Clean coal ash content control method, device and computer equipment

Abstract

The present application relates to a clean coal ash content control method, device, and computer equipment. The method comprises: inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data for a first preset time period, and the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data for the next moment; updating the clean coal ash content prediction data for the next moment to the clean coal ash content prediction data for the current moment, updating the flotation agent dosage stacking data for each flotation agent dosage scheme at the next moment to the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment, repeating the above clean coal ash content prediction steps until a clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme is obtained; and selecting the optimal flotation agent dosage scheme based on multiple clean coal ash content prediction trajectories to perform clean coal ash content control. This method can accurately control the clean coal ash content.

Description

Technical Field

[0001] The present application relates to the technical field of coal slime flotation, and in particular to a method, device and computer equipment for controlling ash content in clean coal. Background Art

[0002] Coal slime flotation is a coal preparation method that uses differences in the surface properties of ore particles to separate coal particles from gangue. During the process, flotation agents are added to the flotation machine. Adjusting the dosage of flotation agents controls the ash content of the clean coal, which refers to the gangue content in the clean coal.

[0003] Traditionally, the ash content of clean coal during the coal slime flotation process has been controlled through manual dosing. However, this manual dosing method suffers from lags and inaccuracies, making it impossible to precisely control the ash content of clean coal. Summary of the Invention

[0004] Based on this, it is necessary to provide a clean coal ash control method, device and computer equipment that can accurately control the clean coal ash content in response to the above technical problems.

[0005] In a first aspect, the present application provides a method for controlling ash content in clean coal, comprising:

[0006] Inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data for the next moment;

[0007] Updating the clean coal ash content prediction data at the next moment to the clean coal ash content prediction data at the current moment, updating the flotation agent dosage stacking data of each flotation agent dosage scheme at the next moment to the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment, returning to the step of inputting the clean coal ash content prediction data at the current moment, the historical flotation slime input data of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into the pre-built transfer model, and continuing to predict the clean coal ash content until a clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme is obtained;

[0008] According to the clean coal ash content prediction trajectories corresponding to multiple flotation agent dosage schemes, the optimal flotation agent dosage scheme is selected to control the clean coal ash content.

[0009] In one embodiment, before inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stack data for each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data at the next moment, the method further includes:

[0010] Determine the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment according to the flotation agent dosage stacking time and the clean coal ash content prediction step at the current moment;

[0011] According to the stacking time of the floating coal slime data and the clean coal discharging time, the historical floating coal slime data for the first preset time period is determined.

[0012] In one embodiment, the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stack data for each flotation agent dosage scheme at the current moment are input into a pre-built transfer model to obtain the clean coal ash content prediction data for the next moment, including:

[0013] The clean coal ash content prediction data at the current moment is used as the observation quantity prediction value at the current moment, the historical flotation slime data of the first preset time period is used as the external variable of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment is used as the action quantity stacking data of each flotation agent dosage scheme at the current moment;

[0014] The predicted value of the observation quantity at the current moment, the external variables of the first preset time period, and the stacked data of the action quantity of each flotation agent dosage scheme at the current moment are input into the pre-built transfer model to obtain the predicted value of the observation quantity at the next moment.

[0015] In one embodiment, the method further includes:

[0016] When the current moment is the first moment, the actual data of clean coal ash content at the current moment is obtained;

[0017] The actual clean coal ash content data at the current moment, the historical flotation slime data of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment are input into the pre-built transfer model to obtain the clean coal ash content prediction data at the next moment.

[0018] In one embodiment, before inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stack data for each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data at the next moment, the method further includes a transfer model construction step, including:

[0019] Obtaining clean coal ash content prediction data at the current training moment, and obtaining historical flotation slime data and flotation agent dosage stacked data that match the current training moment; inputting the clean coal ash content prediction data at the current training moment, the historical flotation slime data and flotation agent dosage stacked data that match the current training moment into the transfer model to be trained, and obtaining clean coal ash content prediction data for the next training moment;

[0020] The clean coal ash content prediction data for the next training moment is updated to the clean coal ash content prediction data for the current training moment, and the process returns to the step of obtaining the historical floating coal slime data and flotation agent dosage stacking data that match the current training moment, and continues to predict the clean coal ash content until a prediction trajectory of a preset length is obtained;

[0021] The transfer model to be trained is updated according to the predicted trajectory of the preset length until the preset conditions are met, thereby obtaining the pre-built transfer model.

[0022] In one embodiment, the method further includes:

[0023] When the current training moment is the initial training moment, obtaining actual clean coal ash content data at the current training moment;

[0024] The actual data of clean coal ash content at the current training moment, the historical flotation slime data and flotation agent dosage stacked data matching the current training moment are input into the transfer model to be trained to obtain the predicted data of clean coal ash content at the next training moment.

[0025] In one embodiment, updating the transfer model to be trained according to the predicted trajectory of the preset length includes:

[0026] Determine the error between the predicted trajectory of a preset length and the true trajectory through the discriminant network;

[0027] The transfer model to be trained is updated based on the error between the predicted trajectory of a preset length and the true trajectory.

[0028] In one embodiment, updating the transfer model to be trained according to the predicted trajectory of a preset length until a preset condition is met to obtain a pre-built transfer model includes:

[0029] The transfer model to be trained is updated according to the predicted trajectory of the preset length until the training conditions are met, thereby obtaining the trained transfer model;

[0030] The trained transfer model is subjected to model testing. When the model test passes, the pre-built transfer model is obtained; the model test includes deduction test, response curve test and data distribution test.

