Temperature field prediction method and device of rotary kiln, electronic equipment and storage medium

Abstract

The invention provides a temperature field prediction method and device, electronic equipment and a storage medium, and relates to the technical field of rotary kilns.The method comprises the steps that target working condition parameters of a rotary kiln are obtained; the target working condition parameters comprise a target coal feeding amount and a target pulverized coal heat value; inputting the eigenvalue matrix corresponding to the target coal feed quantity and the target coal powder calorific value, the flow field identification matrix and the sign distance function matrix into a rotary kiln temperature field prediction model to obtain a temperature field of the rotary kiln; the flow field identification matrix is used for identifying a fluid domain and a solid domain in a geometric body of the rotary kiln, the sign distance function is used for representing the distance from a geometric space point of the rotary kiln to the geometric body of the rotary kiln, and the rotary kiln temperature field prediction model is obtained by training based on the sample coal feeding amount, the sample pulverized coal heat value, the flow field identification matrix and the sign distance function matrix. Through the trained rotary kiln temperature field prediction model, the temperature field of the rotary kiln can be accurately determined, and the prediction accuracy and efficiency of the temperature field of the rotary kiln are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of rotary kiln, and particularly relates to a rotary kiln temperature field prediction method and device, an electronic device and a storage medium. BACKGROUND

[0002] The cement rotary kiln is a kind of lime kiln, wherein the cement rotary kiln is the main equipment of the cement clinker dry and wet production line. With the development of industrial technology, real-time monitoring of the temperature field inside the cement rotary kiln is crucial for production and safety.

[0003] The traditional temperature field simulation technology of the cement rotary kiln needs to rely on a mechanism model, that is, partial differential equations such as combustion reaction and turbulent motion need to be established, and the calculation time is as long as several hours, which cannot meet the online monitoring demand, and the simulation precision of the temperature field is low.

[0004] Therefore, how to break through the real-time limitation of the traditional mechanism model and construct a dynamic temperature field / flow field online simulation system considering the calculation efficiency and precision is an urgent problem to be solved. SUMMARY

[0005] The present application provides a rotary kiln temperature field prediction method and device, an electronic device and a storage medium to solve the problem of low calculation efficiency and precision of the temperature field.

[0006] The present application provides a rotary kiln temperature field prediction method, comprising: obtaining target working condition parameters of a rotary kiln; the target working condition parameters include a target coal feeding amount and a target coal powder calorific value; inputting a feature value matrix corresponding to the target coal feeding amount and the target coal powder calorific value, a flow field identification matrix and a signed distance function matrix into a rotary kiln temperature field prediction model to obtain a temperature field of the rotary kiln output by the rotary kiln temperature field prediction model; wherein the flow field identification matrix is used to identify a fluid domain and a solid domain in the rotary kiln geometry, the signed distance function is used to represent the distance from the rotary kiln geometric space point to the rotary kiln geometry, and the rotary kiln temperature field prediction model is obtained by training based on sample coal feeding amount, sample coal powder calorific value, the flow field identification matrix and the signed distance function matrix.

[0007] According to the rotary kiln temperature field prediction method provided by the present application, the feature value matrix corresponding to the target coal feeding amount and the target coal powder calorific value, the flow field identification matrix and the signed distance function matrix are input into the rotary kiln temperature field prediction model to obtain the temperature field of the rotary kiln output by the rotary kiln temperature field prediction model, which comprises: input the target coal feeding amount and the characteristic value matrix corresponding to the target coal powder calorific value, the flow field identification matrix and the signed distance function matrix to the rotary kiln temperature field prediction model to obtain the fluid domain temperature and the solid domain temperature output by the rotary kiln temperature field prediction model; based on the solid domain temperature, determine the temperature of the solid domain at the next time; based on the fluid domain temperature and the temperature at the next time, determine the temperature field of the rotary kiln.

[0008] According to the rotary kiln temperature field prediction method provided by the application, the temperature of the solid domain at the next time is determined based on the solid domain temperature, which comprises: based on the solid domain temperature and the temperature at the current time, the temperature of the solid domain at the next time is determined by formula (1) ; (1) wherein T0 represents the temperature at the current time, represents a time constant, represents a predicted time interval, represents the solid domain temperature.

[0009] According to the rotary kiln temperature field prediction method provided by the application, the rotary kiln temperature field prediction model is obtained based on the following steps: construct a training data set; the training data set comprises the sample coal feeding amount and the sample coal powder calorific value; based on the sample coal feeding amount, the sample coal powder calorific value, the flow field identification matrix and the signed distance function matrix, train an initial rotary kiln temperature field prediction model to obtain the rotary kiln temperature field prediction model.

[0010] According to the rotary kiln temperature field prediction method provided by the application, the training data set is constructed, which comprises: determine the working condition parameters and the value range of the working condition parameters; the working condition parameters comprise the coal feeding amount and the coal powder calorific value; based on the working condition parameters, determine the number of working conditions; the number of working conditions is the same as the number of training samples in the training data set; based on the working condition parameters, the number of working conditions and the value range of the working condition parameters, determine the three-dimensional temperature field distribution information in the rotary kiln; based on the three-dimensional temperature field distribution information, determine the sample coal feeding amount and the sample coal powder calorific value.

[0011] According to the application, a temperature field prediction method of a rotary kiln is provided, and the three-dimensional temperature field distribution information in the rotary kiln is determined based on the working condition parameters, the working condition quantity and the value range of the working condition parameters, and the method comprises the following steps: The parameter value of the working condition parameter is determined based on the working condition parameter, the working condition quantity and the value range of the working condition parameter. The three-dimensional temperature field distribution information in the rotary kiln is determined based on the parameter value of the working condition parameter and the clinker reaction heat modeling technology.

[0012] According to the application, a temperature field prediction method of a rotary kiln is provided, and the three-dimensional temperature field distribution information in the rotary kiln is determined based on the working condition parameters, the working condition quantity and the value range of the working condition parameters, and the method comprises the following steps: The temperature of a plurality of points in the rotary kiln is extracted based on the three-dimensional temperature field distribution information. The temperatures of the plurality of points in the rotary kiln are respectively subjected to coordinate conversion to obtain converted temperatures. The converted temperatures are subjected to down-sampling processing to obtain processed temperatures. Sample parameters are determined based on the processed temperatures, and the sample parameters comprise the sample coal feeding quantity and the sample coal powder calorific value.

[0013] The application further provides a temperature field prediction device of a rotary kiln, which comprises: An acquisition module is configured to acquire target working condition parameters of a rotary kiln, and the target working condition parameters comprise a target coal feeding quantity and a target coal powder calorific value. A prediction module is configured to input a feature value matrix corresponding to the target coal feeding quantity and the target coal powder calorific value, a flow field identification matrix and a signed distance function matrix into a rotary kiln temperature field prediction model to obtain a temperature field of the rotary kiln output by the rotary kiln temperature field prediction model, wherein the flow field identification matrix is used to identify a fluid domain and a solid domain in a geometric body of the rotary kiln, the signed distance function is used to represent a distance from a geometric space point of the rotary kiln to the geometric body of the rotary kiln, and the rotary kiln temperature field prediction model is obtained by training based on a sample coal feeding quantity, a sample coal powder calorific value, the flow field identification matrix and the signed distance function matrix.

