This works focuses on using neural networks and expert systems to control a
gas/solid sorption chilling machine. In such systems, the cold production
changes cyclically with time due to the batchwise operation of the gas/soli
d reactors. The accurate simulation of the dynamic performance of the chill
ing machine has proven to be difficult for standard computers when using de
terministic models. Additionally, some model parameters dynamically change
with the reaction advancement. A new modelling approach is presented here t
o simulate the performance of such systems using neural networks. The backp
ropagation learning rule and the sigmoid transfer function have been applie
d in feedforward, full connected, single hidden layer neural networks. Over
all control of this system is divided in three blocks: control of the machi
ne stages, prediction of the machine performance and fault diagnosis. (C) 1
998 Published by Elsevier Science Ltd and IIR. All rights reserved.