The capability and process characteristics of ultrafiltration membrane technology treating soy whey wastewater were investigated by using regenerated fiber membrane (RC) and PES membrane (PES) with molecular weight cutoff (MWCO) of 5000

10000 and 30000

respectively.The permeation flux

membrane attenuation coefficient

rejection rate of protein

carbohydrate permeation rate of all kinds of membranes were studied through the test.The results showed that by comprehensively considering flux

decline factor

protein rejection coefficient and carbohydrates permeation rate

RC-10000 was the suitable membrane for separating soybean whey.The experimental results demonstrated that the optimal operation pressure of ultrafiltration membrane was 30 kPa

the optimal pH was 9

concentration ratio was 3∶1

and temperature was 20℃.The permeation flux was 31.2 L·m-2·h-1 and rejection rate of protein and protein content were 78.46% and 51.37%

respectively.In order to optimize operation parameters and predict the permeation flux of the ultrafiltration membrane for soy whey wastewater treatment

a simulation model of operation parameters was established based on curve fitting and back-propagation neural network (BPNN).Simulated results by the simulation model analysis were basically consistent with experimental values.This indicated that the simulation model on the basis of curve fitting and BPNN was practical.