Transformer tank optimization using design of experiments

Abstract

Background / Objective: Present study deals with the optimization of 6.6 MVA 66/11.5 KV transformer tank using Design Of Experiments (DOE) concepts such as Screening Methodology, Central Composite Design (CCD) and Response Surface Methodology (RSM). Objective is to reduce the weight of the transformer tank. Methods / Analysis: To achieve objective mentioned above individual wall analysis and optimization approach was used to reduce complexity and time requirement. In fact DOE simplified the number of FEA simulations. Screening Methodology is used for filter out the parameters or design variables which are not necessary for optimization. CCD technique helps to create quadratic model for response or objective variable and hence preparation of DOE matrix. Response surface was generated from the resulted DOE matrix to obtain optimum design point. Findings: Individual wall analysis is found time saving and verified with whole analysis. Approximately 15% reduction in weight of the wall resulted from the study. Same approach can be used for other walls of the transformer to get optimization of whole tank. Also methodology can be used to study different stiffener arrangement of the tank wall to get optimum response such as weight, stresses and deformation.