Vulnerability and Threat Based Measures of Building Protection from Airborne Chemical and Biological Agents

Open Access
Aumpansub, Ponkamon
Graduate Program:
Architectural Engineering
Master of Science
Document Type:
Master Thesis
Date of Defense:
September 23, 2013
Committee Members:
  • William P Bahnfleth, Thesis Advisor
  • Vulnerability
  • threat
  • Risk
  • Assessment
  • Airborne Contaminants
  • HVAC
To mitigate risk in the event of extraordinary incidents including chemical and biological attacks, risk assessment is a necessary process to identify threats and systems vulnerabilities. Most available risk assessment methodologies are qualitative and lack building specific components to assess building system protection level. It is important that quantitative or semi-quantitative methods are developed to evaluate building systems, control, and remediation measures. These methods require reliable building performance measures to determine occupant health consequences of a building attack. This research investigated available threat-based and vulnerability-based measures specific for building mechanical systems and determine their accuracy and usefulness in risk assessment. Threat-based measures are event and agent specific measures indicating an absolute severity levels of building occupants in an attack event. Vulnerability-based measures are relative and non-agent specific measures which evaluate the impact of an event relative to a base event. Both approaches were carried out on a presumed contaminant release of 6 x 108 cfu of Bacillus anthracis inside or into an OA intake of a three-story, 16,791 ft2 (1,560 m2) residential building to determine the impact of mechanical system and event characteristics including filter efficiency, ventilation rate, mechanical system types, zoning strategy, and stack effect. The multizone modeling program CONTAM was used to simulate the spread of the contaminants after a release. The contaminant transport simulation was coupled with the appropriate epidemiological methodology to determine total infections and casualties which were used as reference measures. Vulnerability-based measures were also applied to the same scenarios and compared their interpretation to reference measures. Uncertainty analysis was performed on three filter scenarios (MERV 6 filter, MERV 13 filter and no-filter) to determine the effects of variations in input parameters. Total infections, total casualties, percent infections, and percent casualties provided absolute numbers of occupants that had health adverse consequences from contaminant exposure. These measures were easy to compare between cases. Fraction of building protection, another threat-based measure, were straightforward to compute and also easy to compare; however, it lacked ability to differentiate risk in some scenarios especially one with uniform air distribution. A number of vulnerability-based measures (i.e., nondimensional building average concentration, transient concentration, and improvement score) were not complicated to obtain compared to the threat-based measures. They were suitable in identifying risk in scenarios with uniform contaminant distribution but different mechanical system characteristics. On the other hand, nondimensional occupancy dose curves and decile breakdown of nondimensional exposure dose indicated risk levels accurately for scenarios with uniform and non-uniform airborne contaminants. The nondimensional occupancy dose curve included only results in occupied space; while the decile breakdown of nondimensional exposure dose covered the contaminant distribution for the entire building. Based on results from both measures, the impact of different mechanical system and release characteristics could be evaluated in a building attack event. A higher efficiency filter provides better protection than the less efficient one. For a release into an air handler, the DOAS system was worse than the CAV and VAV systems due to much lower air change rates in contaminated spaces. However, for an interior release event, a DOAS system was much more protective than the CAV and VAV systems because of its ability to limit the spread of contaminant from recirculation. Also, the zoning strategy could contain the contaminants within zones connected to the mechanical system that the release occurred. Nonetheless, zoning created more possible release scenarios. A winter stack effect could reverse air direction in elevator shaft and lower air change rates in stairwell; its effect on severity level of building occupant were found to be minimal for this building. In conclusion, the research evaluated each measure and discussed its application and usefulness. Sensitivity analysis indicated that uncertainties for most input parameters affected results in varying degree depending on the mechanical system and type of results. The confidence in the mitigation measure depends on the confidence in the risk assessment results. Future work would embody the development of risk assessment methodology using building-specific performance measure as a part of design procedure.