Intelligent optimization method and system for operation of waste combustion device, and medium

Abstract

The present application provides an intelligent optimization method and system for the operation of a waste combustion device, and a medium. The method comprises: acquiring a preset number of pieces of historical input variable data and historical output variable data of a waste combustion device, then classifying the historical input variable data and the corresponding historical output variable data to obtain a training data set and a test data set; acquiring key parameters of a preset gradient-boosted tree, and performing training to obtain an updated gradient-boosted tree model; performing calculation on the basis of the updated gradient-boosted tree model to obtain an input variable importance index, and obtaining an optimized input variable by comparing the input variable importance index with a threshold; performing processing by means of a multi-layer BP neural network and a particle swarm optimization algorithm to obtain a thermal efficiency prediction model; processing the thermal efficiency prediction model by means of a reinforcement learning algorithm to obtain optimal input variable feature data. Therefore, intelligent optimization of the waste combustion device is realized by means of the gradient-boosted tree model, the multi-layer BP neural network and the particle swarm optimization algorithm, thereby reducing the power generation costs of the waste combustion device.

Description

A method, system, and medium for intelligent optimization of waste incineration equipment operation.

[0001] Technical Field

[0002] This application relates to the field of waste incineration, and more specifically, to an intelligent optimization method, system, and medium for the operation of waste incineration equipment. Background Technology

[0003] With the acceleration of urbanization, the amount of waste is increasing daily, and waste incineration has become a widely used effective waste treatment method. However, current waste incineration equipment faces numerous problems during operation. For example, the operation of waste incinerators relies heavily on individual experience, leading to incomplete combustion and low effective conversion rates. Furthermore, traditional waste incineration equipment operation and control methods are often based on fixed parameters or simple feedback mechanisms, making it difficult to adapt to the dynamic changes in waste characteristics. They cannot effectively adjust key parameters such as temperature, air supply, and combustion time in real time and with precision, thus affecting the efficiency and environmental benefits of waste incineration. In addition, the combustion process involves numerous parameters, and optimizing the control of these parameters is also a challenge. Effective technical solutions are urgently needed to address these problems. Summary of the Invention

[0004] The purpose of this application is to provide an intelligent optimization method, system, and medium for the operation of waste incineration equipment. This application first analyzes the parameters in the combustion process through a gradient boosting tree model to screen out important parameters, thereby reducing the number of parameters to be optimized. In addition, it realizes intelligent optimization of waste incineration equipment through multi-layer BP neural network and particle swarm optimization algorithm technology, and further optimizes the parameters in the combustion process through reinforcement learning to obtain the highest boiler thermal efficiency and reduce the power generation cost of waste incineration equipment.

[0005] This application also provides an intelligent optimization method for the operation of waste incineration equipment, including the following steps:

[0006] Obtain historical operational data of a preset number of waste incineration devices, including historical input variable data and historical output variable data. Establish a correspondence between the historical input variable data and the corresponding historical output variable data and classify them to obtain training datasets and test datasets. Obtain the key parameters of the preset gradient boosting tree, and train the gradient boosting tree based on the training dataset, test dataset and key parameters to obtain an updated gradient boosting tree model.

[0007] The importance index of the input variable is calculated based on the updated gradient boosting tree model, and the importance index of the input variable is compared with the preset optimized input variable threshold to obtain the optimized input variable;

[0008] An initial thermal efficiency prediction model is established by combining the optimized input variable data and output variable data with a preset multi-layer BP neural network, and the thermal efficiency prediction model is obtained by processing with a preset particle swarm optimization algorithm.

[0009] The thermal efficiency prediction model is processed by a preset reinforcement learning algorithm to obtain the optimal input variable feature data corresponding to the optimized input variables.

[0010] Optionally, in the intelligent optimization method for the operation of waste incineration equipment described in this application, the step of obtaining a preset number of historical operating data of the waste incineration equipment, including historical input variable data and historical output variable data, establishing a correspondence between the historical input variable data and the corresponding historical output variable data, and classifying them to obtain training datasets and test datasets, specifically includes:

[0011] Establish a correspondence between historical input variable data and corresponding historical output variable data, and group them according to a preset ratio to obtain training dataset and test dataset;

[0012] The training dataset includes training history input variable data and corresponding training history output variable data;

[0013] The test dataset includes historical test input variable data and corresponding historical test output variable data.

[0014] Optionally, in the intelligent optimization method for the operation of waste incineration equipment described in this application, the step of obtaining the key parameters of a preset gradient boosting tree and training the gradient boosting tree based on the training dataset, the test dataset, and the key parameters to obtain an updated gradient boosting tree model specifically includes:

[0015] Obtain the key parameters of the preset gradient boosting tree, including the number of decision trees, learning rate data, and maximum depth data of the decision trees;

[0016] Obtain the average value of the historical output variable data during training and mark it as the initial value for the model;

[0017] The model is iterated based on the initial values, historical input variables, and corresponding historical output variables, combined with the number of decision trees. The model is then tested using historical input variables and corresponding historical output variables to obtain an updated gradient boosting tree model.

[0018] Optionally, in the intelligent optimization method for the operation of waste incineration equipment described in this application, the step of calculating the importance index of the input variable based on the updated gradient boosting tree model and comparing the importance index of the input variable with a preset optimization input variable threshold to obtain the optimized input variable specifically includes:

[0019] The importance index of each input variable is obtained by processing it using a gradient boosting tree model;

[0020] Calculate the sum of the importance indices of each input variable and record it as the total importance.

[0021] The importance percentage of each input variable is obtained by dividing the importance index of each input variable by the sum of the importance indices.

[0022] Sort the important percentage data of the input variables in descending order;

[0023] Select input variables that are greater than a preset importance threshold and mark them as optimized input variables.

[0024] Optionally, in the intelligent optimization method for the operation of waste incineration equipment described in this application, the step of establishing an initial thermal efficiency prediction model based on the optimization input variable data and output variable data corresponding to the optimization input variables, combined with a preset multi-layer BP neural network, and obtaining the thermal efficiency prediction model through a preset particle swarm optimization algorithm, specifically includes:

[0025] An initial thermal efficiency prediction model is established by using a pre-set multi-layer BP neural network and combining optimized input and output variable data.

[0026] Obtain the predicted and actual values ​​of the output variables, calculate the mean squared error based on the predicted and actual values ​​of the output variables, and mark it as the fitness value;

[0027] The initial thermal efficiency prediction model is optimized using a preset method based on the fitness value to obtain the thermal efficiency prediction model.

[0028] Optionally, in the intelligent optimization method for the operation of waste incineration equipment described in this application, the step of processing the thermal efficiency prediction model through a preset reinforcement learning algorithm to obtain the optimal input variable feature data corresponding to the optimized input variables specifically includes:

[0029] S51. Initialize the input variables to be optimized;

[0030] S52. The agent obtains the predicted output variable based on the input variable to be optimized;

[0031] S53. The output variable reward value and the sum of the output variable reward values ​​are obtained by processing each output variable data through a preset reward function; the agent of the reinforcement learning is the thermal efficiency prediction model.

