Multi-parameter dynamic optimization industrial dust control method
By using the Transformer-BP network to predict dust concentration in dust control and combining genetic algorithms to optimize air volume control parameters, the problem of insufficient adaptability of dust concentration prediction and control in complex mine environments is solved, high-precision prediction and optimization control are achieved, and safety and efficiency are improved.
Patent Information
- Application Number
- CN202510357069.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-25
- Publication Date
- 2025-06-27
AI Technical Summary
The existing dust control methods are difficult to achieve accurate dust concentration prediction and optimization control in complex mine environments, especially under dynamic operating conditions. The traditional methods are insufficiently adaptable, resulting in large deviations in dust concentration prediction and unstable air volume control effect.
The dust concentration prediction model based on the Transformer-BP network is adopted, and the air volume control parameters are dynamically adjusted in combination with the genetic algorithm to form an industrial dust control method with multi-parameter dynamic optimization. This method extracts timing features through Transformer, performs feature modeling by BP neural network, and optimizes air volume control parameters by genetic algorithm to achieve high-precision prediction and optimization control of dust concentration.
It improves the accuracy and adaptability of dust concentration prediction, enhances the stability and efficiency of air volume control, reduces energy consumption, improves the intelligence level of mine ventilation and dust removal system, and ensures the safety of mine operations.
Smart Images
Figure CN120215374A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of dust control and relates to an industrial dust control method with multi-parameter dynamic optimization. Background Art
[0002] In complex working conditions such as mine roadways, the accurate prediction and efficient control of dust concentration have always been a key challenge. Existing dust control methods mainly rely on fixed parameter configurations and traditional empirical formulas, but these methods have significant limitations in practical applications. For example, fixed parameters are difficult to adapt to dynamically changing working conditions, resulting in large deviations in dust concentration prediction and affecting the regulation effect of the ventilation and dust removal system. In addition, traditional prediction models are difficult to fully capture the complex relationships between various input variables, especially under the action of multi-parameter non-linearity, there are large errors in prediction accuracy. On the other hand, most existing air volume control strategies are optimized in a single target direction, lacking comprehensive consideration of multiple factors such as energy consumption and dust removal efficiency, and it is difficult to achieve intelligent and adaptive precise regulation.
[0003] These drawbacks lead to poor performance of current dust concentration prediction and control methods in complex mine environments, often unable to effectively reduce dust concentration and provide a safe working environment for mine workers. Especially in the face of dynamically changing environmental conditions, traditional methods are difficult to adjust air volume control parameters in a timely manner, resulting in poor dust removal effect and further increasing the safety risks of mine operations. Therefore, there is an urgent need for a method that can achieve accurate dust concentration prediction and optimal control in complex and variable environments to improve the efficiency and safety of the mine ventilation and dust removal system. Summary of the Invention
[0004] In view of this, the purpose of the present invention is to provide an industrial dust control method with multi-parameter dynamic optimization.
[0005] To achieve the above purpose, the present invention provides the following technical solutions:
[0006] An industrial dust control method with multi-parameter dynamic optimization, comprising the following steps:
[0007] S1: Collect data related to dust prediction in the mine roadway, perform preprocessing, and form a data set;
[0008] S2: Input the data set into a dust concentration prediction model based on the Transformer-BP network to predict the dust concentration;
[0009] S3: Use the genetic algorithm GA to dynamically adjust the air volume control parameters to minimize the dust concentration;
[0010] S4: Output the optimized air volume control parameters to control the dust concentration in the roadway.
[0011] Furthermore, the dust prediction related data includes the roadway cross-sectional area, the forced air volume, the extracted air volume, the distance from the dust extraction port to the heading face, the length of the dust control device, the distance from the dust control device to the heading face, and the ratio of radial to axial air volume.
[0012] Furthermore, the preprocessing includes the following steps:
[0013] Sort out the data and convert the data into a CSV file;
[0014] Divide the input and output and perform data normalization;
[0015] Perform data cleaning, feature engineering, and data augmentation on the normalized data to obtain the final dataset.
[0016] Furthermore, the dust concentration prediction model based on the Transformer-BP network includes a Transformer module and a BP neural network;
[0017] The Transformer module processes historical dust concentration data through the self-attention mechanism, extracts long-term dependence relationships, enhances the prediction ability, and makes the model more adaptable to environmental changes;
[0018] The BP neural network uses three hidden layers, with 64, 32, and 16 neurons respectively. Each layer is followed by a ReLU activation function to introduce non-linear characteristics; Dropout regularization is adopted to randomly discard some neurons.
