• Pergamon PII: 50273-1223(98)00561-7 Wat Sci Tech. Vol. 38. No.6. pp. 1-8.1998. IAWQ C 1998 Published by Elsevier Science Ud. Prillled in Great Britain. All rights rcscrvcd 0273-1223198 $19'00 + 0-00 FRAMEWORKS FOR ASSESSING RELIABILITY OF MULTIPLE, INDEPENDENT BARRIERS IN POTABLE WATER REUSE Charles N. Haas* and R. Rhodes Trussell** * School of Environmental Science. Engineering and Policy. Drexel University. Philadelphia. PA 19104. USA •• Montgomery Watson Engineers. 250 N Madison Avenue. Pasadena, CA 9Jl01·7009. USA ABSTRACf The development of potable water reuse systems (systems for the treatment of wastewater to quality adequate to augment other surface water supplies) requires careful attention to the consistent production of product water which has low levels of contaminants. such as infectious pathogens, capable of causing human health effects from acute exposure. Little consideration has been given to the formal specification of the degree of reliability of such systems. In this paper we present two ways to approach the problem. The first is a formal extension of the 'multiple baIrier' concept often cited in the water treatment literature. The second is an application of probabilistic analysis. With either method. it is clear that much more information should be obtained with respect to the failure modes, and the frequency with which individual processes achieve a particular level of performance. However the conceptual framework presented here should enable a formal analysis of the problem to be conducted. iCI 1998 Published by Elsevier Science Ltd. All rights reserved KEYWORDS Water reuse; reliability; risk; pathogens; barriers. INTRODUCfION A water reclamation plant is an engineered system. Although it has been stated that 'Every water treatment facility must be so designed that, when properly operated, it can produce continuously the design rate of flow and meet the established water quality standards.... (James M. Montgomery Consulting Engineers Inc., 1985. p63S), as an engineered system the potential for out of specification performance always exists. Regardless of the adequacy of design of a set of treatment processes, and the nature of the water or reclaimed wastewater processed. the qUality of the product will exhibit some degree of variability. This is due to variability in the influent stream (e.g., secondary wastewater effluent), as well as variability in performance of the individual process elements. For individual contaminants of public health significance. if the potential health effects are associated only with long-term exposure (such as carcinogens which may produce an effect upon chronic exposure). the presence of temporal variability is relatively unimportant,