[0014] The application further provides an electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor implements the temperature field prediction method of the rotary kiln according to any one of the above-mentioned methods when executing the computer program.

[0015] The application further provides a non-transitory computer readable storage medium having a computer program stored thereon, and the computer program is executable by a processor to implement the temperature field prediction method of the rotary kiln according to any one of the above-mentioned methods.

[0016] The application also provides a computer program product comprising a computer program which, when executed by a processor, implements the temperature field prediction method of the rotary kiln as described above.

[0017] The application provides a rotary kiln temperature field prediction method, device, electronic equipment and storage medium. The target working condition parameters of the rotary kiln are obtained. The target working condition parameters include a target coal feeding amount and a target pulverized coal calorific value. The feature value matrix corresponding to the target coal feeding amount and the target pulverized coal calorific value, the flow field identification matrix and the symbolic distance function matrix are input into a rotary kiln temperature field prediction model to obtain the temperature field of the rotary kiln output by the rotary kiln temperature field prediction model. The flow field identification matrix is used to identify the fluid domain and the solid domain in the rotary kiln geometry. The symbolic distance function is used to represent the distance from the rotary kiln geometric space point to the rotary kiln geometry. The rotary kiln temperature field prediction model is obtained based on the sample coal feeding amount, the sample pulverized coal calorific value, the flow field identification matrix and the symbolic distance function matrix. The trained rotary kiln temperature field prediction model can accurately determine the temperature field of the rotary kiln and improve the prediction accuracy and efficiency of the temperature field of the rotary kiln. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0019] Figure 1 is one of the flowcharts of the rotary kiln temperature field prediction method provided by the application.

[0020] Figure 2 is a schematic diagram of the simulation system provided by the application.

[0021] Figure 3 is the second flowchart of the rotary kiln temperature field prediction method provided by the application.

[0022] Figure 4 is a schematic diagram of the three-dimensional temperature field distribution information provided by the application.

[0023] Figure 5 is a schematic diagram of the temperature distribution of the x=0 section provided by the application.

[0024] Figure 6(a) is a schematic diagram of the temperature distribution of the z=10m section provided by the application.

[0025] Figure 6(b) is a schematic view of temperature distribution of a cross-section at z = 20m according to the present application.

[0026] Figure 6(c) is a schematic view of temperature distribution of a cross-section at z = 40m according to the present application.

[0027] Figure 6(d) is a schematic view of temperature distribution of a cross-section at z = 70m according to the present application.

[0028] Figure 7 is a schematic view of temperature curve of a cross-section of hot material and fluid domain according to the present application.

[0029] Figure 8 is a schematic view of data front view of data after rotation of three-dimensional temperature field distribution information according to the present application.

[0030] Figure 9 is a schematic view of data after down-sampling difference according to the present application.

[0031] Figure 10(a) is a schematic view of coal input of one of results of coal dust calorific value characteristics according to the present application.

[0032] Figure 10(b) is a schematic view of coal input of two of results of coal dust calorific value characteristics according to the present application.

[0033] Figure 10(c) is a schematic view of coal input of three of results of coal dust calorific value characteristics according to the present application.

[0034] Figure 10(d) is a schematic view of coal input of four of results of coal dust calorific value characteristics according to the present application.

[0035] Figure 10(e) is a schematic view of coal input of five of results of coal dust calorific value characteristics according to the present application.

[0036] Figure 11 is a schematic view of flow of temperature field prediction method of a rotary kiln according to the present application.

[0037] Figure 12 is a schematic view of structure of temperature field prediction device of a rotary kiln according to the present application.

[0038] Figure 13 is a schematic view of structure of electronic device according to the present application. DETAILED DESCRIPTION

[0039] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in conjunction with the drawings in the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0040] First, in order to facilitate a clearer understanding of the embodiments of the present application, first, some related knowledge is introduced as follows.

[0041] The temperature field simulation technology of the traditional cement rotary kiln still has the following defects: 1. Poor adaptability of static empirical model. Among them, the regression model constructed based on experimental design data is difficult to capture the dynamic changes in the kiln (such as temperature field distortion caused by material composition fluctuation); 2. Low utilization rate of sensor data. Among them, the multi-source heterogeneous data such as infrared thermal imager and gas composition analyzer are not fused, resulting in insufficient flow field reconstruction accuracy.

[0042] Based on the above problems, the present application provides a rotary kiln temperature field prediction method, which can accurately determine the temperature field of the rotary kiln through the trained rotary kiln temperature field prediction model, and improve the prediction accuracy and efficiency of the rotary kiln temperature field.

[0043] The rotary kiln temperature field prediction method of the present application will be described below. Figures 1-11

[0044] Figure 1 is one of the flowcharts of the rotary kiln temperature field prediction method provided by the present application, as Figure 1 shown, the method comprises steps 101-102.

[0045] Step 101, obtaining target working condition parameters of the rotary kiln; the target working condition parameters include target coal feeding amount and target pulverized coal calorific value.

[0046] It should be noted that the rotary kiln temperature field prediction method provided by the present application can be applied to the scene of internal temperature field prediction of industrial intelligent equipment, for example, the temperature field prediction inside the cement rotary kiln. The execution subject of the method can be a rotary kiln temperature field prediction device, such as an electronic device, or a control module in the rotary kiln temperature field prediction device for executing the rotary kiln temperature field prediction method.

[0047] Specifically, Figure 2 is a schematic diagram of the simulation system provided by the present application, as Figure 2 ​As shown, under the support of the field data platform service, the system includes two parts of web front-end service and back-end service, wherein the front-end service is mainly used for visualizing the system page to present to the user operation, and the back-end service mainly includes three parts of data service, visualization service and model service; wherein the data service mainly includes functions such as collecting platform data provided by the platform service, pushing data to the web front-end (browser client), and checking the authority of the web front-end; the visualization service mainly includes functions such as providing an initialization model, obtaining model probe data (i.e. position points, coordinates of points, temperature data) and calculating a result cloud map; the model service mainly provides functions of calculating based on real-time data on site and giving a result of a three-dimensional dynamic temperature field inside the rotary kiln in real time.

[0048] Based on the above description, the target working condition parameters of the rotary kiln can be obtained in real time through the platform service; the target working condition parameters include a target coal feeding amount and a target coal powder calorific value, and can further include at least one of the following: a temperature and an air volume of secondary air, a feeding amount and a temperature of hot raw materials, a negative pressure of a kiln head cover, volatile matter and fixed carbon, and a speed and a temperature of an axial flow air.

[0049] Based on the target coal feeding amount and the target coal powder calorific value, other parameters of the existing working condition associated with the target coal feeding amount and the target coal powder calorific value can be screened through the original working condition parameter list.