[0019] Furthermore, in step S2, first use Transformer for time series feature extraction, and capture the long-term dependence relationship of historical data through the self-attention mechanism;
[0020] Define the BP neural network to predict values of A and B;
[0021] Initialize the Transformer-BP network model and the optimizer;
[0022] Set the maximum number of iterations, perform iterative training on the Transformer-BP network model, and verify the loss.
[0023] Furthermore, the genetic algorithm GA is used to optimize the weight initialization of the BP network, improve the training stability, and is also used to optimize the air volume control parameters, including the ratio of radial to axial air volume, the extracted air volume, and the distance of the dust control device, dynamically adjust the air volume control strategy to minimize the dust concentration and maximize the dust removal efficiency.
[0024] Furthermore, the genetic algorithm GA includes the following steps:
[0025] Initialize the genetic algorithm optimizer;
[0026] The genetic algorithm is used to evolve parameters and select the optimal individual;
[0027] Define the objective function for calculating the C value;
[0028] Dynamically adjust the air volume control parameters;
[0029] The genetic algorithm evolves and searches for the optimal C value and parameters;
[0030] Combined with the prediction results of the Transformer - BP network model, evaluate the parameters optimized by GA;
[0031] Output the optimal C value and control parameters optimized by GA.
[0032] The beneficial effects of the present invention are as follows:
[0033] 1. By introducing a combined model of BP neural network and Transformer, processing multi - parameter inputs such as roadway cross - sectional area, radial - axial air volume ratio, and extracted air volume, using Transformer to extract time - series features, enhancing the modeling ability of the dust concentration change trend, achieving high - precision prediction in complex mine environments, and effectively solving the problem of insufficient adaptability of traditional methods under dynamic working conditions;
[0034] 2. Using the genetic algorithm (GA) to globally optimize parameters such as radial - axial air volume ratio, extracted air volume, and the distance of dust control devices, ensuring that the optimal parameter combination can be quickly found, improving the stability of dust control, reducing energy consumption at the same time, and enhancing the adaptability of the model in complex environments;
[0035] 3. Construct a dust concentration control system integrating data acquisition, feature expansion, Transformer - BP prediction, GA parameter optimization, and dynamic air volume control, enhancing the intelligent level of mine ventilation and dust treatment, ensuring safety, and reducing operating costs.
[0036] Other advantages, objectives, and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on the study of the following text, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following specification. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be described in detail preferably with reference to the accompanying drawings, where:
[0038] Figure 1 It is the schematic diagram of the BP neural network model;
[0039] Figure 2Flowchart of an industrial dust control method for multi-parameter dynamic optimization. Detailed implementation mode
[0040] The following illustrates the implementation mode of the present invention through specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific implementation modes. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.
[0041] It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner. Therefore, only the components related to the present invention are shown in the diagrams, rather than being drawn according to the number, shape, and size of the components in actual implementation. The type, quantity, and proportion of each component in actual implementation can be arbitrarily changed, and the component layout type may also be more complex.
[0042] In the following description, a large number of details are explored to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.
[0043] Embodiment 1:
[0044] As Figure 1 shown, the present invention provides an industrial dust control method for multi-parameter dynamic optimization, which globally optimizes the BP neural network through GA and captures temporal features by combining Transformer to achieve more accurate dust concentration prediction and intelligent air volume regulation. The dust concentration in the mine roadway is greatly affected by factors such as air volume ratio, extraction air volume, and distance of dust control devices. Due to fixed parameters, traditional control methods are difficult to adaptively adjust, resulting in insufficient prediction accuracy and unstable air volume control effect. To address this problem, this technical solution uses GA to optimize the BP network parameters to improve the model training effect. At the same time, Transformer is used to extract the deep temporal relationship in historical environmental data to construct an adaptive dust concentration prediction and optimization control method, improving the accuracy and efficiency of mine dust governance.
[0045] As Figure 1As shown in the figure, this model adopts a hybrid architecture that combines a BP neural network, genetic algorithm (GA) optimization, and Transformer enhanced time series modeling. The input parameters of the model are seven, namely the roadway cross-sectional area, forced air volume, extracted air volume, distance from the dust extraction port to the heading face, length of the dust control device, distance from the dust control device to the heading face, and radial to axial air volume ratio.
