a to-be-processed historical data set; preprocessing the to-be-processed historical data set to obtain a preprocessed historical data set; wherein the preprocessing comprises data cleaning processing; acquiring a first training set and a first test set from the preprocessed historical data set; the second training set and the second test set both comprise historical kiln data of the kiln temperature zone within a preset prediction early stage time period; wherein the preset prediction early stage time period comprises a prediction early stage normal time period and a prediction early stage fault time period.

2. The method of claim 1, wherein, The data cleaning processing specifically comprises: when there is abnormal jump historical kiln data in the to-be-processed historical data set, replacing the abnormal jump historical kiln data with the median value of all the historical kiln data until there is no abnormal jump temperature data in the kiln data; the abnormal jump historical kiln data is temperature data with a difference from nearby historical kiln data exceeding a preset difference value.

3. The method of claim 1, wherein: The initial temperature prediction model comprises: an LSTM model, an RNN model, a GRU model, a CNN model, and a GRN model.

4. The method of claim 1, wherein, The method further comprises: acquiring a plurality of statistical magnitudes from the kiln data of the kiln temperature zone and obtaining temperature distribution characteristics of the kiln temperature zone according to the statistical magnitudes; obtaining kiln temperature zone temperature change trends according to the temperature distribution characteristics.

5. A method of kiln temperature prediction according to claim 4, characterised in that: The statistical magnitudes comprise mean, variance, skewness, and kurtosis.

6. A furnace temperature prediction system characterized by comprising: The system comprises: The data acquisition module is configured to acquire kiln data of any kiln temperature zone within a corresponding temperature zone setting time range of the kiln temperature zone, and form a historical data set; the kiln data includes temperature data and power data; The first screening module is configured to acquire a first training set and a first test set according to the historical data set, train a plurality of initial temperature prediction models by using the first training set, and obtain a plurality of first temperature prediction models; the first screening module is further configured to test the plurality of first temperature prediction models by using the first test set, output a first test set output error, and screen a second temperature prediction model from the first temperature prediction models according to the first test set output error and a first test error preset range; The second screening module is configured to train the plurality of second temperature prediction models by using a second training set acquired in advance, and obtain third temperature prediction models; the second screening module is further configured to test the plurality of third temperature prediction models according to a second test set acquired in advance, output a second test set output error, and select a third temperature prediction model corresponding to a minimum value of the second test set output error as an optimal temperature prediction model of a corresponding kiln temperature zone; The temperature prediction module is configured to input predicted kiln data of the kiln temperature zone collected within a preset prediction early stage time period into the optimal temperature prediction model, and obtain a predicted temperature of the kiln temperature zone in a prediction time period.

7. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to realize the steps of the method in any one of claims 1 to 5.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the method in any one of claims 1 to 5.

Images