[0032] S54. Adjust the optimized input variable data according to the sum of the output variable reward values, and then repeat steps S51-S53 until the optimal input variable feature data is obtained.

[0033] Secondly, this application provides an intelligent optimization system for the operation of a waste incineration device. The system includes a memory and a processor. The memory includes a program for an intelligent optimization method for the operation of the waste incineration device. When executed by the processor, the intelligent optimization method program for the operation of the waste incineration device performs the following steps:

[0034] Obtain historical operational data of a preset number of waste incineration devices, including historical input variable data and historical output variable data. Establish a correspondence between the historical input variable data and the corresponding historical output variable data and classify them to obtain training datasets and test datasets.

[0035] Obtain the key parameters of the preset gradient boosting tree, and train the gradient boosting tree based on the training dataset, test dataset, and key parameters to obtain the updated gradient boosting tree model;

[0036] The importance index of the input variable is calculated based on the updated gradient boosting tree model, and the importance index of the input variable is compared with the preset optimized input variable threshold to obtain the optimized input variable;

[0037] An initial thermal efficiency prediction model is established by combining the optimized input variable data and output variable data with a preset multi-layer BP neural network, and the thermal efficiency prediction model is obtained by processing with a preset particle swarm optimization algorithm.

[0038] The thermal efficiency prediction model is processed by a preset reinforcement learning algorithm to obtain the optimal input variable feature data corresponding to the optimized input variables.

[0039] The acquisition of historical operational data for a preset number of waste incineration devices includes historical input variable data and historical output variable data. A correspondence is established between the historical input variable data and the corresponding historical output variable data, and the data is classified to obtain training and testing datasets. Specifically, this includes:

[0040] Establish a correspondence between historical input variable data and corresponding historical output variable data, and group them according to a preset ratio to obtain training dataset and test dataset;

[0041] The training dataset includes training history input variable data and corresponding training history output variable data;

[0042] The test dataset includes historical test input variable data and corresponding historical test output variable data.

[0043] The process of obtaining the key parameters of the preset gradient boosting tree and training the gradient boosting tree based on the training dataset, test dataset, and key parameters to obtain the updated gradient boosting tree model specifically includes:

[0044] Obtain the key parameters of the preset gradient boosting tree, including the number of decision trees, learning rate data, and maximum depth data of the decision trees;

[0045] Obtain the average value of the historical output variable data during training and mark it as the initial value for the model;

[0046] The model is iterated based on the initial values, historical input variables, and corresponding historical output variables, combined with the number of decision trees. The model is then tested using historical input variables and corresponding historical output variables to obtain an updated gradient boosting tree model.

[0047] The step of calculating the importance index of the input variable based on the updated gradient boosting tree model, and comparing the importance index of the input variable with a preset optimization input variable threshold to obtain the optimized input variable, specifically includes:

[0048] The importance index of each input variable is obtained by processing it using a gradient boosting tree model;

[0049] Calculate the sum of the importance indices of each input variable and record it as the total importance.

[0050] The importance percentage of each input variable is obtained by dividing the importance index of each input variable by the sum of the importance indices.

[0051] Sort the important percentage data of the input variables in descending order;

[0052] Select input variables that are greater than a preset importance threshold and mark them as optimized input variables.

[0053] The step of calculating the importance index of the input variable based on the updated gradient boosting tree model, and comparing the importance index of the input variable with a preset optimization input variable threshold to obtain the optimized input variable, specifically includes:

[0054] The importance index of each input variable is obtained by processing it using a gradient boosting tree model;

[0055] Calculate the sum of the importance indices of each input variable and record it as the total importance.

[0056] The importance percentage of each input variable is obtained by dividing the importance index of each input variable by the sum of the importance indices.

[0057] Sort the important percentage data of the input variables in descending order;

[0058] Select input variables that are greater than a preset importance threshold and mark them as optimized input variables.

[0059] The process of establishing an initial thermal efficiency prediction model based on the optimized input variable data and output variable data corresponding to the optimized input variables, combined with a preset multi-layer BP neural network, and obtaining the thermal efficiency prediction model through a preset particle swarm optimization algorithm, specifically includes:

[0060] An initial thermal efficiency prediction model is established by using a pre-set multi-layer BP neural network and combining optimized input and output variable data.

[0061] Obtain the predicted and actual values ​​of the output variables, calculate the mean squared error based on the predicted and actual values ​​of the output variables, and mark it as the fitness value;

[0062] The initial thermal efficiency prediction model is optimized using a preset method based on the fitness value to obtain the thermal efficiency prediction model.

[0063] The step of processing the thermal efficiency prediction model using a preset reinforcement learning algorithm to obtain the optimal input variable feature data corresponding to the optimized input variables specifically includes:

[0064] Initialize the input variables to be optimized;

[0065] The agent obtains the predicted output variable based on the input variable to be optimized;

[0066] The output variable reward value and the sum of the output variable reward values ​​are obtained by processing each output variable data through a preset reward function; the agent of the reinforcement learning is the thermal efficiency prediction model;

[0067] The optimized input variable data is adjusted according to the sum of the output variable reward values, and then the above steps are repeated until the sum of the reward values ​​no longer changes. At this point, the optimal input variable feature data is obtained.

[0068] Optionally, in the intelligent optimization system for the operation of waste incineration equipment described in this application, the step of obtaining a preset number of historical operating data of the waste incineration equipment, including historical input variable data and historical output variable data, establishing a correspondence between the historical input variable data and the corresponding historical output variable data, and classifying them to obtain training datasets and test datasets, specifically includes:

[0069] Establish a correspondence between historical input variable data and corresponding historical output variable data, and group them according to a preset ratio to obtain training dataset and test dataset;

[0070] The training dataset includes training history input variable data and corresponding training history output variable data;

[0071] The test dataset includes historical test input variable data and corresponding historical test output variable data.

[0072] Optionally, in the intelligent optimization system for the operation of the waste incineration equipment described in this application, the step of obtaining the key parameters of the preset gradient boosting tree and training the gradient boosting tree based on the training dataset, the test dataset, and the key parameters to obtain an updated gradient boosting tree model specifically includes:

[0073] Obtain the key parameters of the preset gradient boosting tree, including the number of decision trees, learning rate data, and maximum depth data of the decision trees;

[0074] Obtain the average value of the historical output variable data during training and mark it as the initial value for the model;

[0075] The model is iterated based on the initial values, historical input variables, and corresponding historical output variables, combined with the number of decision trees. The model is then tested using historical input variables and corresponding historical output variables to obtain an updated gradient boosting tree model.

[0076] Thirdly, this application also provides a readable storage medium storing a program for an intelligent optimization method for the operation of a waste incineration device. When the program is executed by a processor, it implements the steps of an intelligent optimization method for the operation of a waste incineration device as described in any of the preceding claims.