[0031] In a second aspect, the present application further provides a clean coal ash content control device, comprising:

[0032] An ash content prediction module is used to input the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data at the next moment;

[0033] a cyclic prediction module, configured to update the clean coal ash content prediction data at the next moment to the clean coal ash content prediction data at the current moment, update the flotation agent dosage stacking data of each flotation agent dosage scheme at the next moment to the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment, return to the step of inputting the clean coal ash content prediction data at the current moment, the historical flotation slime input data for the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into a pre-built transfer model, and continue to predict the clean coal ash content until a clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme is obtained;

[0034] The scheme selection module is used to select the optimal flotation agent dosage scheme according to the clean coal ash content prediction trajectory corresponding to multiple flotation agent dosage schemes to control the clean coal ash content.

[0035] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0036] Inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data for the next moment;

[0037] Updating the clean coal ash content prediction data at the next moment to the clean coal ash content prediction data at the current moment, updating the flotation agent dosage stacking data of each flotation agent dosage scheme at the next moment to the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment, returning to the step of inputting the clean coal ash content prediction data at the current moment, the historical flotation slime input data of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into the pre-built transfer model, and continuing to predict the clean coal ash content until a clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme is obtained;

[0038] According to the clean coal ash content prediction trajectories corresponding to multiple flotation agent dosage schemes, the optimal flotation agent dosage scheme is selected to control the clean coal ash content.

[0039] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the following steps:

[0040] Inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data for the next moment;

[0041] Updating the clean coal ash content prediction data at the next moment to the clean coal ash content prediction data at the current moment, updating the flotation agent dosage stacking data of each flotation agent dosage scheme at the next moment to the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment, returning to the step of inputting the clean coal ash content prediction data at the current moment, the historical flotation slime input data of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into the pre-built transfer model, and continuing to predict the clean coal ash content until a clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme is obtained;

[0042] According to the clean coal ash content prediction trajectories corresponding to multiple flotation agent dosage schemes, the optimal flotation agent dosage scheme is selected to control the clean coal ash content.

[0043] In a fifth aspect, the present application further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the following steps:

[0044] Inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data for the next moment;

[0045] Updating the clean coal ash content prediction data at the next moment to the clean coal ash content prediction data at the current moment, updating the flotation agent dosage stacking data of each flotation agent dosage scheme at the next moment to the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment, returning to the step of inputting the clean coal ash content prediction data at the current moment, the historical flotation slime input data of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into the pre-built transfer model, and continuing to predict the clean coal ash content until a clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme is obtained;

[0046] According to the clean coal ash content prediction trajectories corresponding to multiple flotation agent dosage schemes, the optimal flotation agent dosage scheme is selected to control the clean coal ash content.

[0047] The above-described clean coal ash content control method, apparatus, computer device, storage medium, and computer program product, by introducing a transition model, predict the clean coal ash content forecast for the next moment based on the current clean coal ash content forecast data, historical flotation slime data for a first preset time period, and the stacked flotation agent dosage data for each flotation agent dosage scenario at the current moment, until the clean coal ash content forecast trajectory corresponding to each flotation agent dosage scenario is predicted, that is, the clean coal ash content data for a period of time in the future is predicted. This method accurately captures the large inertia and long time lag characteristics of the coal slime flotation process and fully considers the impact of fluctuations in key parameters such as flotation slime data and flotation agent dosage data on clean coal ash content, achieving accurate, efficient, and comprehensive simulation of the actual coal slime flotation process. By combining a model control strategy algorithm, the optimal flotation agent dosage scenario can be calculated in real time based on the predicted trajectory of the transition model. This dynamic adjustment strategy effectively overcomes the hysteresis and inaccuracy caused by human factors in traditional manual dosing methods, achieving precise control of clean coal ash content. Furthermore, this method improves the quality of coal products and enhances market competitiveness by precisely controlling the ash content of clean coal. It also reduces resource waste and increased costs caused by large fluctuations in clean coal ash content. Furthermore, automated and intelligent dosing control reduces operator workload and improves clean coal production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 Schematic diagram of a process for controlling ash content in clean coal according to an embodiment;

[0050] Figure 2 A schematic diagram of a process for obtaining flotation agent dosage stacking data and historical flotation slime data for a first preset time period in one embodiment;

[0051] Figure 3 A schematic diagram of a process for constructing a transfer model in one embodiment;

[0052] Figure 4 This is a structural block diagram of a clean coal ash content control device in one embodiment;

[0053] Figure 5 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0055] In one embodiment, Figure 1 As shown, a method for controlling the ash content of clean coal is provided. This embodiment uses the method applied to a terminal as an example for illustration. It is understandable that the method can also be applied to a server, or to a system including a terminal and a server, and implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0056] Step 102: input the clean coal ash content prediction data at the current moment, the historical flotation slime data of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data at the next moment.

[0057] The clean coal ash content prediction data at the current moment refers to the clean coal ash content prediction data obtained based on the prediction. The historical flotation slime data for the first preset time period refers to the historical flotation slime data for the most recent known time period, which includes the ash content and concentration of the flotation slime. Each flotation agent dosage plan refers to the flotation agent dosage, where flotation agents include collectors and frothers. The flotation agent dosage stacked data for each flotation agent dosage plan at the current moment refers to the flotation agent dosage data for multiple consecutive moments that match the current moment when each flotation agent dosage plan is used to control the clean coal ash content. The multiple consecutive moments include the current moment.

