CN-121993431-A - Intelligent optimization control method and system for mine ventilator based on GA-SVR

CN121993431ACN 121993431 ACN121993431 ACN 121993431ACN-121993431-A

Abstract

The invention relates to the technical field of ventilation regulation and control, and provides an intelligent optimizing control method and system of a mine ventilator based on GA-SVR, wherein the intelligent optimizing control method comprises the steps of generating an initial population, and encoding a combination of a blade angle and a motor frequency by each individual in the population; based on the required ventilation quantity and wind pressure, a trained support vector regression model is adopted, after the total pressure efficiency and wind pressure of each individual are predicted, a second fitness value of each individual is calculated, and self-adaptive genetic operation and elite retention strategy are executed until convergence is achieved, so that the optimal total pressure efficiency and wind pressure are obtained, and the mine ventilator is controlled, wherein the support vector regression model uses a genetic algorithm to carry out super-parameter optimization in the training process. An intelligent regulation and control mechanism taking 'on-demand air supply' as a core is established, and the limitation of small sample data is effectively overcome.

Inventors

LU GUOLIANG
LIU YANG

Assignees

山东大学

Dates

Publication Date: 20260508
Application Date: 20260105

Claims (10)

1. The intelligent optimizing control method of the mine ventilator based on the GA-SVR is characterized by comprising the following steps: Acquiring required ventilation quantity, wind pressure and a selected mode; Generating an initial population, each individual in the population encoding a combination of blade angle and motor frequency; Based on the required ventilation quantity, wind pressure and mode, a trained support vector regression model is adopted, after the total pressure efficiency and wind pressure of each individual are predicted, inverse optimization solution is started, a second fitness value of each individual is calculated, and adaptive genetic operation and elite retention strategy are executed until convergence is achieved, so that the optimal total pressure efficiency and wind pressure are obtained, and the mine ventilator is controlled; the support vector regression model uses a genetic algorithm to perform super-parameter optimization in the training process.
2. The intelligent optimization control method for the mine ventilator based on the GA-SVR as set forth in claim 1, wherein said calculation of the second fitness value comprises the product of the wind pressure error and the penalty coefficient and the total pressure efficiency of the ventilator.
3. The intelligent optimizing control method for mine ventilator based on GA-SVR as set forth in claim 2, wherein the penalty coefficient is set according to the selected optimizing mode, the penalty coefficient in the air volume priority mode is smaller than the penalty coefficient in the equalizing mode, and the penalty coefficient in the equalizing mode is smaller than the penalty coefficient in the air pressure priority mode.
4. The intelligent optimizing control method for mine ventilator based on GA-SVR as set forth in claim 1, wherein in the adaptive genetic operation, when the number of the population convergence stagnation algebra exceeds a set value, the mutation probability is increased and the crossover probability is decreased.
5. The intelligent optimizing control method for mine ventilator based on GA-SVR as set forth in claim 1, wherein in the adaptive genetic operation, in the normal convergence phase, the linear adjustment strategy is adopted to adjust the crossover probability Probability of variation : ; ; Wherein, the For the initial probability of variation, the first time, For the initial cross-over probability to be the same, And g is the current iteration number, wherein g is the preset total evolution algebra.
6. The intelligent optimization control method for the mine ventilator based on the GA-SVR as set forth in claim 1, wherein said super parameters include regularization parameters, nuclear parameters and insensitive loss parameters.
7. The intelligent optimization control method for the mine ventilator based on the GA-SVR as set forth in claim 1, wherein the training process of the support vector regression model comprises: generating an initial population, each individual in the population encoding a hyper-parametric combination; For each individual, configuring 3 support vector regression models by using super parameters, wherein the 3 support vector regression models are respectively used for predicting total pressure efficiency, wind pressure and input power, training the support vector regression models on a training set, calculating a first fitness value of each individual on a test set, and executing genetic operation and elite retention strategy until convergence to obtain an optimal super parameter combination; retraining the support vector regression model with the optimal hyper-parametric combination and evaluating its performance using all training data; and if the performance reaches the standard, the support vector regression model training is completed.
8. GA-SVR-based intelligent optimizing control system for mine ventilator is characterized by comprising the following components: A model training module configured to perform a hyper-parametric optimization of the support vector regression model using a genetic algorithm during training; a ventilation demand acquisition module configured to acquire a required ventilation amount, wind pressure, and a selected mode; An initialization module configured to generate an initial population, each individual in the population encoding a combination of blade angle and motor frequency; the reverse optimization control module is configured to predict the total pressure efficiency and the wind pressure of each individual by adopting a trained support vector regression model based on the required ventilation quantity, wind pressure and mode, then start reverse optimization solution, calculate the second fitness value of each individual, and execute self-adaptive genetic operation and elite retention strategy until convergence to obtain the optimal total pressure efficiency and wind pressure so as to control the mine ventilator.
9. A computer-readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, realizes the steps in the GA-SVR-based intelligent optimizing control method of a mine ventilator according to any one of claims 1-7.
10. A computer device comprising a computer readable storage medium, a processor and a computer program stored on the computer readable storage medium and executable on the processor, characterized in that the steps in the GA-SVR based intelligent optimization control method of a mine ventilator according to any one of claims 1-7 are implemented by the processor when executing the program.

