Prediction of induced stress in concrete liner of swellable rocks using Adaptive-Network Based Fuzzy Inference System

Concrete liner of excavations in swellable rocks is under time dependent stress. One of the most famous methods for determining the stress is compensation method using flat-jacks. As the first step، measuring pins are cemented on the surface of the concrete lining in an appropriate arrangement on both sides of the planned cut. The distances between the pins are recorded by an accurate displacement sensor. Following the zero measurement، a slot is cut with a diamond-tipped circular saw blade. Hydraulic pressure cell (flat-jack) is inserted into the slot and is connected with a hydraulic pump fitted to a precision manometer. Finally، the flat jack is loaded until the relief-induced deformations are compensated. Since swelling is a time dependent behavior، the stress of liner changes with time. This needs a periodic measurement using heavy pumping device and requires a time consuming process. In this paper، the ability of adaptive network based fuzzy inference system is examined to predict the mentioned stresses. The performed compensation test at Masjed-Soleiman underground power house has been used as an example to illustrate the method. Using this method an engineer is enabled to predict the lining stress by reading the slot closure across a flat jack device.