Applying Neural Network in Evaporative Cooler Performance Prediction

The back-propagation （BP） neural network is created to predict the performance of a direct evaporative cooling （DEC） air conditioner with GLASdek pads. The experiment data about the performance of the DEC air conditioner are obtained. Some experiment data are used to train the network until these data can approximate a function, then, simulate the network with the remanent data. The predicted result shows satisfying effects.