Journal
NEURAL COMPUTING & APPLICATIONS
Volume 30, Issue 7, Pages 2145-2158Publisher
SPRINGER LONDON LTD
DOI: 10.1007/s00521-016-2809-3
Keywords
Longwall mining; Height of caving-fracturing zone; Artificial neural network; Multivariable regression analysis
Categories
Ask authors/readers for more resources
After the ore (seam) extraction in longwall mining, the immediate roof layers over the extracted panel are strained and suspended downward. This process expands upward and causes the caving and fracturing of damaged roof rock strata. The combination height of the caved and interconnected fractured zones is considered as the height of caving-fracturing zone (HCFZ) in this research. Precise estimation of this height is crucial to the exact determination of directed loads toward the front and sides abutments. The paper describes an intelligent model based on the artificial neural network (ANN) to predict HCFZ. To validate the ability of ANN model, its results are compared to the multivariable regression analysis (MVRA) results. For models construction and evaluation, a wide range of datasets comprising of geometrical and geomechanical characteristics of mined panel and roof strata have been gathered. Performance evaluation indices including determination coefficient (R-2), variance account for, mean absolute error (E-a) and mean relative error (E-r) have been utilized to assess the models' capability. Comparison results show that the ANN model performance is considerably better than the MVRA model. Moreover, obtained results are further compared with the results of available in situ, empirical, analytical, numerical and physical models reported in the literature. This comparison confirms that a reasonable agreement exists between the ANN model and the previous comparable methods. Finally, the sensitivity analysis of ANN results shows that the overburden depth has the maximum effect, whereas the Poisson's ratio has the minimum effect on the HCFZ in this research.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available