Abstract
Three types of extreme value distributions are fitted to the maximum daily temperature and solar radiation. It is found that type III distribution for the largest value fits the data most closely. A methodology using the maximum yearly temperature data and extremal distributions is developed for the optimum design of refrigerated warehouses. The use of the concept of return period in the optimum design of thermal systems is also suggested. The interior penalty function method with Davidon-Fletcher-Powella method of unconstrained minimization is used as the optimization technique for solving the problems. A sensitivity analysis is conducted about the optimum design point to find the influence of changes in various design parameters on the cooling load and total cost.