[0058] Since coal slime flotation requires a long period of time from feeding to discharging, this period is referred to as the clean coal discharging time, denoted as T. The clean coal ash content discharged at time t is not determined solely by the coal slime data fed to the flotation process at time tT and the flotation agent dosage at that time, but rather by the combined effect of all coal slime data fed to the flotation process and the flotation agent dosage over that period. Specifically, the clean coal ash content discharged at time t is affected by the coal slime data fed to the flotation process before and after time tT, as well as the flotation agent dosage for the period before time tT. Furthermore, in the actual ash content control process, future coal slime data fed to the flotation process is unknown. Therefore, historical coal slime data fed to the flotation process for the most recent known time period, namely, historical coal slime data fed to the flotation process for the first preset time period, is obtained as future coal slime data fed to the flotation process. This future coal slime data remains unchanged during the clean coal ash content control process using each flotation agent dosage scheme, and the flotation agent dosage also remains unchanged during the clean coal ash content control process using each flotation agent dosage scheme.

[0059] Optionally, after obtaining the clean coal ash content prediction data at the current moment, the historical flotation coal slime data for the first preset time period, and the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment, the terminal inputs the clean coal ash content prediction data at the current moment, the historical flotation coal slime data for the first preset time period, and the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment into a pre-built transfer model, and predicts the clean coal ash content data at the next moment through the pre-built transfer model to obtain the clean coal ash content prediction data at the next moment.

[0060] Step 104: Update the clean coal ash content prediction data at the next moment to the clean coal ash content prediction data at the current moment, update the flotation agent dosage stacking data of each flotation agent dosage scheme at the next moment to the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment, and return to the step of inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into the pre-built transfer model, and continue to predict the clean coal ash content until the clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme is obtained.

[0061] Optionally, after obtaining the clean coal ash content prediction data for the next moment, the clean coal ash content prediction data for the next moment is updated to the clean coal ash content prediction data for the current moment. The flotation agent dosage stacking data for each flotation agent dosage scheme at the next moment is obtained, and the flotation agent dosage stacking data for each flotation agent dosage scheme at the next moment is updated to the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment. The process then returns to step 102, where the clean coal ash content prediction data for the current moment, the historical flotation slime input data for the first preset time period, and the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment are input into a pre-built transfer model to obtain the clean coal ash content prediction data for the next moment. The clean coal ash content prediction is then continuously predicted until clean coal ash content prediction data of a preset length is obtained, thereby obtaining a clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme.

[0062] When each flotation agent dosage scheme is adopted to control the clean coal ash content, the clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme is obtained according to the above clean coal ash content control method, thereby obtaining the clean coal ash content prediction trajectories corresponding to multiple flotation agent dosage schemes.

[0063] Step 106 : selecting the optimal flotation agent dosage scheme according to the clean coal ash content prediction trajectories corresponding to the multiple flotation agent dosage schemes to control the clean coal ash content.

[0064] Optionally, the clean coal ash content prediction trajectories corresponding to multiple flotation agent dosage schemes are compared with a preset clean coal ash content trajectory, and the error between the clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme and the preset clean coal ash content trajectory is determined. The flotation agent dosage scheme with the clean coal ash content prediction trajectory having the smallest error is selected as the optimal flotation agent dosage scheme, i.e., the optimal clean coal ash content control scheme. Clean coal ash content is controlled using the optimal clean coal ash content control scheme.

[0065] Furthermore, the method for determining the error between the clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme and the preset clean coal ash content trajectory may include error calculation methods such as mean absolute error and root mean square error.

[0066] Furthermore, multiple flotation agent dosage schemes can be selected through a model control strategy algorithm by uniform sampling.

[0067] Since the coal slime flotation process requires a long processing time from the entry of coal slime into the flotation to the final clean coal output, it has the characteristics of large inertia and large time lag. In addition, factors such as the data of the coal slime entering the flotation fluctuate greatly, which has a great impact on the ash content of the clean coal. As a result, manual dosing is still widely used in traditional technologies, making it difficult to achieve precise control of the ash content of the clean coal.

[0068] The above-mentioned clean coal ash content control method, by introducing a transfer model, predicts the clean coal ash content forecast for the next moment based on the current clean coal ash content forecast data, historical flotation slime data for the first preset time period, and the stacked flotation agent dosage data for each flotation agent dosage scheme at the current moment. This method predicts the clean coal ash content forecast trajectory for each flotation agent dosage scheme, thereby predicting the clean coal ash content data for a period of time in the future. This method accurately captures the large inertia and time lag characteristics of the coal slime flotation process and fully considers the impact of fluctuations in key parameters such as flotation slime data and flotation agent dosage data on clean coal ash content, achieving accurate, efficient, and comprehensive simulation of the actual coal slime flotation process. By combining the model control strategy algorithm, the optimal flotation agent dosage scheme can be calculated in real time based on the predicted trajectory of the transfer model. This dynamic adjustment strategy effectively overcomes the hysteresis and inaccuracy caused by human factors in traditional manual dosing methods, achieving precise control of clean coal ash content. Furthermore, this method improves the quality of coal products and enhances market competitiveness by precisely controlling the ash content of clean coal. It also reduces resource waste and increased costs caused by large fluctuations in clean coal ash content. Furthermore, automated and intelligent dosing control reduces operator workload and improves clean coal production efficiency.

[0069] In an exemplary embodiment, Figure 2As shown, before inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stacked data for each flotation agent dosage scheme at the current moment into the pre-built transfer model to obtain the clean coal ash content prediction data at the next moment, the method further includes: a step of acquiring the flotation agent dosage stacked data and the historical flotation slime data for the first preset time period, including:

[0070] Step 202 : determining the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment according to the flotation agent dosage stacking time and the clean coal ash content prediction step at the current moment.