[0050] In step 102, the characteristic value matrix corresponding to the target coal feeding amount and the target coal powder calorific value, the flow field identification matrix and the signed distance function matrix are input into the rotary kiln temperature field prediction model to obtain the temperature field of the rotary kiln output by the rotary kiln temperature field prediction model; wherein the flow field identification matrix is used to identify the fluid domain and the solid domain in the rotary kiln geometry, and the signed distance function is used to represent the distance from the rotary kiln geometry point to the rotary kiln geometry, and the rotary kiln temperature field prediction model is obtained based on the sample coal feeding amount, the sample coal powder calorific value, the flow field identification matrix and the signed distance function matrix.

[0051] Specifically, the rotary kiln temperature field prediction model is a rotary kiln internal reduced-order model, for example, the rotary kiln internal reduced-order model is a U-Net network, which is a classic convolutional neural network architecture. Due to its unique U-shaped structure and symmetric design, the U-Net network performs well in various computer vision tasks. The Net network structure is composed of an encoder (down-sampling path) and a decoder (up-sampling path), and the whole is a U-shaped symmetric structure. The training framework is built using pytorch to realize the model structure.

[0052] The rotary kiln temperature field prediction model is trained based on sample coal input, sample coal powder calorific value, flow field identification matrix and signed distance function matrix. After the rotary kiln temperature field prediction model is trained, the rotary kiln temperature field prediction model is deployed to the system in the field for application, and the system embedded with the rotary kiln temperature field prediction model is connected with the data and platform in the field, so that the data of the present house can be obtained in real time.

[0053] The characteristic value matrix corresponding to the target coal input and the target coal powder calorific value, the flow field identification matrix and the signed distance function matrix are input into the rotary kiln temperature field prediction model, so that the temperature field of the rotary kiln output by the rotary kiln temperature field prediction model can be obtained, that is, the steady-state temperature field. It should be noted that the characteristic value matrix corresponding to the target coal input and the target coal powder calorific value is the characteristic value matrix of the target coal input The characteristic value matrix of the target coal powder calorific value; the characteristic value matrix, the flow field identification matrix and the signed distance function matrix can all be 64 64 512; for the flow field identification matrix, the points in the fluid domain are marked as 2, the points in the solid domain are marked as 1, and the points not in the calculation domain are marked as 0; for the signed distance function matrix, the distance from the sampling point to the inner wall or outer wall of the rotary kiln is calculated, and the sampling points in the gas domain and the material layer domain inside the rotary kiln take positive values, and the sampling points in the refractory, outer cylinder domain and not in the calculation domain take negative values. Since the sampling is preceded by coordinate transformation, the axis of the rotary kiln is parallel to the z coordinate axis, and the signed distance function can be calculated according to the distance from the sampling point to the axis, which is expressed by formula (2) as follows: (2) Wherein, represents the signed distance function; 2.6 represents the radius of the inner wall of the refractory; x and y represent the coordinate values of the sampling point, which are the values after coordinate transformation.

[0054] The application provides a rotary kiln temperature field prediction method, which comprises the following steps: obtaining target working condition parameters of a rotary kiln; the target working condition parameters comprise a target coal feeding amount and a target coal powder calorific value; inputting a characteristic value matrix corresponding to the target coal feeding amount and the target coal powder calorific value, a flow field identification matrix and a signed distance function matrix into a rotary kiln temperature field prediction model to obtain a temperature field of the rotary kiln output by the rotary kiln temperature field prediction model; wherein the flow field identification matrix is used for identifying a fluid domain and a solid domain in a rotary kiln geometric body, the signed distance function is used for representing the distance from a rotary kiln geometric space point to the rotary kiln geometric body, and the rotary kiln temperature field prediction model is obtained by training based on sample coal feeding amounts, sample coal powder calorific values, the flow field identification matrix and the signed distance function matrix. The trained rotary kiln temperature field prediction model can accurately determine the temperature field of the rotary kiln and improve the prediction accuracy and efficiency of the temperature field of the rotary kiln.

[0055] Figure 3 Figure 2 is a flowchart of the rotary kiln temperature field prediction method provided by the application, as shown in the figure, the method comprises steps 301-304. Figure 3

[0056] Step 301, obtaining target working condition parameters of a rotary kiln; the target working condition parameters comprise a target coal feeding amount and a target coal powder calorific value.

[0057] Step 302, inputting a characteristic value matrix corresponding to the target coal feeding amount and the target coal powder calorific value, a flow field identification matrix and a signed distance function matrix into a rotary kiln temperature field prediction model to obtain fluid domain temperature and solid domain temperature output by the rotary kiln temperature field prediction model.

[0058] Specifically, inputting the characteristic value matrix corresponding to the target coal feeding amount and the target coal powder calorific value, the flow field identification matrix and the signed distance function matrix into the rotary kiln temperature field prediction model can obtain the fluid domain temperature and the solid domain temperature output by the rotary kiln temperature field prediction model; wherein the fluid domain temperature is the temperature output predicted by the rotary kiln temperature field prediction model because the air heat capacity is small and can quickly reach a steady state value.

[0059] Step 303, determining the temperature of the solid domain at the next moment based on the solid domain temperature.

[0060] Specifically, because the solid domain comprises raw materials, refractory bricks and outer cylinder steel plates and the like, the temperature of the solid domain at the next moment can be further determined based on the solid domain temperature.

[0061] Step 304, determining the temperature field of the rotary kiln based on the fluid domain temperature and the temperature at the next moment.

[0062] ​Specifically, based on the fluid domain temperature and the temperature at the next moment, the temperature field of the rotary kiln can be determined.

[0063] The temperature field prediction method of the rotary kiln provided by the application can quickly predict the internal temperature field of the cement rotary kiln by using the rotary kiln temperature field prediction model based on the real-time target coal input and target pulverized coal calorific value data obtained on site, and can give the prediction result of the three-dimensional dynamic temperature field within 1 minute, thereby improving the prediction accuracy and efficiency of the temperature field of the rotary kiln.

[0064] Optionally, the specific implementation of the step 303 comprises: Based on the solid domain temperature and the temperature at the current moment, the temperature of the solid domain at the next moment is determined by using formula (1) ; (1) Wherein, T0 represents the temperature at the current moment, represents a time constant, represents a predicted time interval, represents the solid domain temperature. It should be noted that, represents a time constant, which can be set to 3600 seconds (s).

[0065] Optionally, the rotary kiln temperature field prediction model is obtained based on the following steps: Construct a training data set; the training data set includes the sample coal input and the sample pulverized coal calorific value; based on the sample coal input, the sample pulverized coal calorific value, the flow field identification matrix and the signed distance function matrix, an initial rotary kiln temperature field prediction model is trained to obtain the rotary kiln temperature field prediction model.

[0066] Specifically, the construction of the training data set comprises determination of working conditions and parameter ranges, generation of sample parameters, determination of temperature field distribution of each working condition and temperature field data preprocessing. The training data set includes sample coal input and sample pulverized coal calorific value.

