Air conditioning in a hospital room must be very important to prevent infection. To design the air conditioning system, it is necessary to measure air flow distributions all over the room. The whole flow field measurement, however, is very difficult in a real space such as a hospital room. Air flow distributions are usually measured with some sensors whose measurement points are limited in a large space. It is required, therefore, to estimate an entire flow velocity vector field from limited measured vectors. One promising approach for the estimation is to make an appropriate model of the field by using measured information. In this paper the entire velocity vector distribution is reconstructed using an artificial neural network from the measured data. The concentration distribution in the room is calculated from the estimated velocity vectors and then is compared with the concentration data measured by using smoke.