[0071] Step 204 : determining historical floating coal slime data for a first preset time period based on the stacking time of the floating coal slime data and the clean coal discharge time.

[0072] The flotation agent dosage stacking duration refers to the time span of flotation agent dosage before time t that affects the clean coal ash content at time t. The flotation slime data stacking duration refers to the time span of flotation slime data before and after time tT that affects the clean coal ash content at time t.

[0073] Specifically, the clean coal ash content discharged at time t is affected by the incoming coal slime data for the period before and after time tT, as well as the flotation agent dosage for the period before time tT. The spans of the two time periods are referred to as the first incoming coal slime data stacking duration and the second incoming coal slime data stacking duration, respectively. The incoming coal slime data stacking duration is determined based on the first incoming coal slime data stacking duration and the second incoming coal slime data stacking duration. Based on the incoming coal slime data stacking duration and the clean coal discharge duration, the historical incoming coal slime data for the first preset time period is determined. The span of the period before time t is referred to as the flotation agent dosage stacking duration. Based on the flotation agent dosage stacking duration and the clean coal ash content prediction step size at the current moment, the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment is determined.

[0074] For example, the current time is recorded as , the flotation agent dosage and stacking time are recorded as , ( >0) the prediction step of clean coal ash content is i, that is, The ash content of clean coal at the moment is affected by Time has come The influence of the flotation agent dosage during the period of +i, the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment includes Time has come The flotation agent dosage at time +i. The stacking time of the first flotation slime data and the second flotation slime data are recorded as and , ( ), that is, the ash content of the clean coal at time t1 is affected by Time has come The impact of the floating coal slime data during this period of time, the historical floating coal slime data of the first preset time period includes Time has come Therefore, the historical floating coal slime data and flotation agent dosage input into the transfer model are respectively + The stacking and of stacking.

[0075] In this embodiment, the influence of fluctuations in key parameters such as flotation agent dosage and floated coal slime data on the clean coal ash content is fully considered, thereby achieving a comprehensive simulation of the actual clean coal production process.

[0076] In an exemplary embodiment, since the input of the transfer model includes observations, action quantities and external variables, the observations refer to the quantities that need to be predicted, the action quantities refer to the quantities that control the observations, and the external variables refer to other quantities that have an impact on the observations. In this embodiment, the output clean coal ash content is the observation quantity. , the amount of flotation agent added is the action amount , the floating coal slime data is an external variable , predicting the clean coal ash content at the next moment. Therefore, step 102 inputs the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data at the next moment, including: using the clean coal ash content prediction data at the current moment as the observed quantity prediction value at the current moment, using the historical flotation slime data for the first preset time period as the external variable for the first preset time period, and using the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment as the action quantity stacking data of each flotation agent dosage scheme at the current moment; and inputting the observed quantity prediction value at the current moment, the external variable for the first preset time period, and the action quantity stacking data of each flotation agent dosage scheme at the current moment into the pre-built transfer model to obtain the observed quantity prediction value at the next moment.

[0077] Specifically, the clean coal ash content prediction data at the current moment is used as the observation quantity prediction value at the current moment, the historical flotation slime data of the first preset time period is used as the external variable of the first preset time period, the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment is used as the action quantity stacking data of each flotation agent dosage scheme at the current moment, and the observation quantity prediction value at the current moment, the external variable of the first preset time period and the action quantity stacking data of each flotation agent dosage scheme at the current moment are input into a pre-constructed transfer model, and the observation quantity at the next moment is predicted by the transfer model to obtain the observation quantity prediction value at the next moment.

[0078] In this embodiment, the clean coal ash content prediction data at the current moment is used as the observed quantity prediction value at the current moment, the historical flotation slime data of the first preset time period is used as the external variable of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment is used as the action quantity stacking data of each flotation agent dosage scheme at the current moment, and is input into a pre-built transfer model. The input data that meets the model requirements can be input into the transfer model, thereby enabling the coal slime flotation process to be accurately and efficiently simulated in a virtual environment.

[0079] In an exemplary embodiment, the above method further includes: when the current moment is the first moment, obtaining the actual data of the clean coal ash content at the current moment; inputting the actual data of the clean coal ash content at the current moment, the historical floating coal slime data of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the predicted data of the clean coal ash content at the next moment.

[0080] Among them, the actual data of clean coal ash content refers to the actual output clean coal ash content data.

[0081] Optionally, when the current moment is the first moment, since there is no clean coal ash content prediction data, the actual clean coal ash content data at the current moment can be directly used as the observed value at the current moment. The historical floating coal slime data of the first preset time period is used as the external variable of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment is used as the action amount stacking data of each flotation agent dosage scheme at the current moment. The observed value at the current moment, the external variable of the first preset time period, and the action amount stacking data of each flotation agent dosage scheme at the current moment are input into the pre-built transfer model to obtain the observed value prediction value at the next moment, that is, the clean coal ash content prediction data at the next moment. When the current moment is the first moment, the clean coal ash content is predicted according to the following formula:

[0082]

[0083] in, Represents the predicted value of the observation at time t1+1; Represents a pre-built transfer model; Represents the observation value at the current time t1; Indicates the nth flotation agent dosage scheme corresponding to the current time t1 Time has come The action amount stacking data at each moment; N represents the number of samples of the flotation agent dosage scheme. The larger N is, the more flotation agent dosage schemes there are. express Time has come It should be noted that when the current moment is the first moment, the clean coal ash content prediction step length i is 0, so the above formula does not need to be calculated based on the clean coal ash content prediction step length i.

