Experimental Investigation of the Effect of Natural Convection on Heat Transfer in Mineral Wool

Abstract

In Denmark and a number of other countries convection in fibrous materials is considered non-existent when calculating heat transmission as well as when designing building structures. The current investigation serves to examine whether considering natural convection non-existent in fibrous materials is a fair assumption to use in present and future building structures. In order to facilitate the experimental work, an apparatus with a metering area of 3 × 1 m2 was designed for measuring the effect of convection. With this apparatus it is possible to measure a specimen with thickness ranging from 0.1 to 0.5 m. The contribution of natural convection to the total heat transfer has been experimentally examined for ‘‘perfectly’’ installed fibrous materials of different air-flow permeability and thermal conductivity. To support the interpretation of the measurements, the experimental work has been supplemented with numerical calculations. The measurements and the concurrent computations both show a clear convection-induced redistribution of the heat flow in the material. When measured from the hot side in the apparatus this occurs in the form of an increased heat flow, primarily through the lower part of the vertical structure, and a decreased heat flow primarily through the upper part. The risk of convection causing the total heat flow in the material to increase is primarily related to the highly permeable material (corresponding with low density). In the low permeable material the redistribution of the heat flow has no influence on the total heat flow through the material. However, the redistribution of the convection-induced heat flow is distinct at a material thickness of 0.2 m and a temperature gradient of 20 C across the material. This research has established that when the permeability reaches a certain level, convection is able to increase the total heat flow by more than three percent, which is considered here as a significant level raised above the uncertainty in the measurement, even under the most ideal conditions even at small temperature differences. This means that the commonly accepted assumption that convection is non-existent is invalid.