[0077] As described above, this application provides an intelligent optimization method, system, and medium for the operation of waste incineration equipment. This method acquires a preset number of historical input and output variable data for the waste incineration equipment. Then, it classifies the historical input and output variable data to obtain training and test datasets. Key parameters for a gradient boosting tree are obtained and trained to obtain an updated gradient boosting tree model. Based on the updated gradient boosting tree model, the importance index of the input variables is calculated, and optimized input variables are obtained through threshold comparison. A thermal efficiency prediction model is obtained through multi-layer BP neural network and particle swarm optimization algorithm. The thermal efficiency prediction model is then processed by a reinforcement learning algorithm to obtain the optimal input variable feature data. Thus, intelligent optimization of the waste incineration equipment is achieved through the gradient boosting tree model, multi-layer BP neural network, and particle swarm optimization algorithm, reducing the power generation cost of the waste incineration equipment.

[0078] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings. Attached Figure Description

[0079] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0080] Figure 1 is a flowchart of an intelligent optimization method for the operation of a waste incineration device provided in an embodiment of this application;

[0081] Figure 2 is a flowchart illustrating the process of obtaining training and test datasets for an intelligent optimization method for the operation of a waste incineration device according to an embodiment of this application.

[0082] Figure 3 is a flowchart of obtaining and updating the gradient boosting tree model of an intelligent optimization method for the operation of a waste incineration device provided in an embodiment of this application;

[0083] Figure 4 is a flowchart of the optimization input variables for an intelligent optimization method for the operation of a waste incineration device provided in an embodiment of this application;

[0084] Figure 5 is a schematic diagram illustrating the importance of the input parameters of the incinerator in an embodiment of this application;

[0085] Figure 6 is a schematic diagram of the number of iterations in reinforcement learning according to an embodiment of this application;

[0086] Figure 7 is a schematic diagram of the fitness values ​​of particles in an embodiment of this application. Detailed Implementation

[0087] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0088] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0089] Please refer to Figure 1, which is a flowchart of an intelligent optimization method for the operation of a waste incineration device according to some embodiments of this application. This intelligent optimization method for the operation of the waste incineration device is used in terminal devices, such as computers and mobile phones. The intelligent optimization method for the operation of the waste incineration device includes the following steps:

[0090] S11. Obtain historical operating data of a preset number of waste incineration equipment, including historical input variable data and historical output variable data. Establish a correspondence between the historical input variable data and the corresponding historical output variable data and classify them to obtain training dataset and test dataset.

[0091] S12. Obtain the key parameters of the preset gradient boosting tree, and train the gradient boosting tree according to the training dataset, test dataset and key parameters to obtain the updated gradient boosting tree model.

[0092] S13. Calculate the importance index of the input variable according to the updated gradient boosting tree model, and compare the importance index of the input variable with the preset optimized input variable threshold to obtain the optimized input variable;

[0093] S14. Based on the optimized input variable data and output variable data corresponding to the optimized input variable, and combined with the preset multi-layer BP neural network, an initial thermal efficiency prediction model is established, and the thermal efficiency prediction model is obtained by processing it through the preset particle swarm optimization algorithm.

[0094] S15. The thermal efficiency prediction model is processed by a preset reinforcement learning algorithm to obtain the optimal input variable feature data corresponding to the optimized input variable.

[0095] It should be noted that the operation of waste incineration equipment is a relatively complex process involving many factors. Therefore, a neural network model is required in the data processing. First, a predetermined amount of historical operational data from the waste incineration equipment must be obtained, including historical input and output variable data. There are multiple historical input and output variables, and a correspondence exists between them. To better utilize this historical data, classification is performed to obtain training and testing datasets. Key parameters for a predetermined gradient boosting tree are then obtained, and the model is trained to obtain an updated gradient boosting tree model. Some input variables may have insignificant effects; therefore, the importance index of the input variables needs to be calculated by updating the gradient boosting tree model. Then, the optimized input variables are obtained by comparing the threshold of the input variable importance index. The optimized input variables contain multiple influencing factors, but the number of influencing factors is less than the number of influencing factors in the input variable data. Then, a thermal efficiency prediction model is obtained through a predetermined multi-layer backpropagation neural network and a predetermined particle swarm optimization algorithm. Finally, a predetermined reinforcement learning algorithm is used to process the thermal efficiency prediction model to obtain the optimal input variable feature data, thereby clarifying the adjustment target of the input variables.

[0096] Please refer to Figure 2, which is a flowchart illustrating the acquisition of training and test datasets in an intelligent optimization method for the operation of waste incineration equipment according to an embodiment of this application. According to this embodiment, acquiring a preset number of historical operating data of the waste incineration equipment, including historical input variable data and historical output variable data, and establishing a correspondence between the historical input variable data and the corresponding historical output variable data and classifying them to obtain the training and test datasets, specifically includes:

[0097] S21. Establish a correspondence between historical input variable data and corresponding historical output variable data, and group them according to a preset ratio to obtain training dataset and test dataset;

[0098] S22. The training dataset includes training history input variable data and corresponding training history output variable data;

[0099] S23. The test dataset includes test history input variable data and corresponding test history output variable data.

[0100] It should be noted that historical input variable data is the independent variable, and historical output variable data is the dependent variable; therefore, there is a corresponding relationship between the two. In this embodiment, the historical input variable data includes historical feeder speed data, historical main steam flow rate data, historical main steam temperature data, historical main steam pressure data, historical furnace temperature data, historical feedwater temperature data, historical feedwater flow rate data, historical primary air temperature data, historical induced draft fan frequency data, historical secondary air volume data, historical flue gas temperature data, historical primary air volume data of the first grate, historical primary air volume data of the second grate, and historical primary air volume data of the third grate. The historical primary air volume data for the 4th and 5th grate sections, the historical operating speed data for the 1st, 2nd, 3rd, 4th, and 5th grate sections, and the historical output variable data include historical flue gas oxygen content data, historical furnace negative pressure data, historical furnace outlet temperature data, and historical unit steam volume data. The preset ratio can be set according to user needs. In this embodiment, the ratio of the corresponding data quantity in the training dataset to the corresponding data quantity in the test dataset is 7:3.

[0101] In this embodiment, the combustion process has 21 input variables and 4 output variables (i.e., flue gas oxygen content, furnace negative pressure, furnace outlet temperature, and unit steam quantity). This embodiment aims to adjust the incinerator combustion parameters to maximize the unit steam quantity while maintaining the flue gas oxygen content, furnace negative pressure, and furnace outlet temperature within a reasonable range. However, with 21 combustion parameters, using all of them would significantly increase the training time for subsequent neural networks and reinforcement learning, resulting in a large computational load. This embodiment aims to reduce the number of incinerator parameters to decrease the computational load.