[0084] When the current moment is not the first moment, the clean coal ash content prediction data at the current moment is used as the observed quantity prediction value at the current moment, the historical flotation slime data for the first preset time period is used as the external variable for the first preset time period, and the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment is used as the action quantity stacking data for each flotation agent dosage scheme at the current moment; the observed quantity prediction value at the current moment, the external variable for the first preset time period, and the action quantity stacking data for each flotation agent dosage scheme at the current moment are input into the pre-built transfer model to obtain the observed quantity prediction value for the next moment. When the current moment is not the first moment, the clean coal ash content prediction is performed according to the following formula:

[0085]

[0086]

[0087]

[0088] Where i represents the prediction step length of clean coal ash content, the maximum value is L-1, and L represents the preset length, that is, the predicted trajectory length; express The predicted value of the observation at time; represents the predicted value of the observation at time t1+i, Indicates the nth flotation agent dosage scheme in Corresponding Time has come The amount of action at each moment is stacked.

[0089] In this embodiment, when the current moment is the first moment, the actual clean coal ash content data at the current moment is input into the transfer model as an observation quantity, which can more accurately predict the clean coal ash content in the future based on the actual clean coal ash content.

[0090] In an exemplary embodiment, before inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data at the next moment, the above method also includes a transfer model construction step and a model control strategy algorithm construction step, wherein the transfer model is a simulation of the coal slime flotation process, and the virtual modeling is completed by driving the transfer model through historical data. The model control strategy algorithm construction step refers to the construction process of a method for selecting the optimal flotation agent dosage scheme based on the clean coal ash content prediction trajectory of the pre-built transfer model. Further, the transfer model construction step is as follows. Figure 3 Shown, including:

[0091] Step 302: Obtain the clean coal ash content prediction data at the current training moment, and obtain the historical floating coal slime data and flotation agent dosage stacked data that match the current training moment; input the clean coal ash content prediction data at the current training moment, the historical floating coal slime data and flotation agent dosage stacked data that match the current training moment into the transfer model to be trained, and obtain the clean coal ash content prediction data at the next training moment.

[0092] In step 304, the clean coal ash content prediction data for the next training moment is updated to the clean coal ash content prediction data for the current training moment, and the process returns to the step of obtaining the historical floating coal slime data and flotation agent dosage stacking data that match the current training moment, and continues to predict the clean coal ash content until a prediction trajectory of a preset length is obtained.

[0093] Step 306 : Update the transfer model to be trained according to the predicted trajectory of the preset length until the preset conditions are met, thereby obtaining a pre-built transfer model.

[0094] Because the flotation agent added to the flotation machine requires a certain reaction time, simply predicting the clean coal ash content at the next moment and controlling based on this result may not necessarily achieve optimal control results. To improve control effectiveness, the transfer model must have a certain degree of inference capability, that is, the ability to infer the clean coal ash content value for a period of time in the future. If inference is performed using a transfer model trained using a traditional supervised learning algorithm, any errors will be continuously amplified as the inference progresses. Control based on erroneous results will inevitably lead to poor control results. Therefore, when training the transfer model in this embodiment, the transfer model is first allowed to generate a prediction trajectory.

[0095] Specifically, the clean coal ash content at a specific moment in time is selected from the historical clean coal ash content data as the clean coal ash content prediction data for the current training moment. Simultaneously, the historical flotation slime data and flotation agent dosage stacked data matching the current training moment are obtained. The historical flotation slime data and flotation agent dosage stacked data matching the current training moment refer to historical flotation slime data and flotation agent dosage stacked data that are correctly matched and stacked with the current training moment. In other words, the historical flotation slime data and flotation agent dosage stacked data matching each training moment are determined based on the clean coal ash content prediction step size at each training moment.

[0096] The clean coal ash content prediction data for the current training moment, along with historical flotation slime data and flotation agent dosage stacked data matching the current training moment, are input into the transfer model to be trained to obtain the clean coal ash content prediction data for the next training moment. The clean coal ash content prediction data for the next training moment is updated to the clean coal ash content prediction data for the current training moment. The process then returns to the step of obtaining historical flotation slime data and flotation agent dosage stacked data matching the current training moment, continuing to predict clean coal ash content until a prediction trajectory of a preset length is obtained. The preset length refers to the total clean coal ash content prediction step length, which can be the same as the total clean coal ash content prediction step length used in actual applications.

[0097] After obtaining a predicted trajectory of a preset length, the transfer model to be trained is updated according to the predicted trajectory of the preset length until a preset condition is satisfied, thereby obtaining a pre-built transfer model. Furthermore, the preset condition may be a training termination condition, specifically, reaching a Nash equilibrium, at which point the transfer model has approximately simulated the coal slime flotation process. When the training termination condition is reached, the transfer model at that point is used as the pre-built transfer model.

[0098] In this embodiment, the transfer model is trained based on the predicted clean coal ash content data at the current training moment, historical flotation slime data matching the current training moment, and stacked flotation agent dosage data. This fully accounts for the impact of fluctuations in key parameters such as the flotation slime data and flotation agent dosage data on the clean coal ash content, enabling the pre-built transfer model to accurately, efficiently, and comprehensively simulate the actual coal slime flotation process. By predicting a trajectory of a preset length and training the transfer model, the pre-built transfer model accurately captures the large inertia and long time lag characteristics of the coal slime flotation process.

