This report summarises the work accomplished in this one-year project in which the fire performance of wood-frame buildings was to be documented. A detailed analysis of Canadian and American fire loss statistics for residential occupancies was undertaken in order to assess the impact of the choice of building materials and the nature of fire-safety provisions in building codes on the overall fire safety in buildings. It was expected that the knowledge gained would enable the wood industry to argue more effectively during deliberations of codes and standards committees.
This study demonstrated that buildings constructed in compliance with current North American building code requirements are among the safest in the world. It was found that the fire loss record of wood-frame houses is about the same as that of large apartment buildings of non-combustible construction. It was shown that the ignition of upholstered furniture or mattresses by smokers’ materials is far and away the leading cause of residential fires involving deaths. Most of these deaths occur before the structure of the building becomes damaged by or involved in fire. Enacting more stringent building code requirements is unlikely to pay a large dividend in terms of life safety. In fact, the statistics suggest that significant improvements in fire safety in buildings would be more easily achieved by limiting the flammability of upholstered furniture and mattresses.
This report summarises progress in the second year of this project. Significant progress has been made towards achieving the original objectives of the project. In addition, several other applications of fire models have been identified that would further the interests of the Canadian wood industry and so appropriate research was initiated.
An objective of this project was to identify wood-stud walls that qualify as being of fireproof construction in Japan. To be classified as fireproof construction, a wood-stud wall must pass the 1 + 3 test in which it is subjected to a one hour fire-resistance test and then must support its load for another 3 hours as the furnace cools. Attempts were made to revise WALL2D to model the response of walls during the heating and cooling phases of an arbitrary fire. The revised model was to be used to model the response of walls in the 1 + 3 test and in furnished house fire tests run in Kemano. However, it turned out to be a major revision to include a cooling phase in WALL2D, but revisions were made to model a heating phase of an arbitrary fire. This was sufficient to get good agreement with temperatures measured within walls in Kemano. Revision of WALL2D to model the 1 + 3 test has been deferred until 2004-2005.
The Japan 2 x 4 Home Builders Association and the Council of Forest Industries have identified, by testing, wood-stud walls and wood-joist floors that pass the 1 + 3 test. These assemblies have been granted Ministerial Approval as being of fireproof construction. It is therefore possible to build 4-storey wood-frame apartment buildings in high-density urban areas. Employing models to identify assemblies that pass the 1 + 3 test is now less urgent, but will continue as models may suggest ways to optimise assemblies meeting the 1 + 3 test.
Another objective of this project was to undertake performance-based design of a building as a showcase study. Carleton University is developing a model to evaluate fire safety designs for 4-storey wood-frame commercial buildings. The first building to be analysed is a wood-frame version of the Carleton Technology Training Centre. The Carleton University model does not yet model the response of the structure of the building. To supplement Carleton University’s efforts, Forintek will undertake performance-based design for fire resistance of a wood-frame version of this building in 2004-2005.
While the initial completion date for this project was to be March 2004, it was intended that if other applications of fire models were identified that would further the goals of the Canadian wood industry, the project would be extended. During 2003-2004, several new applications of fire models were initiated:
A fire resistance model developed jointly by Forintek the National Research Council Canada is being employed to estimate the impact on fire-resistance ratings of the load applied to wood-stud walls during a test. This information would be useful when quoting the fire-resistance ratings of Canadian assemblies in export markets where lower loads are applied during fire tests.
A collaborative venture has been initiated with Australian researchers to model fires in large compartments (found in non-residential buildings) and the resultant response of wood-frame walls.
Data generated in fire tests conducted in furnished houses in Kemano is being used to assess the ability of current fire models to predict fire development in these houses and to predict the performance of a variety of building assemblies. If the models do a good job, one would have increased confidence in applying fire models in a performance-based design environment.
To demonstrate the good fire performance of wood-frame assemblies, three fire tests were run for visiting Chinese fire experts. Fire models were used to design the experiments to ensure that wood-frame assemblies were selected that could withstand the fire exposures envisioned in the tests.
This final report summarises progress in this multi-year project in which fire models have been applied to address a number of market access issues of interest to the Canadian wood industry.
Promoting wood-frame construction in Asian countries has been hindered by the fact that fire-resistance ratings assigned to wood-frame assemblies in Canada are lower than those assigned to similar assemblies in many other countries. Computer models developed at Forintek have been used to assess fire-resistance ratings of wood-frame walls subjected to test methods employed in a number of Asian countries. It was found that differences in fire-resistance ratings quoted in different countries are due to the different loads applied during the tests. Given the same exposures and loads, Canadian assemblies perform as well those from any other country.
A methodology for delivering performance-based design for fire-resistance of wood-frame buildings was developed. The methodology entails modelling the anticipated fire severity and using computer models to predict the performance of wood-frame assemblies protected by gypsum board. The methodology has been applied to two wood-frame buildings: a three-storey hotel and a three-storey office building.
Fire models have been used to assess the performance of wood products in a variety of practical applications in domestic and international markets.
Fire models have been shown to simulate the results of fire experiments conducted in wood-frame houses in Kemano thereby supporting the use of modelling in performance-based fire-safety design.
Forintek provided third-party review for the performance-based design of the expansion to the Vancouver Convention Centre. Modelling demonstrated that suspended glulam ceilings can be safely employed in a ballroom and pre-function areas despite non-compliance with building codes.
A fire protection firm is assessing the viability of utilizing wood trusses to create a pitched roof assembly on existing concrete buildings in Beijing. Performance-based design techniques employing fire models are being employed in the work. Forintek is providing advice to the fire protection firm.
Efforts have also been made to promote performance-based design at home and in Canada’s export markets as a strategy to eliminate the inequitable treatment afforded wood products by prescriptive codes.
Forintek scientists made presentations during an APEC Seminar convened to inform regulators of approaches to managing fire risks so as not to impede the use of wood products unnecessarily.
Forintek scientists have co-authored a chapter in the 4th Edition of the SFPE Handbook of Fire Protection Engineering which will be published in 2006. Participation in writing such documents is part of the Fire Group’s strategy to foster acceptance of performance-based design for fire safety.
Forintek scientists are participating in ISO deliberations addressing the performance of structures in fires. The methodology developed in this project is to be considered for inclusion in design guides.
Efforts are well advanced to develop improved fire models for predicting the thermal and structural response of wood-frame assemblies. These improved models are required for performance-based design in which fires typically grow quickly and after burning at a steady rate for a period of time undergo a decay phase.
The report concludes with the recommendation that it would be useful to run an extensive set of cone calorimeter tests on SCL, glue-laminated timber and CLT products. The fundamental data could be most useful for validating models for predicting flame spread ratings of massive timber products and useful as input to comprehensive computer fire models that predict the course of fire in buildings. It is also argued that the cone calorimeter would be a useful tool in assessing fire performance during product development and for quality control purposes.