Abstract
In recent years, use of artificial neural networks have increased for estimation of Hardgrove grindability index (HGI) of coals. For training of the neural networks, gradient descent methods such as Backpropagaition (BP) method are used frequently. However they originally showed good performance in some non-linearly separable problems, but have a very slow convergence and can get stuck in local minima. In this paper, to overcome the lack of gradient descent methods, a novel particle swarm optimization and artificial neural network was employed for predicting the HGI of Kentucky coals by featuring eight coal parameters. The proposed approach also compared with two kinds of artificial neural network (generalized regression neural network and back propagation neural network). Results indicate that the neural networks - particle swarm optimization method gave the most accurate HGI prediction.
Original language | English |
---|---|
Pages (from-to) | 201-212 |
Number of pages | 12 |
Journal | Energy Exploration and Exploitation |
Volume | 27 |
Issue number | 3 |
DOIs | |
State | Published - Jun 2009 |
Keywords
- Coal petrography
- Hardgrove grindability index
- Neural networks
- Particle swarm optimization
ASJC Scopus subject areas
- Renewable Energy, Sustainability and the Environment
- Nuclear Energy and Engineering
- Fuel Technology
- Energy Engineering and Power Technology