[0067] It should be noted that in the process of training the rotary kiln temperature field prediction model, the AdamW optimizer is used to update the model parameters. The AdamW optimizer is an improvement of the Adam optimizer. The AdamW optimizer can effectively regularize by decoupling weight decay and gradient update, thereby improving the generalization ability of the model. In the present application, in order to dynamically adjust the learning rate, the cosine annealing learning rate scheduler (Cosine AnnealingLR) is used, which periodically adjusts the learning rate in the form of a cosine function during the training process; in terms of loss function, mean square error loss (MSE Loss) is used.

[0068] Based on the sample coal input, the sample coal powder calorific value, the flow field identification matrix and the signed distance function matrix, the initial rotary kiln temperature field prediction model is trained using the sample coal input, the sample coal powder calorific value, the product of the sample coal input and the sample coal powder calorific value, and the flow field identification matrix and the signed distance function matrix, respectively. Three different models are obtained by training. Finally, the three models are evaluated by using the validation set data, and the final rotary kiln temperature field prediction model is obtained.

[0069] Optionally, the training data set is constructed, comprising: determining the working condition parameters and the value range of the working condition parameters; the working condition parameters include the coal input and the coal powder calorific value; based on the working condition parameters, the number of working conditions is determined; the number of working conditions is the same as the number of training samples in the training data set; based on the working condition parameters, the number of working conditions and the value range of the working condition parameters, the three-dimensional temperature field distribution information in the rotary kiln is determined; based on the three-dimensional temperature field distribution information, the sample coal input and the sample coal powder calorific value are determined.

[0070] Specifically, the working condition parameters and the value range of the working condition parameters are determined, wherein the working condition parameters include the coal input and the coal powder calorific value, and the value range of the working condition parameters can be [-2.5, 2.5]; the working condition parameters can also include, but are not limited to, at least one of the following: the temperature and air volume of the secondary air, the feeding amount and temperature of the hot raw material, the negative pressure of the kiln head cover, the volatile matter and fixed carbon, and the speed and temperature of the axial flow air; based on the working condition parameters, the number of working conditions can be determined, and a multi-working condition simulation model file can be generated, for example, the number of working conditions is 200, and the number of working conditions is the same as the number of training samples in the training data set, including the training set, the validation set and the test set. Based on the working condition parameters, the number of working conditions and the value range of the working condition parameters, the three-dimensional temperature field distribution information in the rotary kiln can be further determined; the three-dimensional temperature field distribution information represents the temperature distribution of each position point in the rotary kiln, as shown in Figure 4 Figure 4 is a schematic diagram of the three-dimensional temperature field distribution information provided by the application. Finally, based on the three-dimensional temperature field distribution information, the sample coal input and the sample coal powder calorific value can be further determined.

[0071] Optionally, the three-dimensional temperature field distribution information in the rotary kiln is determined based on the working condition parameters, the number of working conditions and the value range of the working condition parameters, comprising: determining the parameter values of the working condition parameters based on the working condition parameters, the number of working conditions and the value range of the working condition parameters; determining the three-dimensional temperature field distribution information in the rotary kiln based on the parameter values of the working condition parameters and the clinker reaction heat modeling technology.

[0072] ​Specifically, based on the working condition parameters, the number of working conditions and the value range of the working condition parameters, Latin hypercube sampling is used to sample in the parameter space represented by the value range of the working condition parameters, so as to determine the parameter values of each working condition. The specific generation process is as follows: assuming that m samples are extracted in an n-dimensional vector space, that is, m sample parameters are extracted in the parameter space of the working condition parameters corresponding to the n working conditions. The steps of Latin hypercube sampling are as follows: each dimension is divided into m non-overlapping intervals, so that each interval has the same probability (usually a uniform distribution is considered, so that the lengths of the intervals are the same), that is, the parameter space of the working condition parameters corresponding to each working condition is divided into m non-overlapping intervals, for example, [-2.5, 2.5] is divided into 10 intervals; then a point is randomly extracted in each interval of each dimension; then a point is randomly extracted from the points extracted in each interval of each dimension, and a vector is formed by combining the points extracted in all dimensions, to obtain the parameter values of the working condition parameters.

[0073] Based on the parameter values of the working condition parameters and the clinker reaction heat modeling technology, the fluid simulation calculation software is used to run simulation calculation to obtain and output the basic working condition temperature field data, and to obtain the three-dimensional temperature field distribution information in the rotary kiln. The calculation platform is the cpu resource of the dawn cluster, and table 1 is the configuration of the calculation resource.

[0074] Table 1. Configuration of calculation resource

[0075] The clinker reaction heat modeling technology is as follows: since the cement clinker sintering process is relatively complex in the rotary kiln, it is roughly divided into four regions from the kiln head to the kiln tail, i.e. cooling zone, sintering zone, transition zone (temperature rising and exothermic reaction zone) and decomposition zone, and the main reactions in each region are different, so the reaction heat is also different. Therefore, the length of each region, the temperature range of the reaction and the reaction heat source term (reaction heat coefficient) are set as shown in table 2. Through testing, the upper limit of the reaction temperature and the axial position limit are set for the decomposition zone, only the axial position limit is set for the transition zone, and the lower limit of the reaction temperature and the axial position limit are set for the sintering zone and the cooling zone.

[0076] Table 2. Temperature range and reaction heat source term setting of different regions

[0077] In the calculation process of multiple working conditions, the present invention simplifies the calculation model by adjusting the parameters and calculation settings, thereby solving the problems of long calculation time and poor convergence results. The main measures include: (1) changing the hot raw material fluid domain to a solid domain, and replacing the axial movement of the hot raw material with solid motion; (2) changing the steady-state calculation to a transient calculation, and setting the solid time step definition to automatic; (3) setting the ash mass fraction in the coal powder to 0, and synchronously adjusting the calorific value and the coal powder mass flow rate. After the above adjustments and tests, after 2000 steps of transient calculation, the temperature distribution of the calculation model basically reached a thermal equilibrium state, as shown in Figure 4 As shown in the figure, the iteration time for each working condition was shortened from 7 days to 2 to 3 days.

[0078] based on Figure 4 The three-dimensional temperature field distribution information shown can be intercepted in the x direction. Figure 5 This is a schematic diagram of the temperature distribution of the x=0 section provided by the present invention, such as Figure 5 As shown, at the x=0 section, the temperature range is [125-1790]. Figure 6 (a) is a schematic diagram of the temperature distribution of the cross section at z=10m provided by the present invention, Figure 6 (b) is a schematic diagram of the temperature distribution of the cross section at z=20m provided by the present invention, Figure 6 (c) is a schematic diagram of the temperature distribution of the cross section at z=40m provided by the present invention, and Figure 6 (d) is a schematic diagram of the temperature distribution of the cross section at z=70m provided by the present invention. Figure 7 This is a schematic diagram of the temperature curve of the hot raw material and fluid domain cross section provided by the present invention, such as Figure 7 As shown, the horizontal axis 0 represents the kiln head, 78 represents the kiln tail, line_bot represents the temperature curve of the inner wall of the refractory brick, line_mid represents the temperature curve in the middle of the material layer, line_jm is the temperature curve of the upper surface of the material layer, and line_top represents the gas temperature curve 50mm away from the upper surface of the material layer. Figure 7 It can be seen that the highest temperature is between 10m and 20m. The temperature of the hot raw material is relatively low at the kiln tail, then rapidly rises in the solid phase reaction zone, and remains above 1300°C in the transition zone and sintering zone. The highest temperature of the hot raw material is around 1500°C, and the highest temperature above the hot raw material is around 1650°C.