[0046] The model outputs the dust concentration values at six positions, namely the concentrations at 0 meters, 5 meters, 10 meters, 15 meters, 20 meters, and 25 meters from the heading face.
[0047] The specific structure includes:
[0048] Input layer: It adopts multi-feature input, including the roadway cross-sectional area, forced air volume, extracted air volume, etc., and enhances the model's ability to model complex non-linear relationships through feature expansion (square terms, interaction features).
[0049] Transformer module: It processes historical dust concentration data through the self-attention mechanism to extract long-term dependence relationships, enhance the prediction ability, and make the model more adaptable to environmental changes.
[0050] BP neural network: It adopts three hidden layers with 64, 32, and 16 neurons respectively. Each layer is followed by a ReLU activation function to introduce non-linear characteristics and enhance the model's learning ability. Dropout regularization is used to randomly discard some neurons to prevent overfitting and improve the generalization ability.
[0051] GA for parameter optimization: The genetic algorithm (GA) is used to optimize the weight initialization of the BP network, improve the training stability, and avoid falling into local optima. GA can also be used to optimize the air volume control parameters to maximize the dust removal efficiency and reduce the system energy consumption.
[0052] Output layer: It predicts the dust concentration values at different positions and provides more accurate dust removal control suggestions in combination with the optimized air volume control parameters.
[0053] Example 2:
[0054] As Figure 2 shown, the working process of this method is as follows:
[0055] First, use a deep learning model to predict the dust concentration in the mine roadway. By inputting parameters such as the roadway cross-sectional area, radial to axial air volume ratio, extracted air volume, and distance of the dust control device, the Transformer is used for time series feature extraction to capture the long-term dependence relationships of historical data, and combined with the BP neural network for feature modeling to enhance the model's prediction ability and generalization ability.
[0056] Then, the control parameters such as the radial-axial air volume ratio, the extraction air volume, and the distance of the dust control device are globally optimized by the genetic algorithm (GA), and the air volume control strategy is dynamically adjusted to minimize the dust concentration and improve the dust removal effect. In the application, this method preprocesses the input data, then uses the Transformer-BP combined model to predict the dust concentration, and then uses GA to find the optimal control parameter combination.
[0057] Finally, the optimized air volume control parameters are output to effectively control the dust concentration in the roadway.
[0058] By combining the BP neural network, the genetic algorithm (GA), and Transformer, this invention realizes the high-precision prediction and dynamic optimization control of the dust concentration in the mine roadway:
[0059] The BP neural network combines with Transformer for feature extraction, and captures the temporal dependence relationship in historical data through the self-attention mechanism, improving the prediction accuracy by about 30% compared with traditional methods, effectively guiding mine ventilation and dust removal, reducing prediction errors, and improving air quality;
[0060] Data augmentation and anomaly detection techniques enhance the robustness of the model, ensure high accuracy even in complex environments, solve the problem of inaccuracy of traditional methods under data fluctuations, and improve the adaptability of the model;
[0061] The GA algorithm globally optimizes the dust control parameters, dynamically adjusts the radial-axial air volume ratio, the extraction air volume, and the distance of the dust control device, ensures that the optimized parameter combination achieves efficient dust removal, and reduces the energy consumption by 15%-20%, improving the control efficiency;
[0062] Combining cross-validation and data augmentation techniques improves the generalization ability of the model, enables it to adapt to different mine working conditions, maintains high stability and low error, and is more flexible in dealing with complex environments.
[0063] Optionally, the CNN-LSTM combined model can be used to replace the BP-Transformer prediction structure, using CNN for feature extraction and LSTM for temporal modeling to improve the dynamic prediction ability of the dust concentration.
[0064] Optionally, increase the number of data sets and combine synthetic data augmentation to further improve the prediction accuracy. At the same time, find the mathematical relationship between the dust concentration and key environmental parameters through data analysis to enhance the interpretability and industrial applicability of the model.
[0065] In the above embodiments, the reference in the specification to "this embodiment" means that the specific features, structures, or characteristics described in connection with the embodiment are included in at least some embodiments, but not necessarily all embodiments. Multiple occurrences of "this embodiment" do not necessarily all refer to the same embodiment.
[0066] In the above embodiments, although the present invention has been described in connection with specific embodiments of the present invention, many alternatives, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art based on the foregoing description. For example, other storage structures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed. Embodiments of the present invention are intended to cover all such alternatives, modifications, and variations that fall within the broad scope of the appended claims.
