Coal Air-Dry Basis Sulfur Content Detection Model Based on Hippopotamus Optimization Algorithm and Two Hidden Layer Extreme Learning Machine

Coal is a key energy resource in China. The accurate classification and component content detection of coal are vital for efficient use and industrial production. Traditional coal air-dry basis sulfur content detection methods are destructive, time-consuming, and costly, failing to meet modern industry's rapid and precise testing needs. This study innovatively applies machine learning algorithms to the data processing of laser-induced breakdown spectroscopy (LIBS) for in-depth detection of sulfur content in coal based on air drying. LIBS acquires coal sample spectral data, which is then denoised via wavelet transform and dimensionality-reduced through principal component analysis (PCA). A detection model based on a two hidden layer extreme learning machine (TELM) and the hippopotamus optimization (HO) algorithm is constructed, and a sensitivity analysis is finished. Results show the HO-TELM algorithm significantly improves detection accuracy of coal component content compared to traditional ELM algorithms, offering an efficient and reliable solution for intelligent coal resource detection.