A study on the development of an artificial neural network model for the prediction of ground subsidence over abandoned mines in Korea. Issue 3 (4th May 2017)

• Record Type: Journal Article
• Authors: Kim, Yangkyun; Son, Min; Moon, Hyun-Koo; Lee, Seung-Ah
• Is Part Of: Geosystem engineering. Volume 20:Issue 3(2017)
• Page Start: 163
• Page End: 171
• Publication Date: 2017-05-04
• Subjects: Abandoned mine -- artificial neural network -- subsidence -- prediction; Mining engineering -- Periodicals; Petroleum engineering -- Periodicals; Gas engineering -- Periodicals; Geology, Economic -- Periodicals; 620
• DOI: 10.1080/12269328.2016.1254573
• Languages: English
• ISSNs: 1226-9328

Abstract

Abstract: The subsidence over abandoned mines has led to big problems in the local areas. Hence, many attempts have been made to predict subsidence and to check the stability of an abandoned mine. This paper focuses on the development of an artificial neural network (ANN) model to predict mine subsidence, and the verification of prediction ability, based upon the investigation of 247 subsidence areas over 27 abandoned mines in Korea. The analysis is divided into two stages. In stage 1 where the data of 229 cases are used for learning as well as prediction of ANN, it is found that the correct answer percentage of final output values to the actual values shows 99, 95, 100 and 100% for four output parameters of subsidence type, final decision, filling method, and local reinforcement respectively. In stage 2 where the information of 18 other cases is used for prediction after learning with the 229 cases, the correct answer percentage of final output values is 89, 78, 100, and 94% for the four parameters, respectively. Finally, it is thought that this ANN model can be used for the prediction of unknown situations regarding subsidence over an abandoned mine with high reliability.