EXPERT SYSTEM FOR SEISMIC RESISTANT ANALYSIS AND DESIGN OF ELEVATED WATER TANK.

Abstract

Artificial intelligence has two important applications in the field of Engineering analysis and design. They are expert systems and Neural Networks. Expert systems are computer programs that emulate the behaviour of a human expert within a specific domain of knowledge. A neural network is a computational model that is a directed graph composed of nodes and connections between the nodes which replicate the working of human brain. The ideal application of these two tools are for solving problems which require levels of expertise that is not always available and which have imprecise data. The design codes provide the guide lines for analysis and design of civil engineering structures. The knowledge in the design codes is usually widespread and ill structured which in turn means obtaining the necessary information is a tedious job. The knowledge in the field of civil engineering design is gained through experience. So, the information available for the analysis and design is meagre and ill structured. Thus expert systems and neural networks can prove to be effective tools for civil engineering analysis and design. In this dissertation seismic analysis and design of an elevated intze water tank has been examined in depth for the applicability of Al tools and techniques. The problem of analysis and design of intze tank has been dealt with using both the techniques of AI i.e. expert systems for replacement of decision making and neural networks for simulating analytical and design processes. In the first part "ESETAD" expert system for elevated tank analysis and design has been developed which assists the user in step-by-step analysis and design of intze water tank. For the analysis and design of various components of tank modules have been written in C++. The knowledge bases of this package have been derived from IS:1893-1984 and IS:875-1987 and standard literature partially based on thumb rules in existence. In the second part three neural network. models have been developed which are for the dimensioning, analysis and design of the intze tank. On the basis of comparison of results from conventional procedures and artificial neural networks it has been concluded that the use of expert systems and neural networks in the field of engineering analysis and design can eliminate rigorous conventional procedures with acceptable accuracy.