[0102] Please refer to Figure 3, which is a flowchart illustrating the process of obtaining an updated gradient boosting tree model using an intelligent optimization method for the operation of a waste incineration device according to an embodiment of this application. According to this embodiment, obtaining the key parameters of a preset gradient boosting tree and training the gradient boosting tree based on the training dataset, test dataset, and key parameters to obtain an updated gradient boosting tree model specifically includes:

[0103] S31. Obtain the key parameters of the preset gradient boosting tree, including the number of decision trees, learning rate data, and maximum depth data of the decision trees;

[0104] S32. Obtain the average value of the historical output variable data during training and mark it as the initial value of the model;

[0105] S33. Iterate based on the initial values ​​of the model, the historical input variable data of training, and the corresponding historical output variable data of training, combined with the number of decision trees, and test with the historical input variable data of testing and the corresponding historical output variable data of testing to obtain the updated gradient boosting tree model.

[0106] It's important to note that the number of decision trees refers to the number of decision trees to be generated, which is the number of iterations required in the entire training process; the learning rate data controls the impact of each newly generated decision tree on the previously inaccurate predictions (residuals); the maximum depth of the decision tree data indicates the maximum depth a decision tree can reach; the average of the historical output variable data is calculated by taking the average of historical flue gas oxygen content, historical furnace negative pressure, historical furnace outlet temperature, and historical unit steam volume data from the training dataset; during the iteration process, residuals need to be calculated by subtracting the predicted value from the gradient boosting tree from the actual target value, and the difference is the residual to be processed in this round; simultaneously, new decision trees are trained based on historical input variable data or training historical input variable data, the corresponding residuals, and the maximum depth of the decision tree data. Then, the predicted value of the gradient boosting tree model is added to the learning rate data and multiplied by the predicted value of the new decision tree to obtain a new predicted value. After repeating the process for the number of decision trees, the updated gradient boosting tree model is obtained.

[0107] Please refer to Figures 4 and 5. Figure 4 is a flowchart illustrating the intelligent optimization method for the operation of a waste incineration device according to an embodiment of this application, which involves obtaining optimized input variables. The step of calculating the importance index of the input variables based on the updated gradient boosting tree model, and comparing the importance index with a preset optimized input variable threshold to obtain optimized input variables, specifically includes:

[0108] S41. Obtain the input variable importance index of each input variable through gradient boosting tree model;

[0109] S42. Calculate the sum of the importance indices of each input variable and record it as the total importance.

[0110] S43. Divide the importance index of each input variable by the sum of importance to obtain the importance percentage of each input variable.

[0111] S44. Sort the important proportion data of the input variables in descending order;

[0112] S45. Select input variables that are greater than the preset importance threshold and mark them as optimized input variables.

[0113] It should be noted that in this embodiment, the initial value of the importance index of each input variable is set to 0. Then, the gradient boosting tree model is traversed in a preset manner to calculate the contribution of each input variable to reducing the error, i.e., the importance index of each input variable. After calculating the proportion of important input variables, a threshold comparison is performed to obtain the optimized input variables, i.e., the input variables that have a greater impact on the prediction results.

[0114] Referring to Figure 5, this embodiment assigns a preset importance threshold to the 21 parameters of the incinerator. Input variables exceeding this threshold are marked as optimized input variables. Referring to Figure 5, this embodiment reduces the 21 variables to 13 variables: primary air flow rate of left unit 1 (Nm³ / h), primary air flow rate of left unit 2 (Nm³ / h), primary air flow rate of left unit 3 (Nm³ / h), primary air flow rate of left unit 4 (Nm³ / h), primary air flow rate of left unit 5 (Nm³ / h), temperature measurement of the left primary air fan suction pipe (°C), secondary air fan frequency (%), feeder speed (mm / s), unit conveying speed (mm / s), unit 2 conveying speed (mm / s), unit 3 conveying speed (mm / s), unit 4 conveying speed (mm / s), and unit 5 conveying speed (mm / s).

[0115] This embodiment first analyzes the parameters in the combustion process using a gradient boosting tree model, selecting key parameters to reduce the number of parameters to be optimized. This reduces the number of input parameters for subsequent multi-layer BP neural networks, thus reducing training load. Furthermore, in subsequent reinforcement learning, the reduced number of parameters to be optimized also helps to reduce the action selection space and the number of training iterations. This embodiment, by using the gradient boosting tree model to obtain key parameters, reduces the number of parameters to be optimized. Moreover, even with a reduced number of parameters, the BP neural network can still effectively predict the combustion process of the combustion device. Because of the nonlinear relationship between the parameters to be optimized during combustion (i.e., the parameters of the incinerator mentioned above) and the output variables (i.e., flue gas oxygen content, furnace negative pressure, furnace outlet temperature, and unit steam quantity), (if dimensionality reduction is not performed, too many parameters to be optimized will require a large amount of computation, which may also lead to getting stuck in local optima during optimization), this embodiment first establishes the contribution of the parameters to be optimized during combustion to the output variables through a gradient boosting tree model, thereby filtering the parameters to be optimized with a larger contribution, thus reducing the number of input variables of the subsequent BP neural network model and the selection space of reinforcement learning actions, which can greatly reduce the training workload. On the other hand, it can also reduce the local optima caused by learning the parameters to be optimized with a small contribution, which would lead to a large selection space of reinforcement learning actions.

[0116] According to an embodiment of the present invention, the step of establishing an initial thermal efficiency prediction model based on the optimized input variable data and output variable data corresponding to the optimized input variables and combining them with a preset multi-layer BP neural network, and obtaining the thermal efficiency prediction model through a preset particle swarm optimization algorithm, specifically includes:

[0117] An initial thermal efficiency prediction model is established by using a pre-set multi-layer BP neural network and combining optimized input and output variable data.

[0118] Obtain the predicted and actual values ​​of the output variables, calculate the mean squared error based on the predicted and actual values ​​of the output variables, and mark it as the fitness value;

[0119] The initial thermal efficiency prediction model is optimized using a preset method based on the fitness value to obtain the thermal efficiency prediction model.

[0120] It should be noted that BP neural networks are a type of feedforward neural network widely used in machine learning. They use the backpropagation algorithm to adjust the connection weights between neurons to minimize the error between the predicted output and the actual output. A multi-layer BP neural network includes an input layer, hidden layers (which can be one or more), and an output layer. The input layer receives external data, which undergoes complex nonlinear transformations through the hidden layers, ultimately yielding the prediction result at the output layer. In this embodiment, the input layer refers to the data corresponding to the optimized input variables, and the output layer refers to the output variable data corresponding to the output variables. The number of neurons and layers in the hidden layers can be adjusted according to actual conditions. Through continuous trial and optimization, the most suitable network structure for modeling the operational data of waste incineration equipment is found. Particle swarm optimization is a swarm intelligence-based optimization algorithm. In this embodiment, the fitness value is obtained by calculating the mean squared error (MSE), which is a metric for measuring the "squared average error." It is used to evaluate the performance of the prediction model. It calculates the average of the squares of the difference (error) between the predicted value and the true value. The smaller the fitness value, the better the corresponding neural network parameters. The preset method refers to the PSO update mechanism and the particle optimal value update method commonly used in particle swarm optimization algorithm. The thermal efficiency prediction model is obtained after the mean square error reaches the set accuracy target.

