Collecting, Analysing, and Presenting Reliability Data for Automatic Sprinkler Systems
Abstract
Much work has gone into studying the reliability of sprinkler systems, but there are large level differences between the studies. For example, a recent study in the United States (National Fire Protection association Research, 2017) sets reliability level at 88%, while a study in Australia and New Zealand finds 99.5% (Maybee, 1988) and one in the UK says 93% (Optimal Economics, 2017).
For this Master’s thesis, I conducted a critical review of the following studies: National Fire Protection Association (NFPA) reports from 1970, 2019, 2017, "Fire - A Century of Automatic Sprinkler Protection in Australia and New Zealand - 1886 to 1986" (Marryat, Rev. 1988) and "Efficiency and Effectiveness of Sprinkler Systems in the United Kingdom: An Analysis from Fire Service Data" (Optimal Economics, 2017). The review of these gave many questions and answers.
I validated the studies using document analysis, basing the analysis on how a scientific investigation should be carried out. All had problems in four out of seven possible areas: 1. unclear issues, including missing definitions and intentions of the investigations; 2. uncertain data collection process; 3. varying quality of analysis and lack of quality assurance; 4. lack of systematic presentation and discussion.
Based on this finding, I concluded that none of the reports on sprinkler reliability can be taken into account for a general documentation on reliability or on future probability for sprinkler systems to function as designed.
Document analysis has primarily been a tool for social science, but as this thesis shows, it is very useful in the field of fire science. Based on the findings from the document analysis, I propose a modified methodology adapted to the scientific principles of fire science. This is exemplified by two proposals for studies, the first descriptive and the second explanatory.
Efforts to provide reliable data for sprinkler systems have major implications for the reliability of all passive and active fire protection as well as performance-based design.