[0079] Through test calculations of a combustion model and a hot raw material movement model, the present invention determines a calculation form in which air and hot raw material respectively occupy different areas of the calculation domain. The fluid domain and the hot raw material domain complete heat transfer through coupled walls. The reaction heat in the hot raw material is set to different reaction heat source item values ​​according to the position of the hot raw material rotary kiln. This calculation method shortens the calculation time while ensuring calculation accuracy.

[0080] Optionally, the determining the sample coal input quantity and the sample coal powder calorific value based on the three-dimensional temperature field distribution information comprises: extracting temperatures of multiple points in the rotary kiln based on the three-dimensional temperature field distribution information; performing coordinate conversion on the temperatures of the multiple points in the rotary kiln respectively to obtain converted temperatures; performing down-sampling processing on each of the converted temperatures to obtain processed temperatures; determining sample parameters based on the processed temperatures; the sample parameters comprise the sample coal input quantity and the sample coal powder calorific value.

[0081] Specifically, the pre-processing process of the three-dimensional temperature field distribution information comprises temperature extraction, coordinate conversion, down-sampling difference, segmented temperature field, feature selection and saved data file, etc. Wherein, based on the three-dimensional temperature field distribution information, the temperatures of multiple points in the rotary kiln are extracted, that is, the structure point coordinates and temperature features are extracted from the three-dimensional temperature field distribution information stored in the original cgns format and saved as an npy file. Then, the temperatures of multiple points in the rotary kiln are converted respectively to obtain converted temperatures, that is, the data coordinates rotation axis is adjusted to the z axis by rotating, translating and other changes of the npy point coordinates, and at the same time, the clinker is at the bottom, as shown in Figure 8 , Figure 8 is a schematic diagram of the data front view after data rotation of the three-dimensional temperature field distribution information provided by the application.

[0082] The down-sampling processing is performed on each of the converted temperatures, that is, the linear interpolation processing is performed on each of the converted data to interpolate to 64 64 512 of a uniform cuboid, wherein the temperature of the outer point of the circle with a radius distance greater than 3.3 is set to 0, and finally the processed temperature is obtained, as shown in Figure 9 , Figure 9 is a schematic diagram of the data obtained by the down-sampling difference.

[0083] Based on the processed temperatures, the three-dimensional schematic diagram (as shown in Figure 9 ) corresponding to the processed temperatures is sliced along the x axis to obtain 64 two-dimensional sections of 64 512; based on the temperature distribution represented by each two-dimensional section, the coal input quantity and the coal powder calorific value which mainly affect the temperature distribution parameters are determined as the feature reference points, the specific values of the coal input quantity and the coal powder calorific value are determined, and the values of the coal input quantity and the coal powder calorific value are determined as the values of the sample parameters; the sample parameters comprise the sample coal input quantity and the sample coal powder calorific value. The parameter conditions of the existing working conditions are screened from the original working condition parameter list and saved as a csv file.

[0084] It should be noted that the feature distribution is obtained by linearly combining the feature reference points, that is, the sample parameters massflow_coal and qnet are combined to obtain the sample massflow_coal, the sample qnet, and the sample massflow_coal qnet of the sample coal powder, the feature value matrix based on the sample massflow_coal, the feature value matrix of the sample qnet, the feature value matrix of the sample massflow_coal qnet, the feature value matrix based on the sample massflow_coal, the feature value matrix of the sample qnet, the feature value matrix of the sample massflow_coal qnet, and the flow field identification matrix and the symbolic distance function matrix, the initial rotary kiln temperature field prediction model is trained, three different models are trained, and finally the three models are evaluated using the validation set data to obtain the final rotary kiln temperature field prediction model.

[0085] In the process of model training, 64 cross sections of 64 512 are obtained along the x direction, and the 64 cross sections are trained, and in the model inference, the 64 cross sections are spliced back to the 64 64 512 cuboid. At the same time, the training results are tested and evaluated, and if the accuracy meets the requirements, the model parameters are saved. Among them, the present application adopts the normalized root mean squared error (Normalized Root Mean Squared Error, NRMSE) evaluation index, and the NRMSE evaluation index is an index for evaluating the prediction accuracy of the model, which is a normalized version of the root mean squared error (Root Mean Squared Error, RMSE). By normalizing the RMSE to the range of the target variable, the error between different true values and predicted values can be compared, as shown in formula (3).

[0086] (3) Wherein, y represents the training result, that is, the prediction result.

[0087] The three models are evaluated using the validation set data, combined with images and evaluation indexes, and qualitative and quantitative analysis is performed to obtain the best prediction effect under the characteristic massflow_coal qnet.

[0088] The validation set data is as follows: (1) Reference feature: massflow_coal Out_NRMSE: 0.052, 0.085, 0.063, 0.045 , 0.110, 0.052 , 0.136, 0.133, 0.041 , average: 0.0797; In_NRMSE: 0.071, 0.057, 0.042, 0.075 , 0.050, 0.116 , 0.078, 0.080, 0.089 , average: 0.0731; (2) Reference feature: coal powder heat value (qnet) Out_NRMSE: 0.049 , 0.148, 0.141, 0.031, 0.142 , 0.124, 0.154, 0.068 , 0.020 , average: 0.097; In_NRMSE: 0.060 , 0.051, 0.038, 0.053, 0.049 , 0.067, 0.071, 0.048 , 0.046 , average: 0.0537; (3) Reference feature: coal input coal powder heat value (massflow_coal qnet) Out_NRMSE: 0.028, 0.186, 0.041 , 0.048 , 0.086, 0.084, 0.022, 0.021 , 0.120, average: 0.071; In_NRMSE: 0.044 , 0.0081, 0.072 , 0.102 , 0.055, 0.051, 0.050, 0.042 , 0.053, average: 0.053.

[0089] Wherein, Out_NRMSE represents the evaluation index of the prediction result of the cement rotary kiln pipe wall, and In_NRMSE represents the evaluation index of the prediction result of the cement rotary kiln internal fluid. For each reference feature, the same black italicized number in each column represents the evaluation index corresponding to the same sample (except other numbers).