[0121] Referring to Figure 7, which shows the fitness value of the particles in this embodiment, the fitness value of the particles drops to the minimum after 200 iterations, at which point the BP neural parameters are optimal.

[0122] According to an embodiment of the present invention, the step of processing the thermal efficiency prediction model using a preset reinforcement learning algorithm to obtain the optimal input variable feature data corresponding to the optimized input variable specifically includes:

[0123] S51. Initialize the input variables to be optimized; in this embodiment, the input variables to be optimized are the above 13 incinerator combustion parameters.

[0124] S52. The agent obtains the predicted output variable based on the input variable to be optimized; in this embodiment, the agent is the thermal efficiency prediction model.

[0125] S53. The output variable reward value and the sum of the output variable reward values ​​are obtained by processing each output variable data through a preset reward function; the reinforcement learning agent is the thermal efficiency prediction model.

[0126] S54. Adjust the optimized input variable data according to the sum of the output variable reward values, and then repeat steps S51-S53 until the optimal input variable feature data is obtained.

[0127] It should be noted that the requirements for output variables are generally relatively fixed. In this embodiment, the standard data for output variables include standard data for flue gas oxygen content (data variation range), standard data for furnace negative pressure (data variation range), standard data for furnace outlet temperature (data variation range), and standard data for unit steam volume (data variation range). The standard data for flue gas oxygen content is [3,8], the standard data for furnace negative pressure is [20,70], the standard data for furnace outlet temperature is [950,1150], and the standard data for unit steam volume refers to the amount of steam produced after burning a unit weight of waste. The standard data for unit steam volume is [1.5,3.5]. When the flue gas oxygen content data, furnace negative pressure data, and furnace outlet temperature data are within the standard data range, the reward value is -2, indicating that the current state is acceptable but there is still room for improvement. Otherwise, the reward value is 0, indicating that the current state deviates significantly from the ideal range and needs adjustment. The reward value for unit steam volume data needs to be calculated based on its average value and standard deviation using a preset method, as detailed below.

[0128] The reward values ​​corresponding to the flue gas oxygen content data, furnace negative pressure data, furnace outlet temperature data, and unit steam quantity data are added together to obtain the total reward value of the output variable. When the total reward value no longer changes, the optimal solution can be considered to have been obtained. By continuously trying different combinations of input variable data, the BP neural network algorithm gradually learns the strategy that can maximize the reward value, that is, the optimal input variable feature data corresponding to the optimal input variable.

[0129] Referring to Figure 6, in this embodiment, the total reward value does not change after 60 iterations, so the iteration can be terminated at the 80th iteration to obtain the optimal input variable feature data corresponding to the termination of the iteration.

[0130] This embodiment aims to maximize the unit steam output. Therefore, it uses the first three output variables to calculate the reward value, with the reward value ranging as follows:

[0131] 1. Oxygen content (%) in boiler outlet flue gas varies from 3 to 8%.

[0132] 2. First channel flue gas outlet pressure (furnace outlet negative pressure (Pa) variation range: 20-70)

[0133] 3. Average temperature in the upper part of the furnace (furnace temperature) (°C) range: 950-1150. If the value of the output variable is within the given upper and lower limits: reward = -2, otherwise reward = 0.

[0134] The fourth variable (steam production) should be maximized to achieve the optimization objective.

[0135] According to professional heat balance calculations: the maximum steam output is 3.5, and the minimum is 1.5; the reward values ​​are as follows:

[0136] Reward consists of four parts, reward 1, reward 2, reward 3, reward 4Reward_tota l=reward 1+reward 2+reward 3+reward 4

[0137] Reward4 = -norm(Z4 - 3.5); where Z4 is the predicted value of the fourth variable (steam production), that is, the predicted value output by the agent based on the input parameters;

[0138] When Z4 < (1.5 - 2.0871) / 0.1574, Reward_total = -10; when Z4 = (3 - 2.0871) /

[0139] At 0.1574, Reward_total = 0.

[0140] (Z4-2.0871) / 0.1574mean=2.0871,std=0.1574

[0141] All calculations are performed using normalized values: if the predicted value for the fourth term is Z4, then input (Z4 - 2.0871) / 0.1574.

[0142] goa l State= (3-2.0871) / 0.1574=5.8;

[0143] reward4 = -norm((Z4 - 2.0871) / 0.1574 - goal State);

[0144] Minimum value of reward4: -norm((1.5 - 2.0871) / 0.1574 - goal State) = -norm(-3.72 -

[0145] 5.8) = -9.53

[0146] Maximum value of reward4: -norm((3 - 2.0871) / 0.1574 - goal State) = 0

[0147] Oxygen content of flue gas at furnace outlet, reward 1:

[0148] When Z1 (normalized value) > Z1 max (normalized value) or Z1 (normalized value) < Z1 min (normalized value),

[0149] Reward 1 = -10; where Z1 is the predicted value of the oxygen content of flue gas at the furnace outlet;

[0150] When Z1 min (normalized value) < Z1 (normalized value) < Z1 max (normalized value), Reward1 = 0.

[0151] reward 2:

[0152] When Z2 (normalized value) > Z2 max (normalized value) or Z2 < Z2 min (normalized value), Reward 2 = -10;

[0153] When Z2 min (normalized value) < Z2 (normalized value) < Z2 max (normalized value), Reward2 = 0. Where Z2 is the predicted value of the flue gas outlet pressure of the first channel;

[0154] reward 3:

[0155] When Z3 (normalized value) > Z3 max (normalized value) or Z3 < Z3 min (normalized value), Reward 3 = -10;

[0156] When Z3 min (normalized value) < Z3 (normalized value) < Z3 max (normalized value), Reward3 = 0. Where Z3 is the predicted value of the furnace temperature;

[0157] Specifically, based on the oxygen content of flue gas at the furnace outlet, negative pressure at the furnace outlet, furnace temperature, and steam output of the intelligent agent, the corresponding reward values ​​Reward 1, Reward 2, Reward 3, and Reward 4 are calculated respectively. The reward values ​​Reward 1, Reward 2, Reward 3, and Reward 4 are summed to obtain the total reward value Reward total.

[0158] Specifically, Reward4 is calculated as follows: when Z4 < (1.5 - 2.0871) / 0.1574, Reward_total = -10; when Z4 = (3 - 2.0871) / 0.1574, Reward_total = 0; otherwise, national State = (3 - 2.0871) / 0.1574 = 5.8; reward4 = -norm((Z4 - 2.0871) / 0.1574 - national State).

[0159] In this embodiment, the calculation of the reward value considers both steam production and other output variables (i.e., oxygen content in the flue gas at the furnace outlet, negative pressure at the furnace outlet, and furnace temperature) as the total reward value. This ensures that the incinerator operates normally while maximizing its output, reducing the risk of it malfunctioning due to reinforcement learning prioritizing steam production for the reward value. Furthermore, in this embodiment, when the predicted steam production falls outside the preset range, the total reward value is directly set to 0, indicating the need for optimization.

