Reliability Based Analysis of Effects of Fire Service Intervention and Sprinklers on Building Fire Resistance

  • Iziengbe Inerhunwa

Student thesis: Phd


Fire in buildings is one of the major hazards that can cause significant damages and loss of life all over the world. To mitigate losses associated with fire incidents, fire safety protection systems are put in place in buildings to facilitate occupant escape, to slow down fire propagation, to contain fire and to fight fire. These fire protection methods can be grouped broadly into active and passive systems. Active fire protection systems are provided to detect and suppress the growth of fire and passive fire protection systems limit fire spread or prevent structural collapse. There are many challenges in incorporating the beneficial effects of fire protection systems into fire safety design methods, including how to deal with uncertainties that characterise both active and passive fire protection measures, as well as their interaction. This thesis presents the results of detailed research to characterise uncertainty in intumescent coating as a passive fire protection material and develop practical methods of incorporating two active fire protection systems, fire service intervention and sprinklers, in specification of passive fire protection requirements for fire resistance design of structures. The scope of applicability of current methods in design codes is also assessed. Since there are many uncertainties in how different active and passive fire protection systems perform in fire, a probabilistic reliability analysis method was adopted for this research. Statistical analysis of fire test data is used to characterize the uncertainties in temperature dependent thermal conductivity and dry film thickness (DFT) of intumescent coating, and results showed that they are best described by lognormal and normal probability distributions, respectively. Partial safety factors for intumescent coating thermal conductivity to achieve different target probabilities of failure are also derived. A probabilistic analysis method is developed to explicitly include the influences of fire service intervention (FSI) and sprinkler operation in fire resistance design of intumescent coating protected steel. The method considers different scenarios of fire behaviour following FSI and sprinkler operation, based on probabilistic analysis of current UK fire data. These statistical data are then used to generate random fire temperature-time curves to calculate the probability of structural failure of intumescent coating protected steel members using the Monte Carlo Simulation method. To aid deterministic design, a simplified method is developed to derive fire load modification factors to account for the effects of FSI alone (dn,FSI), sprinkler alone (dn,SPRK) and the combined effects of sprinkler activation and FSI (dn,SPRK-FSI). An extensive parametric study covering a wide range of practical design factors is carried out and an Analysis of Variance (ANOVA) is performed on the results to examine whether fire load modification factor is sensitive to any design factor. The application of the fire load modification factors to reinforced concrete structures is also evaluated. The results show that fire growth rate and compartment floor area are the most influential design factors. The effects of fire load, steel critical temperature and fire protection thickness have moderate effects. Furthermore, the results show that multiplication of the individual factors dn,FSI and dn,SPRK to account for the combined effects of sprinkler activation and FSI acting together, as is the current practice in Eurocode 1 Part 1.2, overestimates the effects of both systems combined, give unsafe results and should not be allowed. In addition, the results suggest that the same values of fire load modification factors can be used for both intumescent coating protected steel and reinforced concrete structures. Based on the results, new values of dn,SPRK, dn,FSI and dn,SPRK-FSI are recommended. To achieve a target reliability level, a probabilistic method is also developed and used
Date of Award1 Aug 2022
Original languageEnglish
Awarding Institution
  • The University of Manchester
SupervisorMeini Su (Supervisor) & Yong Wang (Supervisor)


  • Monte Carlo simulation
  • Reliability analysis
  • Probabilistic analysis
  • Sprinkler
  • Fire service intervention
  • Fire load factor
  • Intumescent coating
  • Fire resistance

Cite this