مهدی راضی فرد

Nowadays, earth pressure balance tunnel boring machines (EPB-TBMs) are favorably used in urban mechanized tunneling projects. The availability of these machines is rather low, so their performance, time planning, maintenance scheduling and cost control may need further improvements. The reliability, availability and maintainability (RAM) of any machine could be analyzed by evaluating the actual performance of that machine or system and thereby minimizing maintenance costs. In current research work, RAM analysis of the electrical system of the EPB-TBM in line 1 of Tabriz metro in Iran is performed. For this investigation, failure and repair data were collected during about 26 months of machine operation. Statistical tests including trend and serial correlation tests indicated that the data are independently distributed, consequently the statistical technique was used for modeling. The data analysis also revealed that the time between failures (TBFs) and time to repairs (TTRs) data fit to the Lognormal (3P) and Gamma (3P) distributions, respectively. Reliability analysis showed that there is approximately 80.1% chance for the electrical system of TBM machine to be operational for 15 h without any failure. Moreover, the maintainability curve indicated that most of the failures have been repaired in less than 1 hour and that it is more likely that 80% of these failures would be repaired in 1.88h. Also the availability of electrical system of this machine is found to be equal to 97.8%.

A number of common laboratory rock mechanics tests are carried out in all geotechnical projects such as dams, to determine parameters such as porosity, density, water absorption, sonic velocity, Brazilian tensile strength, uniaxial compressive strength, and triaxial compressive strength. In this paper, data obtained from two dams in Asmari Formation including Khersan 1 and Karun 4 - both located in Chahar MahalVaBakhtiari Province, Iran - have been subjected to a series of statistical analyses. Then, using Multivariate Linear Regression (MLR) and Artificial Neural Networks values of UCS, E, C, and j were predicted using the input parametersincludingdepth, compressionultrasonic velocity, porosity, density, and Brazilian tensile strength. The designed ANN in this research was a feedforwardbackpropagation network which is powerful tool to solve prediction problems. Designed network had two hidden layer (hidden layer 1: 18 neurons and hidden layer 2: 20 neurons). Via comparing designed MLR and ANN models, it was revealed that ANNs (R2 UCS= 0.91, R2 E = 0.87, R2 C =0.78, R2 phi = 0.61) are more efficient than MLR models (R2 UCS= 0.69, R2 E= 0.69, R2 C = 0.66, and R2 phi = 0.50) in predicting strength and shear parameters of the intact rock. Also, to enhance the credibility of this study, some extra tests were carried out to evaluate the efficiency of network designed for prediction of strength parameters. The results obtained from this network were as: R2 UCS= 0.85, R2 E = 0.81.