[0160] Using the method described in this embodiment, the final optimized variables are:

[0161] Optimized values ​​for input variables:

[0162] 1. Primary air flow rate (Nm3 / h) of Unit 1 on the left side = 770.35. Reasonable variation range: 500-1800.

[0163] 2. The primary air flow rate (Nm3 / h) of Unit 2 on the left side is 540.8. The reasonable variation range is 500-1500.

[0164] 3. Primary air flow rate (Nm3 / h) of Unit 3 on the left side = 7160.24. Reasonable variation range: 3000-10000

[0165] 4. Primary air flow rate (Nm3 / h) of Unit 4 on the left side = 4322.86. Reasonable variation range: 2000-10000

[0166] 5. Primary air flow rate (Nm3 / h) of Unit 5 on the left side = 2556.49. Reasonable variation range: 2000-5000.

[0167] 6. Temperature measurement of the left primary air fan suction pipe (°C) = 170.69. Reasonable variation range: 150-230.

[0168] 7. Secondary air fan frequency (%) = 47.59. Reasonable variation range: 30-86.

[0169] 8. Feeder speed (mm / s) = 0.74. Reasonable variation range: 0.3-0.8.

[0170] 9.1 Unit conveying speed (mm / s) = 0.75. Reasonable variation range: 0.5----2.

[0171] 10.2 Unit conveying speed (mm / s) = 1.56 Reasonable variation range: 0.7-2.5

[0172] 11.3 Unit conveying speed (mm / s) = 1.38 Reasonable variation range: 0.3-1.5

[0173] 12.4 Unit conveying speed (mm / s) = 1.17 Reasonable variation range: 0.3-1.2

[0174] 13.5 Unit Conveying Speed ​​(mm / s) = 0.43 Reasonable Variation Range: 0.35----1.3

[0175] (2) Optimal values ​​of output variables: 1. Oxygen content in boiler outlet flue gas (%) = 6.92. Reasonable variation range: 3-8

[0176] 2. First channel flue gas outlet pressure (furnace outlet negative pressure, Pa) = 68.36. Reasonable variation range: 20-70.

[0177] 3. Average temperature in the upper part of the furnace (furnace temperature) (°C) = 1086.86. Reasonable range of variation: 950-1150.

[0178] 4. Unit waste gas production = 2.41

[0179] The current average value in DCS data is 2.01, representing an improvement of approximately 20%.

[0180] This invention also discloses an intelligent optimization system for the operation of a waste incineration device, comprising a memory and a processor. The memory stores a program for an intelligent optimization method for the operation of the waste incineration device. When the processor executes the program for the intelligent optimization method for the operation of the waste incineration device, it performs the following steps:

[0181] Obtain historical operational data of a preset number of waste incineration devices, including historical input variable data and historical output variable data. Establish a correspondence between the historical input variable data and the corresponding historical output variable data and classify them to obtain training datasets and test datasets.

[0182] Obtain the key parameters of the preset gradient boosting tree, and train the gradient boosting tree based on the training dataset, test dataset, and key parameters to obtain the updated gradient boosting tree model;

[0183] The importance index of the input variable is calculated based on the updated gradient boosting tree model, and the importance index of the input variable is compared with the preset optimized input variable threshold to obtain the optimized input variable;

[0184] An initial thermal efficiency prediction model is established by combining the optimized input variable data and output variable data with a preset multi-layer BP neural network, and the thermal efficiency prediction model is obtained by processing with a preset particle swarm optimization algorithm.

[0185] The thermal efficiency prediction model is processed by a preset reinforcement learning algorithm to obtain the optimal input variable feature data corresponding to the optimized input variables.

[0186] It should be noted that the operation of waste incineration equipment is a relatively complex process involving many factors. Therefore, a neural network model is required in the data processing. First, a predetermined amount of historical operational data from the waste incineration equipment must be obtained, including historical input and output variable data. There are multiple historical input and output variables, and a correspondence exists between them. To better utilize this historical data, classification is performed to obtain training and testing datasets. Key parameters for a predetermined gradient boosting tree are then obtained, and the model is trained to obtain an updated gradient boosting tree model. Some input variables may have insignificant effects; therefore, the importance index of the input variables needs to be calculated by updating the gradient boosting tree model. Then, the optimized input variables are obtained by comparing the threshold of the input variable importance index. The optimized input variables contain multiple influencing factors, but the number of influencing factors is less than the number of influencing factors in the input variable data. Then, a thermal efficiency prediction model is obtained through a predetermined multi-layer backpropagation neural network and a predetermined particle swarm optimization algorithm. Finally, a predetermined reinforcement learning algorithm is used to process the thermal efficiency prediction model to obtain the optimal input variable feature data, thereby clarifying the adjustment target of the input variables.

[0187] According to an embodiment of the present invention, the acquisition of historical operating data of a preset number of waste incineration devices includes historical input variable data and historical output variable data. Establishing a correspondence between the historical input variable data and the corresponding historical output variable data, and classifying them to obtain training and testing datasets, specifically includes:

[0188] Establish a correspondence between historical input variable data and corresponding historical output variable data, and group them according to a preset ratio to obtain training dataset and test dataset;

[0189] The training dataset includes training history input variable data and corresponding training history output variable data;

[0190] The test dataset includes historical test input variable data and corresponding historical test output variable data.

[0191] It should be noted that historical input variable data is the independent variable, and historical output variable data is the dependent variable; therefore, there is a corresponding relationship between the two. In this embodiment, the historical input variable data includes historical feeder speed data, historical main steam flow rate data, historical main steam temperature data, historical main steam pressure data, historical furnace temperature data, historical feedwater temperature data, historical feedwater flow rate data, historical primary air temperature data, historical induced draft fan frequency data, historical secondary air volume data, historical flue gas temperature data, historical primary air volume data of the first grate, historical primary air volume data of the second grate, and historical primary air volume data of the third grate. The historical primary air volume data for the 4th and 5th grate sections, the historical operating speed data for the 1st, 2nd, 3rd, 4th, and 5th grate sections, and the historical output variable data include historical flue gas oxygen content data, historical furnace negative pressure data, historical furnace outlet temperature data, and historical unit steam volume data. The preset ratio can be set according to user needs. In this embodiment, the ratio of the corresponding data quantity in the training dataset to the corresponding data quantity in the test dataset is 7:3.

[0192] According to an embodiment of the present invention, obtaining the key parameters of a preset gradient boosting tree and training the gradient boosting tree based on the training dataset, the test dataset, and the key parameters to obtain an updated gradient boosting tree model specifically includes:

[0193] Obtain the key parameters of the preset gradient boosting tree, including the number of decision trees, learning rate data, and maximum depth data of the decision trees;

[0194] Obtain the average value of the historical output variable data during training and mark it as the initial value for the model;

[0195] The model is iterated based on the initial values, historical input variables, and corresponding historical output variables, combined with the number of decision trees. The model is then tested using historical input variables and corresponding historical output variables to obtain an updated gradient boosting tree model.

