The 10th International Conference on Energy, Resources, Environment and Sustainable Development (ERESD 2024)

258 / 2024-03-14 15:52:35

Design of Coal Mine Water Inrush Prediction Model Based on Local Linear Embedding-Fruit Fly Optimization Algorithm-Support Vector Classifier

water inrush from coal mine floor,support vector classification, local,linear embedding algorithm,fruit fly optimization algorithm

Intelligent Equipment Technology > 1．Intelligent Mine Equipment Technology

   To address the issues of low prediction accuracy and slow training speed in coal mine water inrush prediction models, a model based on Locally Linear Embedding (LLE), Fruit Fly Optimization Algorithm (FOA), and Support Vector Classifier (SVC) is designed. Firstly, considering the nonlinear and high-dimensional characteristics of bottom water inrush data, the LLE algorithm is employed to reduce data dimensionality and noise while preserving the local linear properties of the data. Secondly, the FOA algorithm is utilized to optimize the SVC model, avoiding the randomness and blindness in parameter selection. Finally, the necessity and feasibility of combining LLE, FOA, and SVC algorithms are analyzed, and a coal mine water inrush prediction model based on LLE-FOA-SVC is designed. Several commonly used methods for predicting coal mine water inrush, including the water inrush coefficient method, BP neural network and SVC model are compared with LLE-FOA-SVC model through simulation. Experimental results demonstrate that the prediction accuracy of the proposed model is higher than that of the other three models, achieving an accuracy up to 92% with shorter modeling and computation times.