Application of decision analysis to seismic rehabilitation of historic buildings : a case study of rehabilitation of Stanford University Memorial Church

The construction of an expected value decision model for selecting seismic retrofit schemes for a historic building is demonstrated through a case study. Net Present Costs (NPC) are used to rank the retrofit options. NPC is defined as the sum of the initial investment cost and the present expected value of the total future damage costs. Four options are assumed to be proposed to upgrade the Stanford Memorial Church in 1980. Option 1 is to do nothing in terms of seismic strengthening, whereas options 2 to 4 are seismic strengthening the Church with increasing level of safety and costs. Seismic data are selected to derive a set of earthquake probabilities. Damage probabilities and damage costs, including direct costs, indirect costs, and costs of life, are estimated. A discount rate and a life span are chosen to discount the future damage costs to present expected values. Option 3, strengthening the Church to building integrity standard without removing the unreinforced masomy (URM) walls, has the lowest NPC and is considered as the optimal option. If it had been recommended and adopted in 1980, about 40% of the accumulative expense from 1980 to 1990 would have been saved. A sensitivity analysis is carried out in order to determine the importance of variables that are only crudely estimated. In general, the outcome is insensitive to changes in seismic data, damage costs, discount rates and life spans when the changes are within reasonable ranges. To obtain higher accuracy, attention should be paid to estimating the probabilities and damage costs of moderate to large earthquakes (MIvil V to IX) rather than that of great earthquakes (MIvil X and above). The outcome is sensitive to the expected reduction in damage due to retrofits. Better methods for predicting structural behavior is needed. More research in developing guidelines for estimating future damages and social impacts is recommended.Applied Science, Faculty ofCivil Engineering, Department ofGraduat