This project investigated the cost-effectiveness of different fire protection strategies for use in industrial buildings in New Zealand, and developed a risk cost benefit model using @RISK to estimate the cost of fire in industrial buildings. Fire protection options included sprinklers, detection with manual suppression, and compartmentation.

The risk model was probabilistic and accounted for variability and uncertainty in the input data by incorporating probability distributions for inputs. Input data for the model relied on previous research on the cost of industrial fires in New Zealand carried out by Business and Economic Research Ltd (BERL) and supplemented with other data from the literature as well as engineering judgement. Latin hypercube simulation was used to generate an output distribution for the cost of fire. Twenty-five thousand iterations were conducted for each option.

Based on the upper 95% confidence level for the expected cost of fire per building per year, it is concluded that no change to the fire protection system requirements in the New Zealand building code compliance documents for industrial buildings is warranted. However, if buildings of more than 1000 m² in floor area are targeted, then it is recommended that automatic fire detection with manual suppression is the preferred option, closely followed by
fire sprinklers.

The study demonstrates a methodology that addresses uncertainty and provides a more robust analysis for decision-making about Building Code requirements. It also helps identify those parameters that most affect the outcome of interest and those where better data would reduce uncertainty in the results.