Optimisation of ultrafiltration pilot plants requires a better knowled
ge of membrane fouling. Zn the field of drinking water production, phe
nomena involved in fouling are very complex and interdependent because
of the numerous compounds contained in raw waters. As no knowledge mo
del is available for this application, a statistical modelling tool ca
lled neural network is used in this paper to predict the total hydraul
ic resistance at the end of a filtration cycle and after next backwash
, using some parameters concerning water quality (turbidity and temper
ature) and operating conditions, for a given experimental site. Differ
ent network structures have been evaluated, using information concerni
ng the current filtration cycle and the previous cycle. Some of them a
llow a prediction of resistance with a very good accuracy. They take i
nto account as network inlets the permeate flow rate, pressure and wat
er turbidity, and are able to model the effects of reversible fouling
on resistance.