[0090] Figure 10 (a) is one of the results of the coal input coal powder heat value feature provided by the application, as shown in Figure 10 (a), which represents the training result schematic diagram corresponding to the section 08 obtained along the x direction; Figure 10 (b) is one of the results of the coal input The second result diagram of the calorific value characteristics of pulverized coal is shown in FIG10(b), which shows the training result diagram corresponding to the section 16 obtained along the x direction; FIG10(c) shows the coal feed amount provided by the present invention. The third result diagram of the calorific value characteristics of pulverized coal is shown in FIG10 (c), which shows the training process diagram corresponding to the section 32 obtained along the x direction; FIG10 (d) is the coal feed amount provided by the present invention. The fourth result diagram of the calorific value characteristics of pulverized coal is shown in FIG10(d), which shows the training process diagram corresponding to the section 48 obtained along the x direction; FIG10(e) is the coal feed amount provided by the present invention. The fifth result diagram of the calorific value feature of pulverized coal is shown in Figure 10 (e), which shows the training process diagram corresponding to the section 56 obtained along the x direction. Among them, the feature (Feature) represents the amount of coal fed The distribution of the calorific value of pulverized coal. The label (Lablel) represents the cross-sectional temperature distribution obtained along the x-direction, that is, the true value; the prediction result (Pre) represents the predicted temperature distribution output by the model, that is, the predicted value; the difference (PPE({Label-Pred})) represents the error between the true value and the predicted value.

[0091] Figure 11 This is the third flow chart of the temperature field prediction method of the rotary kiln provided by the present invention, as shown in FIG. Figure 11As shown, the method includes a data set construction process, a model training process, and a prediction process; wherein the data set construction process includes: working condition and parameter range determination, sample parameter generation, basic working condition temperature field simulation, and temperature field data preprocessing, that is, determining the working condition parameters and the value range of the working condition parameters, wherein the working condition parameters include the coal input and the coal heat value, and the value range of the working condition parameters can be [-2.5, 2.5], and the working condition parameters can also include but are not limited to at least one of the following: the temperature and air volume of the secondary air, the feeding amount and temperature of the hot raw material, the negative pressure of the kiln head cover, the volatile matter and fixed carbon, and the speed and temperature of the axial flow air; based on the working condition parameters, the number of working conditions can be determined, and a multi-working condition simulation model file can be generated, for example, the number of working conditions is 200, and the number of working conditions is the same as the number of training samples in the training data set, including the training set, the validation set, and the test set. Based on the working condition parameters, the number of working conditions, and the value range of the working condition parameters, Latin hypercube sampling is used to sample in the parameter space represented by the value range of the working condition parameters, and the parameter values of each working condition can be determined. Based on the parameter values of the working condition parameters and the clinker reaction heat modeling technology, the fluid simulation calculation software is used to run simulation calculation to obtain and output the basic working condition temperature field data, and the three-dimensional temperature field distribution information in the rotary kiln is obtained (basic working condition temperature field simulation). The preprocessing process of the three-dimensional temperature field distribution information includes temperature extraction, coordinate conversion, down-sampling interpolation, temperature field segmentation, feature selection, and data file saving. Based on the three-dimensional temperature field distribution information, the temperatures of multiple points in the rotary kiln are extracted, that is, the structure point coordinates and temperature features are extracted from the original three-dimensional temperature field distribution information stored in the cgns format and saved as an npy file. The temperatures of multiple points in the rotary kiln are converted respectively, and the converted temperatures are obtained, that is, the npy point coordinates are rotated, translated, and other changes to adjust the data coordinate rotation axis to the z-axis, and at the same time, the clinker is at the bottom. Each converted temperature is down-sampled, that is, each converted data is linearly interpolated to a uniform cuboid of 64 64 512, wherein the temperature of the outer point with a radius distance greater than 3.3 is set to 0, and finally the processed temperature is obtained. Based on the processed temperature, the three-dimensional diagram corresponding to the processed temperature (as shown in Figure 9 ) is sliced along the x-axis, and 64 two-dimensional sections of 64 512 are obtained; based on the temperature distribution represented by each two-dimensional section, the coal input and the coal heat value, which mainly affect the temperature distribution parameters, are taken as the feature reference points to determine the specific values of the coal input and the coal heat value, and the values of the coal input and the coal heat value are determined as the values of the sample parameters; the sample parameters include the sample coal input and the sample coal heat value. The parameter conditions of the existing working conditions are screened from the original working condition parameter list and saved as a csv file.

[0092] The model training process includes model construction, network parameter setting, optimizer selection, and training and evaluation, that is, an initial rotary kiln temperature field prediction model is constructed, model parameters are initialized, an AdamW optimizer is selected to update model parameters, a feature distribution is obtained by linearly combining feature reference points, that is, sample parameters, mass flow of coal (massflow_coal) and coal powder heat value (qnet) are combined to obtain sample mass flow of coal, sample coal powder heat value, and sample mass flow of coal * sample coal powder heat value sample coal powder heat value (massflow_coal qnet), based on the feature value matrix of the sample mass flow of coal, the feature value matrix of the sample coal powder heat value, the feature value matrix of the sample mass flow of coal * sample coal powder heat value, and the flow field identification matrix and the signed distance function matrix, the initial rotary kiln temperature field prediction model is trained, three different models are obtained by training, and finally the three models are evaluated using the validation set data to obtain the final rotary kiln temperature field prediction model. The normalized root mean squared error (NRMSE) evaluation index is used to test and evaluate the training result, and if the accuracy meets the requirements, the model parameters are saved. It should be noted that the feature value matrix, the flow field identification matrix, and the signed distance function matrix can be 64 64 512; for the flow field identification matrix, the points in the fluid domain are marked as 2, the points in the solid domain are marked as 1, and the points not in the calculation domain are marked as 0; for the signed distance function matrix, the distance from the sampling point to the inner wall or outer wall of the rotary kiln is calculated, and the sampling points in the gas domain and the material layer domain inside the rotary kiln take positive values, and the sampling points in the refractory, outer cylinder domain and not in the calculation domain take negative values. Since the rotary kiln is subjected to coordinate transformation before sampling, the axis of the rotary kiln is parallel to the z coordinate axis, and the signed distance function can be calculated according to the distance from the sampling point to the axis, as shown in the above formula (2).

[0093] The prediction process includes obtaining field operating parameter, running model and temperature field fusion optimization. That is, the target operating parameter of the rotary kiln is obtained; the target operating parameter includes target mass flow of coal and target coal powder heat value; the feature value matrix, the flow field identification matrix and the signed distance function matrix corresponding to the target mass flow of coal and the target coal powder heat value are input into the trained rotary kiln temperature field prediction model to obtain the fluid domain temperature and the solid domain temperature output by the rotary kiln temperature field prediction model; based on the solid domain temperature and the temperature at the current time, the temperature of the solid domain at the next time is determined by formula (1); based on the fluid domain temperature and the temperature at the next time, the temperature field of the rotary kiln is determined.

[0094] The scheme provided by the present application realizes real-time visual presentation of the three-dimensional temperature field in the rotary kiln operation process, supports the operation personnel in analyzing the flame form characteristics based on the visual interface, and further evaluates the working state of the combustion system. Moreover, the rotary kiln temperature field prediction model can be established, the deep learning method is used to calculate the surface temperature distribution in the kiln body, the thermal imaging data generated by the model is highly consistent with the measured values collected by the infrared scanning device of the rotary kiln cylinder in the temperature gradient distribution characteristics, and the reliability of the model and the engineering applicability of the monitoring scheme are fully verified.

