Mold growth modeling of building structures using sensitivity classes of materials

TLDR: Ojanen et al. as mentioned in this paper presented the latest findings of mold growth and the modeling of these factors on different materials, such as pine and spruce sapwood, by using the dynamic temperature and relative humidity histories of the subjected material surfaces.

Abstract: Numerical simulation of mold growth can be used as one of the hygrothermal performance criteria of building structures. Mold growth is one of the first signs of too-high moisture content of materials, and it may affect the indoor air quality and also the appearance of the visible surfaces. Mold growth potential can be predicted by solving a numerical value, mold index, by using the dynamic temperature and relative humidity histories of the subjected material surfaces. The model was originally based on mold growth of wooden materials, but it has now been completed with several other building materials. The model can be used parallel with heat, air, and moisture simulation models or as a post-processing tool. This paper presents the latest findings of mold growth and the modeling of these factors on different materials. The mold growth model has been improved by taking into account the effect of seasonal, long dry or cold periods that do not allow growth. This includes mechanisms for the decrease of mold level (decline of mold index) during unfavorable growth periods and the intensity of the growth after these periods. The laboratory and field results show that the sensitivity of the mold index level may vary in a large range depending on materials. Also, the performance on the interface of two materials has been studied. Instead of modeling the performance separately for each material or product, the materials are presented as different mold sensitivity classes varying from resistant to very sensitive. The sensitive class corresponds to the performance of pine sapwood, which was one basic material in the original model format. Other materials are presented by using the detected correlations between these materials. The mold growth sensitivity classes, decline of the growth level, comparison to detected mold level in materials, and numerical application in practical hygrothermal performance analysis are presented and discussed. MOTIVATION AND OBJECTIVES FOR FURTHER DEVELOPMENT OF THE MOLD GROWTH MODEL Numerical simulation of heat, air, and moisture performance of building structures generates the prediction of hygrothermal conditions in different parts of the analyzed structure. Also, monitoring of laboratory experiments and site investigations produces large amounts of data from critical parts of structures. This data should be post processed in order to evaluate the risks connected to overall performance, service life, interaction with indoor climate conditions, and structural safety. Mold growth is one of the first signs of biological deterioration caused by excess moisture; therefore, mold growth can be used as one of the best hygrothermal performance criteria of building structures. Mold does not deteriorate the material, but it is a sign of too-high moisture content and it represents a risk for other moisture-caused problems, such as decay. Mold affects the appearance of the surface and can severely affect the indoor air quality when the growth is in contact with indoor air or with the leakage air flowing into the room space. The mathematical model of mold growth was developed by Hukka and Viitanen (1999) based on regression analysis of the measured data (Viitanen 1996; Viitanen and Ritschkoff 1991) for calculating the development of mold growth, which is expressed as the mold index. An index value from 0 to 6 is defined to describe the evaluation of mold growth on a surface © 2010 ASHRAE. Tuomo Ojanen is a senior research scientist and team leader at VTT Expert Services Ltd, Finland. Hannu Viitanen is a senior research scientist and team leader at VTT Bioprocessing, Finland. Ruut Peuhkuri worked as a research scientist at VTT during the research and is currently a senior consultant with Passivhus.dk ApS, Næstved, Finland. Kimmo Lähdesmäki and Kati Salminen are research scientists and Juha Vinha is a docent in the Department of Civil Engineering at Tampere University of Technology, Tampere, Finland. a) on a microsopic level (1– 2) and b) when the growth can be detected visually (3–6). This mold index is based on the detectable growth of different mixed mold species. The first version of this model was based on a great number of measurements on pine and spruce sapwood material. This model has been used to analyze (in parallel or in post processing) the result derived from numerical simulation models for the dynamic temperature and relative humidity histories of the critical material surfaces. The mold growth risk analysis based on sensitive wooden materials has been applied also for different material layers that have soiled, dusty surfaces and those surfaces having contact with wood-based materials. Since the first version of the model, the research has included several experimental studies on conditions for mold growth, primarily on wood but also on other building materials. In order to predict the risks of mold growth in varying types of structures made of several building products and materials, it is obvious that an improved model to cover several typical building materials has to be developed. RESEARCH CARRIED OUT TO IMPROVE THE MODEL A three-year research project was carried out at Technical Research Center of Finland (VTT) and Tampere University of Technology. This project included large sets of steady-state and dynamic laboratory experiments for common building materials (Salminen et al. 2009), monitoring of mold growth in material surfaces and structures under real climate conditions, and long-period climate chamber experiments. The results of these findings were used to improve the existing numerical model for mold growth. This paper presents the development of the mold growth model in this project (Peuhkuri et al. 2009; Ojanen et al. 2009), which items were taken into account and how these parameters were studied, and the results interpreted numerically for different materials and conditions. The experiments and their findings are presented and discussed only from the modeling aspect. These results are presented in a concise way, and the main findings are shown as the improvements of the numerical model. The emphasis is on the comparison between experiments and the outcome of the new modeling principles. MATERIALS USED IN THE RESEARCH Some typical building materials were chosen for the experiments: spruce board (with glued edges), concrete (K30, maximum grain size 8 mm), aerated concrete, cellular concrete, polyurethane thermal insulation (PUR, with paper surface and with polished surface), glass wool, polyester wool, and expanded polystyrene (EPS). Pine sapwood was used as a reference material. This set of products cannot entirely represent all the products in the building material group, but it gives improved approximation on the mold growth sensitivity of each. The following results are based on the controlled laboratory and well-monitored site experiments of the chosen materials and structures where these products were used. UPDATING THE EXISTING NUMERICAL MOLD GROWTH MODEL The mold growth model based on experiments with wood was updated to be valid also for the mold growth prediction of other building materials. The idea in this research was to keep the original model structure and to adapt the mold growth parameter values of different materials to the existing model. Some improvements were applied for the model structure to better adjust different growth phenomena. The following sections represent the modeling principles for different mold growth parameters. MOLD GROWTH LEVEL—MOLD INDEX Determination of the mold growth levels is the fundamental element of the whole simulation of this biological phenomenon. This determination sets an interpretation of the visual growth levels as numerical values. This is needed both in the evaluation of the experimental results and in the assessment of the simulation results. Figure 1 represents how mold growth was studied under constant conditions for this research. Closed containers had saturated salt solution vessels to maintain known constant humidity levels. There were nine test samples of each material used in the tests. Some focusing was done to better take into account the different mold growth types with different materials and surfaces. The main difference compared to the version for wood-based materials was in the area that is not visible to the naked eye. It was found out that with some materials the mold growth coverage could be quite high already in microscopic areas (see Figure 2). Therefore, the mold index Figure 1 Laboratory test setup with small samples; there were nine samples of each material.