Citation

Wolin SD, Ryder NL, Leprince F, Milke JA, Mowrer FW, Torero JL. Fire Technol. 2001; 37(4): 363-395.

Affiliation

Department of Fire Protection Engineering, University of Maryland, College Park, MD, USA

Copyright

(Copyright © 2001, Holtzbrinck Springer Nature Publishing Group)

DOI

10.1023/A:1012776916407

PMID

unavailable

Abstract

The characteristics of smoke traveling in an HVAC duct have been observed along with the response of selected duct smoke detectors. The simulated HVAC system consists of a 9 m long duct, 0.45 m in diameter. An exhaust fan is placed at one end of the duct and is capable of inducing airflow rates that range from 0 to 1.5 m 3/s. The flow is controlled by means of a manual damper. On the upstream end of the duct there is a square exhaust hood approximately 2.2 m at the bottom and 0.3 m at the top. The bottom of the hood is approximately 2.5 m above the floor a shroud extends down to approximately 1.5 m above the floor. The test section, placed immediately downstream of the hood, is 3.5 m long duct with a square cross section of 0.4 m on a side. The instrumentation includes oxygen, carbon monoxide and carbon dioxide gas analyzers and a load cell to determine the energy release rate of the fires tested. The smoke within the duct is characterized by means of a laser light sheet and charge coupled device (CCD) camera, two white light source and photocell ensembles, a Pitot tube and an array of eight thermocouples placed on the vertical plane of symmetry. A smoke detector was placed at the downstream end of the test section. Two types of detectors were tested, ionization and photoelectric, with a single sampling probe geometry. The fires tested cover a wide range of fuels (propane, heptane, toluene, toluene/heptane mixture, shredded paper, polyurethane foam, wood cribs) with the peak energy release rates up to 800 kW. The smoke detector performance, temperature, flow field, smoke particle size and particle distributions are dependent on the fire characteristics and airflow through the duct. The different measurements could be scaled by means of the fire size and airflow rate but left a strong dependency on the fuel and burning characteristics (i.e., smoldering, flaming). The optical density and mass optical density are analyzed as metrics for characterizing smoke and smoke detector response. Detailed comparisons between the different metrics used are presented throughout this work. Clear evidence of stratification and aging of the smoke along the duct are also presented. The limitations of the present configuration and the need for a larger scale study are also discussed.