[0067] This embodiment also provides a computer-readable storage medium having a computer program stored thereon, which when executed by a processor implements any one of the methods in this embodiment.
[0068] This embodiment also provides an electronic terminal, including: a processor and a memory;
[0069] The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory so that the terminal executes any one of the methods in this embodiment.
[0070] For the computer-readable storage medium in this embodiment, those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to the computer program. The foregoing computer program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps including the above method embodiments; and the foregoing storage medium includes: various media such as ROM, RAM, magnetic disk, or optical disc that can store program codes.
[0071] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver, and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store a computer program, and the communication interface is used for communication. The processor and the transceiver are used to run the computer program to make the electronic terminal execute each step of the above method.
[0072] In this embodiment, the memory may include a random access memory (Random Access Memory, abbreviated as RAM), and may also include a non-volatile memory, such as at least one disk memory.
[0073] The above-mentioned processor may be a general-purpose processor, including a Central Processing Unit (CPU for short), a Network Processor (NP for short), etc.; it may also be a Digital Signal Processor (DSP for short), an Application Specific Integrated Circuit (ASIC for short), a Field-Programmable Gate Array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
[0074] The present invention can be used in numerous general-purpose or special-purpose computing system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on.
[0075] The present invention can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present invention can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.
[0076] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the present technical solution, and they should all be covered within the scope of the claims of the present invention.
Claims
1. A multi-parameter dynamic optimization industrial dust control method, characterized in that: The following steps are involved: S1: Collect dust prediction related data in the mine tunnel, perform preprocessing, and form a data set; S2: inputting the data set into a dust concentration prediction model based on a Transformer-BP network to predict dust concentration; S3: Genetic algorithm GA is used to dynamically adjust the air volume control parameters to minimize the dust concentration; S4: Output optimized air volume control parameters to control the dust concentration in the tunnel.
2. The multi-parameter dynamic optimization industrial dust control method according to claim 1 is characterized in that: The dust prediction related data include the cross-sectional area of the tunnel, the compressed air volume, the exhausted air volume, the distance between the dust extraction port and the head, the length of the dust control device, the distance between the dust control device and the head, and the radial-axial air volume ratio.
3. The multi-parameter dynamic optimization industrial dust control method according to claim 1 is characterized in that: The pre-processing comprises the following steps: Organize the data and convert it into a CSV file; Divide input and output and normalize data; The normalized data is cleaned, feature engineered, and enhanced to obtain the final data set.
4. The multi-parameter dynamic optimization industrial dust control method according to claim 1 is characterized in that: The dust concentration prediction model based on Transformer-BP network includes a Transformer module and a BP neural network; The Transformer module processes historical dust concentration data through a self-attention mechanism, extracts long-term dependencies, enhances prediction capabilities, and makes the model more adaptable to environmental changes; The BP neural network adopts three hidden layers, with 64, 32 and 16 neurons respectively. Each layer is followed by a ReLU activation function to introduce nonlinear characteristics. Dropout regularization is adopted to randomly discard some neurons.
5. The multi-parameter dynamic optimization industrial dust control method according to claim 1 is characterized in that: In step S2, Transformer is first used to extract temporal features, and the long-term dependencies of historical data are captured through the self-attention mechanism; Define BP neural network to predict A and B values; Initialize the Transformer-BP network model and optimizer; Set the maximum number of iterations, iteratively train the Transformer-BP network model, and verify the loss.
6. The multi-parameter dynamic optimization industrial dust control method according to claim 1 is characterized in that: The genetic algorithm GA is used to optimize the weight initialization of the BP network and improve the training stability. It is also used to optimize the air volume control parameters, including the radial-axial air volume ratio, the exhaust air volume and the dust control device distance, and dynamically adjust the air volume control strategy to minimize the dust concentration and maximize the dust removal efficiency.
7. The multi-parameter dynamic optimization industrial dust control method according to claim 1 is characterized in that: The genetic algorithm GA comprises the following steps: Initialize the genetic algorithm optimizer; Genetic algorithm is used to evolve parameters and select the best individuals; Define the objective function for calculating the C value; Dynamically adjust air volume control parameters; Genetic algorithm evolutionary search for optimal C value and parameters; Combined with the prediction results of the Transformer-BP network model, the parameters optimized by GA were evaluated; Output the optimal C value and control parameters after GA optimization.