کامران گشتاسبی گوهر ریزی

Compressed air storage is one of the methods of energy storage. One of the most important parts of a compressed air storage system is the storage cavern. In this study, the behaviour and the long-term stability of the salt cavern as a storage site for compressed air were investigated. Due to the construction and operation of the salt cavern, the in-situ stress conditions around the cavern will change. The difference between induce stress and pressure of compressed air will cause the cavern to converge and damage around it. In this study, the salt rocks from one of the regions of Iran were used to test creep. To investigate the creep behavior of salt rock, creep test was performed in three stress levels in a stepwise manner. Lubby2 model parameters were calculated. Then, using LOCAS finite element software, the compressed air storage cavern was simulated and its behavior and stability were investigated. The results of these experiments and modelling have shown that with increasing stress in each stage, the slope of the secondary part of the creep increases. Therefore, it can be concluded that with increasing stress, the Maxwell viscosity coefficient decreases. In addition, the amount of movement in the lower half of the cavern was more than the upper half of it. Value of the cavern safety factor according to the Devries dilation criterion also decreased with increasing depth and the probability of dilation in the cavern wall increased.

Observing ground nature and soil properties of the target area, is critical before constructing a structure.Excavation of exploring bore holes and performing required examinations in order to get knowledge of the soil properties, are inevitable necessities. Normally, there is not exact information in region between the bore holes. Excavating more bore holes and repeating the examinations are time consuming and expensive. Neuro Intelligence is a specialized software in the field of artificial neural networks. Neural network has high abilityto estimate non-linear relations. The purpose of this paper is to present a method to estimate the soil properties in the vicinity of bore holes. Layers coordinates and types were selected as input data and soil density as network output. Quick and online propagation algorithms were used for network training. Finally,the network was tested for all used algorithms. Comparing predictions of the network with real values, demonstrates the success of designed network and used algorithms.