Innovative analysis of tailing characterisation for topsoil improvement

This research project explores the potential repurposing of coal tailings to promote sustainable tailings management practices in the mining industry. The study involves comprehensive characterisation of coal tailings samples linked with hyperspectral analysis, delving into their unique characteristics and essential parameters relevant to soil health and plant growth. The application of hyperspectroscopy has been instrumental in distinct coal tailings samples, with particular emphasis on the near-infrared (NIR) region. Analysis of spectral data yields valuable information concerning the chemical composition and mineral content of the tailings, thereby informing subsequent data-driven decision-making and the development of prediction models. This study encompasses the exploration of various machine learning (ML) techniques, including linear regression, logistic regression, polynomial regression, decision trees, and the ML ensemble tree bagger algorithm. Notably, the Ensemble-Bagged Trees Algorithm emerged as the most efficacious tool for predicting key tailings properties. The ML ensemble tree bagger algorithm is adept at handling complex data and emerged as the most apt choice, achieving high prediction accuracy. Its integration with hyperspectral data facilitated precise and reliable predictions for a number of coal tailing elements, which can provide a possible augmenting decision-making capability in tailings management and resource optimisation. The culmination of this research was the construction of a potent ML-Hyperspectral model, harmonising machine learning and spectroscopic data. This model exhibits remarkable accuracy in predicting pivotal tailings characteristics, thereby empowering data-driven decision-making in tailings management and advocating sustainable practices within the mining industry. The model's effectiveness in tailings characterisation holds profound implications for optimising the utilisation of coal tailings as a prospective soil amendment. In summary, this research contributes significant insights into the potential of coal tailings for repurposing and their role in enhancing soil health. The development of a data-driven prediction model serves as a catalyst for informed decision-making, resource optimisation, and responsible tailings management practices in the mining sector, thereby contributing to a more environmentally sustainable future.