[0099] In an optional manner of the above embodiment, the above method further includes: when the current training moment is the initial training moment, obtaining the actual data of the clean coal ash content at the current training moment; inputting the actual data of the clean coal ash content at the current training moment, the historical floating coal slime data and the flotation agent dosage stacked data matching the current training moment into the transfer model to be trained, and obtaining the predicted data of the clean coal ash content at the next training moment.

[0100] Specifically, when the current training moment is the first training moment, since there is no clean coal ash content prediction data, the actual clean coal ash content data at the current training moment can be directly used as the observation value at the current training moment. The historical floating coal slime data that matches the current training moment is used as the external variable that matches the current training moment, and the flotation agent dosage stacking data that matches the current training moment is used as the action quantity stacking data that matches the current training moment. The observation value at the current training moment, the external variable that matches the current training moment, and the action quantity are input into the transfer model to be trained to obtain the observation value prediction value at the next training moment, that is, the clean coal ash content prediction data at the next training moment. When the current training moment is the first training moment, the clean coal ash content prediction is performed according to the following formula:

[0101]

[0102] in, Represents the predicted value of the observation at training time t2+1; represents the transfer model to be trained; Represents the observation value at the current training time t2; Indicates that it matches the current training time t2 Time has come The amount of action at each moment is stacked data; Indicates that it matches the current training time t2 Time has come It should be noted that when the current training moment is the first training moment, the clean coal ash content prediction step length i is 0, so the above formula does not need to be calculated based on the clean coal ash content prediction step length i.

[0103] If the current training moment is not the first training moment, the clean coal ash content prediction data at the current training moment is used as the observed value prediction value at the current training moment, the historical flotation slime data matching the current training moment is used as the external variable matching the current training moment, and the flotation agent dosage stacked data matching the current training moment is used as the action quantity stacked data matching the current training moment. The observed value at the current training moment, the external variables matching the current training moment, and the action quantity are input into the transfer model to be trained to obtain the observed value prediction value for the next training moment. If the current training moment is not the first training moment, the clean coal ash content prediction is performed according to the following formula:

[0104]

[0105]

[0106] Where i represents the prediction step length of clean coal ash content, the maximum value is L-1, and L represents the preset length, that is, the predicted trajectory length; express The predicted value of the observation at the training time; express The predicted value of the observation at the training time; Represents Matching training time Time has come The amount of action at each moment is stacked data; Represents Matching training time Time has come External variables at the moment.

[0107] In this embodiment, when the current training moment is the first training moment, the actual data of the clean coal ash content at the current training moment is input as an observation quantity into the transfer model to be trained, so that the transfer model can make a more accurate prediction of the clean coal ash content in the future based on the actual clean coal ash content.

[0108] In an optional manner of the above embodiment, in step 306, updating the transfer model to be trained based on the predicted trajectory of the preset length includes: determining the error between the predicted trajectory of the preset length and the true trajectory through a discriminant network; and updating the transfer model to be trained based on the error between the predicted trajectory of the preset length and the true trajectory.

[0109] Optionally, the transfer model is trained using a generative adversarial strategy. Specifically, after obtaining a predicted trajectory of a preset length, a discriminant network is used to determine the error between the predicted trajectory of the preset length and the true trajectory. The discriminant network can be any of the types known to those skilled in the art, such as a fully connected neural network, a convolutional neural network, a recurrent neural network, a conditional discriminant network, a self-attention network, etc. Based on the error between the predicted trajectory of the preset length and the true trajectory, the transfer model to be trained is updated. Specifically, the discriminant network is used to distinguish between the predicted trajectory of the preset length and the true trajectory, that is, to correctly identify whether the input trajectory is true or generated by the transfer model. The transfer model is used to achieve the most accurate prediction possible, that is, to generate a predicted trajectory that is as close to the true trajectory as possible. During the training process, the discriminant network and the transfer model compete with each other, continuously playing against each other until a Nash equilibrium is reached. At this point, the discriminant network cannot distinguish whether the predicted trajectory generated by the transfer model is true or generated, indicating that the transfer model has approximately simulated the coal slime flotation process. The transfer model at this point is then used as the pre-built transfer model.

[0110] In this embodiment, the error between the predicted trajectory of a preset length and the actual trajectory is determined by a discriminant network, and the transfer model to be trained is updated based on the error. This enables the transfer model to be constructed based on a generative adversarial strategy, so that the coal slime flotation process can be accurately and efficiently simulated in a virtual environment, and a transfer model that can accurately predict the trajectory of the clean coal ash content is obtained.

[0111] In an optional manner of the above embodiment, step 306, updating the transfer model to be trained according to the predicted trajectory of a preset length until a preset condition is satisfied to obtain a pre-constructed transfer model, includes: updating the transfer model to be trained according to the predicted trajectory of the preset length until the training condition is satisfied to obtain a trained transfer model; performing a model test on the trained transfer model, and obtaining the pre-constructed transfer model when the trained transfer model passes the model test; the model test includes a deduction test, a response curve test, and a data distribution test.

[0112] Optionally, during the construction of the transfer model, after the model training is completed, the trained transfer model may be subjected to a model test. Only by passing the model test can the correctness of the constructed transfer model be fully demonstrated.

[0113] Specifically, model testing can include deduction testing, response curve testing, and data distribution testing. Passing the deduction test means that the generated predicted trajectory is as close to the actual trajectory as possible. Passing the response curve test means that when the action amount or external variable changes, the predicted value of the observed quantity should also change correctly. Passing the data distribution test means that when the transfer model is accurate enough to simulate the actual coal slime flotation process, the data distribution of the output generated by the model based on historical data should be similar to the data distribution of the actual output. For example, the data distribution can be the predicted data of clean coal ash content, that is, the probability distribution of the observed quantity.