[0095] The rotary kiln temperature field prediction device provided by the present application is described below, and the rotary kiln temperature field prediction device described below can be correspondingly referred to the rotary kiln temperature field prediction method described above.

[0096] Figure 12 The structure diagram of the rotary kiln temperature field prediction device provided by the present application is shown in FIG. 1. Figure 12 As shown in FIG. 1, the rotary kiln temperature field prediction device 1200 comprises an acquisition module 1201 and a prediction module 1202, wherein, The acquisition module 1201 is configured to acquire target working condition parameters of the rotary kiln; the target working condition parameters comprise a target coal feeding amount and a target pulverized coal calorific value. The prediction module 1202 is configured to input a feature value matrix corresponding to the target coal feeding amount and the target pulverized coal calorific value, a flow field identification matrix and a signed distance function matrix into a rotary kiln temperature field prediction model, to obtain a temperature field of the rotary kiln output by the rotary kiln temperature field prediction model; wherein the flow field identification matrix is used to identify fluid domains and solid domains in the rotary kiln geometry, the signed distance function is used to represent the distance from the rotary kiln geometric space point to the rotary kiln geometry, and the rotary kiln temperature field prediction model is obtained by training based on sample coal feeding amounts, sample pulverized coal calorific values, the flow field identification matrix and the signed distance function matrix.

[0097] The application provides a rotary kiln temperature field prediction device, which comprises the following steps: obtaining target working condition parameters of a rotary kiln; the target working condition parameters comprise a target coal feeding amount and a target pulverized coal calorific value; inputting a characteristic value matrix corresponding to the target coal feeding amount and the target pulverized coal calorific value, a flow field identification matrix and a signed distance function matrix into a rotary kiln temperature field prediction model to obtain a temperature field of the rotary kiln output by the rotary kiln temperature field prediction model; wherein the flow field identification matrix is used for identifying a fluid domain and a solid domain in a rotary kiln geometric body, the signed distance function is used for representing the distance from a rotary kiln geometric space point to the rotary kiln geometric body, and the rotary kiln temperature field prediction model is obtained based on sample coal feeding amounts, sample pulverized coal calorific values, the flow field identification matrix and the signed distance function matrix. Through the trained rotary kiln temperature field prediction model, the temperature field of the rotary kiln can be accurately determined, and the prediction accuracy and efficiency of the temperature field of the rotary kiln are improved.

[0098] Optionally, the prediction module 1202 is specifically configured to: input the characteristic value matrix corresponding to the target coal feeding amount and the target pulverized coal calorific value, the flow field identification matrix and the signed distance function matrix into the rotary kiln temperature field prediction model to obtain fluid domain temperature and solid domain temperature output by the rotary kiln temperature field prediction model; determine the temperature of the solid domain at the next moment based on the solid domain temperature; determine the temperature field of the rotary kiln based on the fluid domain temperature and the temperature at the next moment.

[0099] Optionally, the prediction module 1202 is further configured to: the determination of the temperature of the solid domain at the next moment based on the solid domain temperature comprises: determining the temperature of the solid domain at the next moment based on the solid domain temperature and the temperature at the current moment by using formula (1) ; (1) wherein T0 represents the temperature at the current moment, represents a time constant, represents a predicted time interval, and T represents the solid domain temperature.

[0100] Optionally, the rotary kiln temperature field prediction device 1200 further comprises: a construction module configured to construct a training data set; the training data set comprises the sample coal feeding amounts and the sample pulverized coal calorific values; The training module is configured to train an initial rotary kiln temperature field prediction model based on the sample coal feeding amount, the sample coal powder calorific value, the flow field identification matrix and the signed distance function matrix, and obtain the rotary kiln temperature field prediction model.

[0101] Optionally, the construction module is specifically configured to: determine a working condition parameter and a value range of the working condition parameter; the working condition parameter includes a coal feeding amount and a coal powder calorific value; determine a working condition number based on the working condition parameter; the working condition number is the same as a number of training samples in the training data set; determine three-dimensional temperature field distribution information in the rotary kiln based on the working condition parameter, the working condition number and the value range of the working condition parameter; determine the sample coal feeding amount and the sample coal powder calorific value based on the three-dimensional temperature field distribution information.

[0102] Optionally, the construction module is further configured to: determine a parameter value of the working condition parameter based on the working condition parameter, the working condition number and the value range of the working condition parameter; determine the three-dimensional temperature field distribution information in the rotary kiln based on the parameter value of the working condition parameter and a clinker reaction heat modeling technology.

[0103] Optionally, the construction module is further configured to: extract temperatures of multiple points in the rotary kiln based on the three-dimensional temperature field distribution information; perform coordinate conversion on the temperatures of the multiple points in the rotary kiln respectively to obtain converted temperatures; perform down-sampling processing on the converted temperatures to obtain processed temperatures; determine sample parameters based on the processed temperatures; the sample parameters include the sample coal feeding amount and the sample coal powder calorific value.

[0104] Figure 13 is a schematic diagram of an entity structure of an electronic device provided by the application, as shown in Figure 13As shown, the electronic device 1300 can include a processor 1310, a communications interface 1320, a memory 1330, and a communications bus 1340, wherein the processor 1310, the communications interface 1320, and the memory 1330 complete mutual communication through the communications bus 1340. The processor 1310 can invoke a logical instruction in the memory 1330 to execute a temperature field prediction method of a rotary kiln, the method comprising: obtaining target operating condition parameters of the rotary kiln; the target operating condition parameters include a target coal input and a target pulverized coal heat value; inputting a feature value matrix corresponding to the target coal input and the target pulverized coal heat value, a flow field identification matrix, and a signed distance function matrix into a rotary kiln temperature field prediction model to obtain a temperature field of the rotary kiln output by the rotary kiln temperature field prediction model; wherein the flow field identification matrix is used to identify a fluid domain and a solid domain in the rotary kiln geometry, the signed distance function is used to represent the distance from the rotary kiln geometry space point to the rotary kiln geometry, and the rotary kiln temperature field prediction model is obtained based on sample coal input, sample pulverized coal heat value, the sample coal input, and the sample pulverized coal heat value.

[0105] In addition, the logical instructions in the memory 1330 described above can be implemented in the form of a software function unit and sold or used as an independent product, which can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various program code storage media.

[0106] In another aspect, the present application also provides a computer program product comprising a computer program, which can be stored on a non-transitory computer readable storage medium, and the computer program can be executed by a processor to enable a computer to perform the rotary kiln temperature field prediction method provided by the above-mentioned methods, which comprises: obtaining target working condition parameters of a rotary kiln; the target working condition parameters comprise a target coal feed rate and a target coal calorific value; inputting a feature value matrix corresponding to the target coal feed rate and the target coal calorific value, a flow field identification matrix and a signed distance function matrix into a rotary kiln temperature field prediction model to obtain a temperature field of the rotary kiln output by the rotary kiln temperature field prediction model; wherein the flow field identification matrix is used to identify fluid domains and solid domains in the rotary kiln geometry, the signed distance function is used to represent the distance from a geometric space point of the rotary kiln to the rotary kiln geometry, and the rotary kiln temperature field prediction model is obtained based on sample coal feed rates, sample coal calorific values, the sample coal feed rates and the sample coal calorific values.