[0196] It's important to note that the number of decision trees refers to the number of decision trees to be generated, which is the number of iterations required in the entire training process; the learning rate data controls the impact of each newly generated decision tree on the previously inaccurate predictions (residuals); the maximum depth of the decision tree data indicates the maximum depth a decision tree can reach; the average of the historical output variable data is calculated by taking the average of historical flue gas oxygen content, historical furnace negative pressure, historical furnace outlet temperature, and historical unit steam volume data from the training dataset; during the iteration process, residuals need to be calculated by subtracting the predicted value from the gradient boosting tree from the actual target value, and the difference is the residual to be processed in this round; simultaneously, new decision trees are trained based on historical input variable data or training historical input variable data, the corresponding residuals, and the maximum depth of the decision tree data. Then, the predicted value of the gradient boosting tree model is added to the learning rate data and multiplied by the predicted value of the new decision tree to obtain a new predicted value. After repeating the process for the number of decision trees, the updated gradient boosting tree model is obtained.

[0197] The step of calculating the importance index of the input variable based on the updated gradient boosting tree model, and comparing the importance index of the input variable with a preset optimization input variable threshold to obtain the optimized input variable, specifically includes:

[0198] The importance index of each input variable is obtained by processing it using a gradient boosting tree model;

[0199] Calculate the sum of the importance indices of each input variable and record it as the total importance.

[0200] The importance percentage of each input variable is obtained by dividing the importance index of each input variable by the sum of the importance indices.

[0201] Sort the important percentage data of the input variables in descending order;

[0202] Select input variables that are greater than a preset importance threshold and mark them as optimized input variables.

[0203] It should be noted that in this embodiment, the initial value of the importance index of each input variable is set to 0. Then, the gradient boosting tree model is traversed in a preset manner to calculate the contribution of each input variable to reducing the error, i.e., the importance index of each input variable. After calculating the proportion of important input variables, a threshold comparison is performed to obtain the optimized input variables, i.e., the input variables that have a greater impact on the prediction results.

[0204] According to an embodiment of the present invention, the step of establishing an initial thermal efficiency prediction model based on the optimized input variable data and output variable data corresponding to the optimized input variables and combining them with a preset multi-layer BP neural network, and obtaining the thermal efficiency prediction model through a preset particle swarm optimization algorithm, specifically includes:

[0205] An initial thermal efficiency prediction model is established by using a pre-set multi-layer BP neural network and combining optimized input and output variable data.

[0206] Obtain the predicted and actual values ​​of the output variables, calculate the mean squared error based on the predicted and actual values ​​of the output variables, and mark it as the fitness value;

[0207] The initial thermal efficiency prediction model is optimized using a preset method based on the fitness value to obtain the thermal efficiency prediction model.

[0208] It should be noted that BP neural networks are a type of feedforward neural network widely used in machine learning. They use the backpropagation algorithm to adjust the connection weights between neurons to minimize the error between the predicted output and the actual output. A multi-layer BP neural network includes an input layer, hidden layers (which can be one or more), and an output layer. The input layer receives external data, which undergoes complex nonlinear transformations through the hidden layers, ultimately yielding the prediction result at the output layer. In this embodiment, the input layer refers to the data corresponding to the optimized input variables, and the output layer refers to the output variable data corresponding to the output variables. The number of neurons and layers in the hidden layers can be adjusted according to actual conditions. Through continuous trial and optimization, the most suitable network structure for modeling the operational data of waste incineration equipment is found. Particle swarm optimization is a swarm intelligence-based optimization algorithm. In this embodiment, the fitness value is obtained by calculating the mean squared error (MSE), which is a metric for measuring the "squared average error." It is used to evaluate the performance of the prediction model. It calculates the average of the squares of the difference (error) between the predicted value and the true value. The smaller the fitness value, the better the corresponding neural network parameters. The preset method refers to the PSO update mechanism and the particle optimal value update method commonly used in particle swarm optimization algorithm. The thermal efficiency prediction model is obtained after the mean square error reaches the set accuracy target.

[0209] According to an embodiment of the present invention, the step of processing the thermal efficiency prediction model using a preset reinforcement learning algorithm to obtain the optimal input variable feature data corresponding to the optimized input variable specifically includes:

[0210] Obtain standard data and real-time data of output variables;

[0211] The output variable reward value and the sum of the output variable reward values ​​are obtained by processing each output variable data using a preset reward function.

[0212] The thermal efficiency prediction model adjusts the output variable reward value and the optimized input variable data corresponding to the sum of the output variable reward values ​​in a preset manner to obtain the optimal input variable feature data.

[0213] It should be noted that the requirements for output variables are generally relatively fixed. In this embodiment, the standard data for output variables include standard data for flue gas oxygen content, furnace negative pressure, furnace outlet temperature, and unit steam volume. The standard data for flue gas oxygen content is [3,8], the standard data for furnace negative pressure is [20,70], the standard data for furnace outlet temperature is [950,1150], and the standard data for unit steam volume refers to the amount of steam produced after burning a unit weight of waste. The standard data for unit steam volume is [1.5,3.5]. When the flue gas oxygen content, furnace negative pressure, and furnace outlet temperature are within the standard data range, the reward value is... A reward of -2 indicates that the current state is acceptable, but there is still room for improvement; otherwise, a reward of 0 indicates that the current state deviates significantly from the ideal range and needs adjustment. The reward value for the unit steam volume data is calculated using a preset method based on the reward values ​​for the flue gas oxygen content data, furnace negative pressure data, and furnace outlet temperature data. The reward values ​​corresponding to the flue gas oxygen content data, furnace negative pressure data, furnace outlet temperature data, and unit steam volume data are added together to obtain the sum of the output variable reward values. By continuously trying different combinations of input variable data, the BP neural network algorithm gradually learns the strategy that maximizes the reward value, that is, the optimal input variable feature data corresponding to the optimal input variable.

[0214] A third aspect of the present invention provides a readable storage medium comprising a program for an intelligent optimization method for operating a waste incineration device, wherein when the program is executed by a processor, it implements the steps of the intelligent optimization method for operating a waste incineration device as described in any of the preceding claims.

[0215] This invention discloses an intelligent optimization method, system, and medium for the operation of waste incineration equipment. It acquires a preset number of historical input and output variable data from the waste incineration equipment, then classifies the historical input and output variable data to obtain training and test datasets. Key parameters for a gradient boosting tree are obtained and trained to acquire an updated gradient boosting tree model. Based on the updated gradient boosting tree model, an input variable importance index is calculated, and optimized input variables are obtained through threshold comparison. A thermal efficiency prediction model is obtained through multi-layer backpropagation neural networks and particle swarm optimization algorithms. The thermal efficiency prediction model is then processed by a reinforcement learning algorithm to obtain optimal input variable feature data. Thus, intelligent optimization of the waste incineration equipment is achieved through gradient boosting tree models, multi-layer backpropagation neural networks, and particle swarm optimization algorithms, reducing the power generation cost of the waste incineration equipment.