[0114] Furthermore, the trained model can be tested in the order of deduction test, response curve test and data distribution test.

[0115] In this embodiment, the correctness of the transfer model can be further ensured by performing deduction testing, response curve testing, and data distribution testing on the trained transfer model.

[0116] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0117] Based on the same inventive concept, embodiments of the present application further provide a clean coal ash control device for implementing the above-mentioned clean coal ash control method. The solution provided by this device is similar to the solution described in the above-mentioned method. Therefore, the specific limitations of one or more embodiments of the clean coal ash control device provided below can be found in the above-mentioned limitations of the clean coal ash control method and will not be further elaborated here.

[0118] In an exemplary embodiment, Figure 4 As shown, a clean coal ash content control device is provided, comprising: an ash content prediction module 402, a cycle prediction module 404 and a scheme selection module 406, wherein:

[0119] The ash content prediction module 402 is used to input the clean coal ash content prediction data at the current moment, the historical flotation slime data of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data at the next moment.

[0120] The cyclic prediction module 404 is used to update the clean coal ash content prediction data at the next moment to the clean coal ash content prediction data at the current moment, update the flotation agent dosage stacking data of each flotation agent dosage scheme at the next moment to the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment, return to the step of inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into the pre-built transfer model, and continue to predict the clean coal ash content until the clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme is obtained.

[0121] The scheme selection module 406 is used to select the optimal flotation agent dosage scheme according to the clean coal ash content prediction trajectories corresponding to the multiple flotation agent dosage schemes to control the clean coal ash content.

[0122] In an exemplary embodiment, the apparatus further comprises:

[0123] The data acquisition module is used to determine the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment based on the flotation agent dosage stacking time and the clean coal ash content prediction step at the current moment; and determine the historical flotation coal slime data for the first preset time period based on the flotation coal slime data stacking time and the clean coal discharge time.

[0124] In an exemplary embodiment, the ash content prediction module 402 is further configured to use the clean coal ash content prediction data at the current moment as the observation quantity prediction value at the current moment, use the historical flotation slime data of the first preset time period as the external variable of the first preset time period, and use the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment as the action quantity stacking data of each flotation agent dosage scheme at the current moment; input the observation quantity prediction value at the current moment, the external variable of the first preset time period, and the action quantity stacking data of each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the observation quantity prediction value at the next moment.

[0125] In an exemplary embodiment, the ash content prediction module 402 is further configured to obtain actual clean coal ash content data at the current moment when the current moment is the first moment; input the actual clean coal ash content data at the current moment, historical flotation slime data for the first preset time period, and flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain predicted clean coal ash content data for the next moment.

[0126] In an exemplary embodiment, the apparatus further comprises:

[0127] A model construction module is used to obtain the clean coal ash content prediction data at the current training moment, and obtain the historical floating coal slime data and flotation agent dosage stacked data that match the current training moment; input the clean coal ash content prediction data at the current training moment, the historical floating coal slime data and flotation agent dosage stacked data that match the current training moment into the transfer model to be trained, and obtain the clean coal ash content prediction data for the next training moment; update the clean coal ash content prediction data for the next training moment to the clean coal ash content prediction data at the current training moment, return to the step of obtaining the historical floating coal slime data and flotation agent dosage stacked data that match the current training moment, and continue to predict the clean coal ash content until a prediction trajectory of a preset length is obtained; and update the transfer model to be trained according to the prediction trajectory of the preset length until a preset condition is met, thereby obtaining a pre-constructed transfer model.

[0128] In an exemplary embodiment, the model building module is further used to obtain the actual data of clean coal ash content at the current training moment when the current training moment is the initial training moment; input the actual data of clean coal ash content at the current training moment, the historical floating coal slime data and flotation agent dosage stacked data matching the current training moment into the transfer model to be trained, and obtain the predicted data of clean coal ash content at the next training moment.

[0129] In an exemplary embodiment, the model building module is further configured to determine an error between a predicted trajectory of a preset length and a true trajectory through a discriminant network; and update the transfer model to be trained according to the error between the predicted trajectory of the preset length and the true trajectory.

[0130] In an exemplary embodiment, the model construction module is further used to update the transfer model to be trained according to the predicted trajectory of a preset length until the training conditions are met, thereby obtaining a trained transfer model; the trained transfer model is subjected to model testing, and when the model test passes, a pre-constructed transfer model is obtained; the model test includes a deduction test, a response curve test, and a data distribution test.

[0131] Each module in the above-mentioned clean coal ash content control device can be implemented in whole or in part through software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of the processor of the computer device in hardware form, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each of the above modules.

[0132] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 5As shown. The computer device includes a processor, memory, an input / output interface, a communication interface, a display unit, and an input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals via wired or wireless means, and the wireless means can be achieved via Wi-Fi, mobile cellular networks, NFC (near-field communication), or other technologies. When executed by the processor, the computer program implements a method for controlling the ash content of clean coal. The display unit of the computer device is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device casing, or an external keyboard, touchpad or mouse.

[0133] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0134] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0135] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0136] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0137] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), data processing logic devices based on quantum computing, and the like.

