Probabilistic design for evaluation of indoor environment

Abstract: In recent years indoor environment and indoor air quality has been subjected to extensive worldwide research efforts. Still, at present there are a lack of methods for prediction of risks and consequences for any defined damage to occur, similar to the probabilistic methods used in modern design codes for structures and buildings. Such a method, if available and usable, could be very beneficial as a tool for decision making at different stages of the building process. In this licentiate thesis a method is developed to estimate the risk of occurrence of high radon concentrations indoors. The method is developed using risk analysis applied on a concrete slab on the ground, which is a foundation method commonly used in Sweden for residential buildings. The undesirable event is "Leakage of radon into the building" and fault tree analysis is used starting in the top with the undesirable event and then working downwards finding the events, which causes the top event. This procedure continues until basic events areestablished and for which failure probabilities are available. Fault tree analysis is a deductive method mainly used for systems built up with electronic equipment where failure probabilities in terms of relative frequency are easy to establish. Failure probabilities are more difficult to establish in the building process since failures are rare. However, this uncertainty can be handled by applying structural reliability analysis on the quantitative analysis whereas several random variables can be taken into account in a single analysis, which makes it possible to analyse a whole branch of a fault tree in a single analysis. The basis for structural reliability methods is probability theory to handle the uncertainties and Monte Carlo simulation and first-order second-moment theory to estimate the risk. To get an indoor environment that is unhealthy to human beings environmental impact is needed together with fault due to human error in some phase of the building process. Examples of faults made in the building process are changes and/or addition of work ordered by the proprietor, unsuitable design or delivery of wrong material to the construction site. Several surveys have come to the same conclusion that a large amount of faults made in the building process depends on e.g. insufficient commitment or lack of knowledge. However, the building could be designed and constructed according to the state-of-the art of knowledge and still has an improper function. Errors cannot be ruled out completely but the use of fault tree technique to develop the causes to an unhealthy indoor environment and the estimation of risk can be a valuable communication tool to the quality management system to get an overview of the entire building process and to identify the week links. The residential building used as an example in this work is situated outside the municipality of Linköping since long-term measurements of radon concentrations in soil air are available from this area. The building has a self-draught ventilation system and the design of the concrete slab, starting from the excavated rock floor, include a geotextile, 150 mm well washed macadam, 50 mm insulation, 100 mm reinforced concrete, a levelling compound and finally a flooring. The causes to radon concentrations indoor has been developed with fault tree analysis where "Leakage of radon contaminated soil air into the building", "Radon contaminated drinking water" and "Building material contains radium disintegrating to radon" constitute the main causes. The event "Leakage of radon contaminated soil air...." is developed further, since it is the major contributor to radon concentrations indoor, into the events "Lower air pressure indoor than outdoor", "Radon contaminated soil air under or around the building" and "Fault in component with regard to air-tightness". A function expressing the relationship between the basic events and their random variables is established and both Monte Carlo simulation and first-order second-moment theory is applied to estimate the probability and the safety index b for the undesirable event to occur. The leakage of radon-contaminated soil air has only been considered through fissures in the concrete slab and it has been difficult to find proper relationships between how fissures occur, fissure width, and concrete and reinforcement properties. Approximations have therefore become necessary. However, the simulation and the analytical calculation gives e.g. safety index beta = 0.30 on normal risk area, which is 70 % of the total area of Sweden, in residential buildings where people stay more than temporary. Comparison can be made with the safety index beta larger or equal to 4.3 for safety class 2 in the Swedish Design Regulations for the structure of residential buildings. In a residential building on normal risk area and designed in accordance with the concrete slab in this work, the probability for radon concentrations indoor to exceed the Swedish Building Regulations threshold value 200 Bq/m3 is over 40.000 times larger than the risk for structural failure.