مقطع : دکتری
دانشگاه : Universiti Teknologi Malaysia
تاریخ دفاع : 2015/06/05
اساتید راهنما : ASSOC.PROF.DR. EDY TONNIZAM BIN MOHAMAD
اساتید مشاور : DR. MOGANA SUNDARAM NARAYANASAMY
اساتید داور : Dr. Nazri Ali
مشاهده سایر پایان نامه های دانیال جاهد ارمغانی
Many works highlight the use of effective parameters in Tunnel Boring Machine (TBM) performance predictive models. However, there is a lack of study considering the effects of tropically weathered rock mass in these models. This research aims to develop several models for predicting Penetration Rate (PR) and Advance Rate (AR) of TBMs in fresh, slightly weathered and moderately weathered zones in granite. To achieve these objectives, an extensive study on 12,649 m of the Pahang- Selangor Raw Water Transfer (PSRWT) tunnel in Malaysia was carried out. The most influential parameters on TBM performance in terms of rock (mass and material) properties and machine specifications were investigated. A database consisting the tunnel length of 5,443 m, 5,530 m and 1,676 m representing fresh, slightly weathered and moderately weathered zones, respectively was analysed. Based on field mapping and laboratory study, a considerable difference of rock mass and material characteristics has been observed. In order to demonstrate the need for developing new models for prediction of TBM performance, two empirical models namely QTBM and Rock Mass Excavatability (RME) were analysed. It was found that empirical models could not predict TBM performance of various weathering zones satisfactorily. Then, multiple regression (i.e. linear and non-linear) analyses were applied to develop new equations for estimating PR and AR. The performance capacity of the multiple regression models could be increased in the mentioned weathering states with overall coefficient of determination (R2) of 0.6. Furthermore, two hybrid intelligent systems (i.e. combination of artificial neural network with particle swarm optimisation and imperialism competitive algorithm) were developed as new techniques in field of TBM performance. By incorporating weathering state as input parameter in hybrid intelligent systems, performance capacity of these models can be significantly improved (R2 = 0.9). With a newly-proposed systems, the results demonstrate superiority of these models in predicting TBM performance in tropically weathered granite compared to other existing and proposed techniques.