[0138] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0139] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for controlling ash content in clean coal, characterized in that: The method comprises: Inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data for the next moment; updating the clean coal ash content prediction data at the next moment to the clean coal ash content prediction data at the current moment, updating the flotation agent dosage stacking data of each flotation agent dosage scheme at the next moment to the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment, returning to the step of inputting the clean coal ash content prediction data at the current moment, the historical flotation slime input data of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into the pre-built transfer model, and continuing to predict the clean coal ash content until a clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme is obtained; According to the clean coal ash content prediction trajectories corresponding to multiple flotation agent dosage plans, the optimal flotation agent dosage plan is selected to control the clean coal ash content; Before inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stack data for each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data at the next moment, the method further includes: Determine the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment according to the flotation agent dosage stacking time and the clean coal ash content prediction step at the current moment; According to the stacking time of the floating coal slime data and the clean coal discharging time, the historical floating coal slime data for the first preset time period is determined.

2. The method according to claim 1, characterized in that The clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stacking data for each flotation agent dosage scheme at the current moment are input into a pre-built transfer model to obtain the clean coal ash content prediction data at the next moment, including: The clean coal ash content prediction data at the current moment is used as the observation quantity prediction value at the current moment, the historical flotation slime data of the first preset time period is used as the external variable of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment is used as the action quantity stacking data of each flotation agent dosage scheme at the current moment; The predicted value of the observed quantity at the current moment, the external variables of the first preset time period, and the stacked data of the action amount of each flotation agent dosage scheme at the current moment are input into a pre-built transfer model to obtain the predicted value of the observed quantity at the next moment.

3. The method according to claim 1, characterized in that The method further comprises: When the current moment is the first moment, obtaining actual data of clean coal ash content at the current moment; The actual clean coal ash content data at the current moment, the historical flotation slime data of the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment are input into the pre-built transfer model to obtain the clean coal ash content prediction data at the next moment.

4. The method according to any one of claims 1 to 3, characterized in that Before inputting the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stack data for each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data at the next moment, the method further includes a transfer model construction step, including: Acquire clean coal ash content prediction data at the current training moment, and acquire historical flotation slime data and flotation agent dosage stacked data that match the current training moment; input the clean coal ash content prediction data at the current training moment, the historical flotation slime data and flotation agent dosage stacked data that match the current training moment into the transfer model to be trained, and obtain clean coal ash content prediction data at the next training moment; Updating the clean coal ash content prediction data for the next training moment to the clean coal ash content prediction data for the current training moment, returning to the step of obtaining historical floating coal slime data and flotation agent dosage stacking data that match the current training moment, and continuing to predict the clean coal ash content until a prediction trajectory of a preset length is obtained; The transfer model to be trained is updated according to the predicted trajectory of the preset length until a preset condition is met, thereby obtaining a pre-constructed transfer model.

5. The method according to claim 4, characterized in that The method further comprises: When the current training moment is the initial training moment, obtaining actual clean coal ash content data at the current training moment; The actual clean coal ash content data at the current training moment, the historical flotation slime data and flotation agent dosage stacked data matching the current training moment are input into the transfer model to be trained to obtain the clean coal ash content prediction data at the next training moment.

6. The method according to claim 4, characterized in that Updating the transfer model to be trained according to the predicted trajectory of the preset length includes: Determine the error between the predicted trajectory of the preset length and the actual trajectory through a discriminant network; The transfer model to be trained is updated according to the error between the predicted trajectory of the preset length and the actual trajectory.

7. The method according to claim 4, characterized in that The step of updating the transfer model to be trained according to the predicted trajectory of the preset length until a preset condition is satisfied to obtain a pre-built transfer model comprises: Updating the transfer model to be trained according to the predicted trajectory of the preset length until the training conditions are met, thereby obtaining a trained transfer model; The trained transfer model is subjected to a model test, and when the model test passes, a pre-built transfer model is obtained; the model test includes a deduction test, a response curve test, and a data distribution test.

8. A clean coal ash content control device, characterized in that: The device comprises: An ash content prediction module is used to input the clean coal ash content prediction data at the current moment, the historical flotation slime data for the first preset time period, and the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment into a pre-built transfer model to obtain the clean coal ash content prediction data at the next moment; a cyclic prediction module, configured to update the clean coal ash content prediction data at the next moment to the clean coal ash content prediction data at the current moment, update the flotation agent dosage stacked data for each flotation agent dosage scheme at the next moment to the flotation agent dosage stacked data for each flotation agent dosage scheme at the current moment, return to the step of inputting the clean coal ash content prediction data at the current moment, the historical flotation slime input data for the first preset time period, and the flotation agent dosage stacked data for each flotation agent dosage scheme at the current moment into a pre-built transfer model, and continue to predict the clean coal ash content until a clean coal ash content prediction trajectory corresponding to each flotation agent dosage scheme is obtained; A scheme selection module is used to select the optimal flotation agent dosage scheme based on the clean coal ash content prediction trajectory corresponding to multiple flotation agent dosage schemes to control the clean coal ash content; The device further comprises: The data acquisition module is used to determine the flotation agent dosage stacking data of each flotation agent dosage scheme at the current moment based on the flotation agent dosage stacking time and the clean coal ash content prediction step at the current moment; and determine the historical flotation coal slime data for the first preset time period based on the flotation coal slime data stacking time and the clean coal discharge time.

9. The device according to claim 8, characterized in that The ash content prediction module is further configured to use the clean coal ash content prediction data at the current moment as the observation quantity prediction value at the current moment, use the historical flotation slime data for the first preset time period as the external variable for the first preset time period, and use the flotation agent dosage stacked data for each flotation agent dosage scheme at the current moment as the action quantity stacked data for each flotation agent dosage scheme at the current moment; and input the observation quantity prediction value at the current moment, the external variable for the first preset time period, and the action quantity stacked data for each flotation agent dosage scheme at the current moment into a pre-constructed transfer model to obtain the observation quantity prediction value for the next moment.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

Images