A multi-resolution wavelet neural network approach for fouling resistance forecasting of a plate heat exchanger

The hidden neurons in dashed and dotted lines are mapped to the resolution of 1 to J spaces whose activation functions are wavelet functions. This is the way to approximate the aimed function from contour to details. If the simulation result is beyond the accuracy requirement, we continue to add a squad of neurons with higher resolution into the network. When the scale parameter of J is obtained, the n basic neurons with scaling function and the added n + 2n + 22n + …2J−1n neurons with a wavelet function in different resolution spaces will approximate the aimed function simultaneously.