[0107] In another aspect, the present application also provides a non-transitory computer readable storage medium having a computer program stored thereon, and the computer program can be executed by a processor to implement the rotary kiln temperature field prediction method provided by the above-mentioned methods, which comprises: obtaining target working condition parameters of a rotary kiln; the target working condition parameters comprise a target coal feed rate and a target coal calorific value; inputting a feature value matrix corresponding to the target coal feed rate and the target coal calorific value, a flow field identification matrix and a signed distance function matrix into a rotary kiln temperature field prediction model to obtain a temperature field of the rotary kiln output by the rotary kiln temperature field prediction model; wherein the flow field identification matrix is used to identify fluid domains and solid domains in the rotary kiln geometry, the signed distance function is used to represent the distance from a geometric space point of the rotary kiln to the rotary kiln geometry, and the rotary kiln temperature field prediction model is obtained based on sample coal feed rates, sample coal calorific values, the sample coal feed rates and the sample coal calorific values.

[0108] The device embodiments described above are merely illustrative, wherein the units illustrated as separate components can or can not be physically separated, and the components illustrated as units can or can not be physical units, i.e., they can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0109] Those skilled in the art can clearly understand the technical solutions of the various embodiments from the above description of the embodiments, and the various embodiments can be implemented by means of software with the necessary general hardware platforms, and of course, can also be implemented by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part of the prior art that makes a contribution, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0110] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for some technical features therein; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for predicting the temperature field of a rotary kiln, characterized in that: include: Obtain target operating parameters of the rotary kiln; The target operating parameters include target coal feed rate and target calorific value of pulverized coal; The eigenvalue matrix, flow field identification matrix and signed distance function matrix corresponding to the target coal feed rate and the target pulverized coal calorific value are input into the rotary kiln temperature field prediction model to obtain the temperature field of the rotary kiln output by the rotary kiln temperature field prediction model; wherein, the flow field identification matrix is ​​used to identify the fluid domain and solid domain within the rotary kiln geometric body, the signed distance function is used to characterize the distance from the rotary kiln geometric space point to the rotary kiln geometric body, and the rotary kiln temperature field prediction model is trained based on the sample coal feed rate, the sample pulverized coal calorific value, the flow field identification matrix and the signed distance function matrix.

2. The temperature field prediction method of a rotary kiln according to claim 1, characterized in that: The step of inputting the eigenvalue matrix, flow field identification matrix, and signed distance function matrix corresponding to the target coal feed rate and the target calorific value of the pulverized coal into a rotary kiln temperature field prediction model to obtain the temperature field of the rotary kiln output by the rotary kiln temperature field prediction model comprises: Inputting the eigenvalue matrix, the flow field identification matrix, and the signed distance function matrix corresponding to the target coal feed rate and the target calorific value of the pulverized coal into a rotary kiln temperature field prediction model to obtain the fluid domain temperature and the solid domain temperature output by the rotary kiln temperature field prediction model; Determining the temperature of the solid domain at a next moment based on the solid domain temperature; The temperature field of the rotary kiln is determined based on the fluid domain temperature and the temperature at the next moment.

3. The temperature field prediction method of a rotary kiln according to claim 2, characterized in that: The determining the temperature of the solid domain at a next moment based on the solid domain temperature includes: Based on the solid domain temperature and the current temperature, the temperature of the solid domain at the next moment is determined using formula (1): ; （1） Among them, T0 represents the temperature at the current moment, represents the time constant, represents the prediction time interval, represents the solid domain temperature.

4. The method for predicting the temperature field of a rotary kiln according to any one of claims 1 to 3, characterized in that: The rotary kiln temperature field prediction model is trained based on the following steps: Constructing a training data set; the training data set includes the sample coal feed amount and the sample pulverized coal calorific value; Based on the sample coal feed amount, the sample pulverized coal calorific value, the flow field identification matrix and the signed distance function matrix, the initial rotary kiln temperature field prediction model is trained to obtain the rotary kiln temperature field prediction model.

5. The temperature field prediction method of a rotary kiln according to claim 4, characterized in that: The constructing of the training data set includes: Determine the operating parameters and their value ranges; the operating parameters include the amount of coal fed and the calorific value of pulverized coal; Determining the number of operating conditions based on the operating condition parameters; the number of operating conditions is the same as the number of training samples in the training data set; Determining three-dimensional temperature field distribution information in the rotary kiln based on the operating condition parameter, the number of operating conditions, and the value range of the operating condition parameter; Based on the three-dimensional temperature field distribution information, the sample coal feed amount and the sample pulverized coal calorific value are determined.

6. The method for predicting the temperature field of a rotary kiln according to claim 5, wherein: The determining of the three-dimensional temperature field distribution information in the rotary kiln based on the operating condition parameter, the number of operating conditions, and the value range of the operating condition parameter includes: Determining a parameter value of the operating condition parameter based on the operating condition parameter, the number of operating conditions, and a value range of the operating condition parameter; Based on the parameter values ​​of the operating parameters and the clinker reaction thermal modeling technology, the three-dimensional temperature field distribution information in the rotary kiln is determined.

7. The method for predicting the temperature field of a rotary kiln according to claim 5, wherein: The determining of the sample coal feed amount and the sample pulverized coal calorific value based on the three-dimensional temperature field distribution information includes: extracting the temperatures of multiple points in the rotary kiln based on the three-dimensional temperature field distribution information; performing coordinate conversion on the temperatures of multiple points in the rotary kiln to obtain converted temperatures; downsampling each of the converted temperatures to obtain a processed temperature; Based on the processed temperature, sample parameters are determined; the sample parameters include the sample coal feed amount and the sample pulverized coal calorific value.

8. A temperature field prediction device for a rotary kiln, characterized in that: include: An acquisition module is used to obtain target operating parameters of the rotary kiln; The target operating parameters include target coal feed rate and target calorific value of pulverized coal; A prediction module is used to input the eigenvalue matrix, flow field identification matrix and signed distance function matrix corresponding to the target coal feed rate and the target coal powder calorific value into the rotary kiln temperature field prediction model to obtain the temperature field of the rotary kiln output by the rotary kiln temperature field prediction model; wherein, the flow field identification matrix is ​​used to identify the fluid domain and solid domain within the rotary kiln geometric body, the signed distance function is used to characterize the distance from the rotary kiln geometric space point to the rotary kiln geometric body, and the rotary kiln temperature field prediction model is trained based on the sample coal feed rate, the sample coal powder calorific value, the flow field identification matrix and the signed distance function matrix.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the temperature field prediction method of the rotary kiln according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the temperature field prediction method of the rotary kiln according to any one of claims 1 to 7 is implemented.

Images