Description

Intelligent optimization control method and system for mine ventilator based on GA-SVR Technical Field The invention belongs to the technical field of ventilation regulation and control, and particularly relates to an intelligent optimizing control method and system for a mine ventilator based on GA-SVR. Background The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art. As a core barrier to maintain safety in downhole operations, the main ventilator in the mine ventilation system needs to be kept continuously running for a long period, which brings about an extremely high energy consumption burden. However, the conventional control means are limited by the conventional manual experience judgment or rigid scheduling mode, and are difficult to accurately follow the complicated and changeable ventilation load demands under the well, so that the mine ventilation system is normalized to be in an inefficient operation state, and obvious energy loss is caused. With the iteration of the industrial internet of things and the intelligent algorithm, data-driven fan optimization becomes a hot spot problem. In the prior mainstream methods such as computational fluid dynamics (Computational Fluid Dynamics, CFD) simulation, machine learning, meta heuristic optimization and the like, many stoppers exist in actual deployment, the CFD calculation cost is high to prevent real-time application, the machine learning model is severely predicted to fluctuate because of manual parameter tuning, and the optimization algorithm is blocked in optimizing under the safety constraint, so that the energy efficiency and the safety target are difficult to be considered. Disclosure of Invention In order to solve the technical problems in the background art, the invention provides the intelligent optimizing control method and system for the mine ventilator based on the GA-SVR, a parameter reverse inversion frame is constructed, the optimal operation condition can be reversely pushed according to the ventilation load demand changing in real time, an intelligent regulation and control mechanism taking 'air supply on demand' as a core is established, the Genetic Algorithm (GA) is utilized to realize the automatic optimizing of the Support Vector Regression (SVR) model super-parameters, and the limitation of small sample data is effectively overcome. In order to achieve the above purpose, the present invention adopts the following technical scheme: the first aspect of the invention provides an intelligent optimizing control method for a mine ventilator based on GA-SVR, which comprises the following steps: Acquiring required ventilation quantity, wind pressure and a selected mode; Generating an initial population, each individual in the population encoding a combination of blade angle and motor frequency; Based on the required ventilation quantity, wind pressure and mode, a trained support vector regression model is adopted, after the total pressure efficiency and wind pressure of each individual are predicted, inverse optimization solution is started, a second fitness value of each individual is calculated, and adaptive genetic operation and elite retention strategy are executed until convergence is achieved, so that the optimal total pressure efficiency and wind pressure are obtained, and the mine ventilator is controlled; the support vector regression model uses a genetic algorithm to perform super-parameter optimization in the training process. Further, the calculation of the second fitness value includes the product of the wind pressure error and the penalty coefficient and the total pressure efficiency of the fan. Further, the punishment coefficient is set according to the selected optimization mode, the punishment coefficient in the air quantity priority mode is smaller than the punishment coefficient in the balance mode, and the punishment coefficient in the balance mode is smaller than the punishment coefficient in the air pressure priority mode. Further, in the adaptive genetic operation, when the population convergence stagnation algebra exceeds a set value, the mutation probability is improved, and the crossover probability is reduced. Further, in the adaptive genetic operation, in the normal convergence stage, the crossover probability is adjusted by adopting a linear adjustment strategyProbability of variation: ; ; Wherein, the For the initial probability of variation, the first time,For the initial cross-over probability to be the same,And g is the current iteration number, wherein g is the preset total evolution algebra. Further, the super parameters include regularization parameters, kernel parameters, and insensitivity loss parameters. Further, the training process of the support vector regression model includes: generating an initial population, each individual in the population encoding a hyper-parametric combination; For each individua