Abstract
Numerical experiments of heat loss through the exterior walls have been performed for two simplified space models, while the indoor thermal environmental conditions have been assumed equal for both space models. All the conditions in both space models have been assumed equal other than in one of the surfaces of one of the models temperature supposed to be different from the similar one of the other model. Based on this assumption (i.e., occurrence of temperature difference in some surface), one of the space models is supposed to be warmed up by a radiant heating system and the other by a warm air heating system. In the 128 experiments, in order to obtain the statistical results for heat loss-when the environmental parameters are assumed to change in the practical ranges-the table of orthogonal arrays L_<128>(2^<127>) has been applied. The purpose of this study may be summerized as follows: 1) In the previous paper, authors have proposed an index, so far called c.c.h.t.c. (critical convective heat transfer coefficient), which relates to the heat loss through the exterior walls. Here, the statistical magnitude of error between the results of the rigorous method and approximate formula for c.c.h.t.c. as well as the effects of main factors upon the error have been depicted. 2) The statistical magnitude of the heat loss rate of two space models and the influences of main factors have been shown. In regard to 1), the Gebhart absorption-factors have been applied in the approximate formula, but at meantime, the similar formula has been set in which the absorption-factors are replaced by shape factors. However, it has been found that in the latter formula the error considerably arises. In relation with 2), as heat loss rate changes due to the conditions, application of a calculation method by which the effects of radiant heat on the heat loss can be correctly evaluated is inevitable. Generally, application of simultaneous linear equations in this kind of calculation method for heat loss is necessary, however, in this paper a method called MRT method is proposed in which mean radiant temperatures relative to the walls and to the global object (i.e., represents the human body), are applied in the iterative computation. The iterative computation used in MRT method has shown that less than two times of iteration is necessary to give the correct results. It is possible to simplify the MRT method by simplifying the method used in obtaining the mean radiant temperatures. In this regard, some more simplified methods have been employed, and have also been statistically compared basd on the Experimental Design method.