[0216] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0217] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0218] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0219] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0220] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

Claims

1. A method for intelligent optimization of the operation of a waste incineration device, characterized in that, include: Obtain historical operational data of a preset number of waste incineration devices, including historical input variable data and historical output variable data. Establish a correspondence between the historical input variable data and the corresponding historical output variable data and classify them to obtain training datasets and test datasets. Obtain the key parameters of the preset gradient boosting tree, and train the gradient boosting tree based on the training dataset, test dataset, and key parameters to obtain the updated gradient boosting tree model; The importance index of the input variable is calculated based on the updated gradient boosting tree model, and the importance index of the input variable is compared with the preset optimized input variable threshold to obtain the optimized input variable; An initial thermal efficiency prediction model is established by combining the optimized input variable data and output variable data with a preset multi-layer BP neural network, and the thermal efficiency prediction model is obtained by processing with a preset particle swarm optimization algorithm. The thermal efficiency prediction model is processed by a preset reinforcement learning algorithm to obtain the optimal input variable feature data corresponding to the optimized input variables.

2. The intelligent optimization method for the operation of waste incineration equipment according to claim 1, characterized in that, The acquisition of historical operational data for a preset number of waste incineration devices includes historical input variable data and historical output variable data. A correspondence is established between the historical input variable data and the corresponding historical output variable data, and the data is classified to obtain training and testing datasets. Specifically, this includes: Establish a correspondence between historical input variable data and corresponding historical output variable data, and group them according to a preset ratio to obtain training dataset and test dataset; The training dataset includes training history input variable data and corresponding training history output variable data; The test dataset includes historical test input variable data and corresponding historical test output variable data.

3. The intelligent optimization method for the operation of waste incineration equipment according to claim 2, characterized in that, The process of obtaining the key parameters of the preset gradient boosting tree and training the gradient boosting tree based on the training dataset, test dataset, and key parameters to obtain the updated gradient boosting tree model specifically includes: Obtain the key parameters of the preset gradient boosting tree, including the number of decision trees, learning rate data, and maximum depth data of the decision trees; Obtain the average value of the historical output variable data during training and mark it as the initial value for the model; The model is iterated based on the initial values, historical input variables, and corresponding historical output variables, combined with the number of decision trees. The model is then tested using historical input variables and corresponding historical output variables to obtain an updated gradient boosting tree model.

4. The intelligent optimization method for the operation of waste incineration equipment according to claim 3, characterized in that, The step of calculating the importance index of the input variable based on the updated gradient boosting tree model, and comparing the importance index of the input variable with a preset optimization input variable threshold to obtain the optimized input variable, specifically includes: The importance index of each input variable is obtained by processing it using a gradient boosting tree model; Calculate the sum of the importance indices of each input variable and record it as the total importance. The importance percentage of each input variable is obtained by dividing the importance index of each input variable by the sum of the importance indices. Sort the important percentage data of the input variables in descending order; Select input variables that are greater than a preset importance threshold and mark them as optimized input variables.

5. The intelligent optimization method for the operation of waste incineration equipment according to claim 4, characterized in that, The process of establishing an initial thermal efficiency prediction model based on the optimized input variable data and output variable data corresponding to the optimized input variables, combined with a preset multi-layer BP neural network, and obtaining the thermal efficiency prediction model through a preset particle swarm optimization algorithm, specifically includes: An initial thermal efficiency prediction model is established by using a pre-set multi-layer BP neural network and combining optimized input and output variable data. Obtain the predicted and actual values ​​of the output variables, calculate the mean squared error based on the predicted and actual values ​​of the output variables, and mark it as the fitness value; The initial thermal efficiency prediction model is optimized using a preset method based on the fitness value to obtain the thermal efficiency prediction model.

6. The intelligent optimization method for the operation of waste incineration equipment according to claim 5, characterized in that, The step of processing the thermal efficiency prediction model using a preset reinforcement learning algorithm to obtain the optimal input variable feature data corresponding to the optimized input variables specifically includes: S51. Initialize the input variables to be optimized; S52. The agent obtains the predicted output variable based on the input variable to be optimized; S53. The output variable reward value and the sum of the output variable reward values ​​are obtained by processing each output variable data through a preset reward function; the reinforcement learning agent is the thermal efficiency prediction model. S54. Adjust the optimized input variable data according to the sum of the output variable reward values, and then repeat steps S51-S53 until the sum of the reward values ​​no longer changes. At this point, the optimal input variable feature data is obtained.

7. An intelligent optimization system for the operation of a waste incineration device, characterized in that, It includes a memory and a processor. The memory includes a smart optimization method program for the operation of a waste incineration device. When the processor executes the smart optimization method program for the operation of the waste incineration device, it performs the following steps: Obtain historical operational data of a preset number of waste incineration devices, including historical input variable data and historical output variable data. Establish a correspondence between the historical input variable data and the corresponding historical output variable data and classify them to obtain training datasets and test datasets. Obtain the key parameters of the preset gradient boosting tree, and train the gradient boosting tree based on the training dataset, test dataset, and key parameters to obtain the updated gradient boosting tree model; The importance index of the input variable is calculated based on the updated gradient boosting tree model, and the importance index of the input variable is compared with the preset optimized input variable threshold to obtain the optimized input variable; An initial thermal efficiency prediction model is established by combining the optimized input variable data and output variable data with a preset multi-layer BP neural network, and the thermal efficiency prediction model is obtained by processing with a preset particle swarm optimization algorithm. The thermal efficiency prediction model is processed by a preset reinforcement learning algorithm to obtain the optimal input variable feature data corresponding to the optimized input variables.

8. The intelligent optimization system for the operation of waste incineration equipment according to claim 7, characterized in that, The acquisition of historical operational data for a preset number of waste incineration devices includes historical input variable data and historical output variable data. A correspondence is established between the historical input variable data and the corresponding historical output variable data, and the data is classified to obtain training and testing datasets. Specifically, this includes: Establish a correspondence between historical input variable data and corresponding historical output variable data, and group them according to a preset ratio to obtain training dataset and test dataset; The training dataset includes training history input variable data and corresponding training history output variable data; The test dataset includes historical test input variable data and corresponding historical test output variable data.

9. The intelligent optimization system for the operation of waste incineration equipment according to claim 8, characterized in that, The process of obtaining the key parameters of the preset gradient boosting tree and training the gradient boosting tree based on the training dataset, test dataset, and key parameters to obtain the updated gradient boosting tree model specifically includes: Obtain the key parameters of the preset gradient boosting tree, including the number of decision trees, learning rate data, and maximum depth data of the decision trees; Obtain the average value of the historical output variable data during training and mark it as the initial value for the model; The model is iterated based on the initial values, historical input variables, and corresponding historical output variables, combined with the number of decision trees. The model is then tested using historical input variables and corresponding historical output variables to obtain an updated gradient boosting tree model.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a smart optimization method program for the operation of a waste incineration device. When the smart optimization method program for the operation of the waste incineration device is executed by a processor, it implements the steps of the smart optimization method for the operation of the waste incineration device as described in any one of claims 1 to 6.

Images