Sulphur Recovery Unit Modelling via Stacked Neural Network

M.G. Xibilia and N. Barbalace (Italy)


Neural networks, stacked neural networks, refineries, monitoring, product quality, soft sensor.


In the paper non linear neural modelling of the acid gases hydrogen sulfide (H2S) and sulphur dioxide (SO2) in the tail stream of a Sulphur Recovery Unit (SRU) in a refinery located in Sicily, Italy, is described. In particular Stacked Neural Networks are used to improve the accuracy of the model. Classical and linear stacking approaches based on simple average, the least square approach and principal component regression are compared with a new non-linear combination strategy based on a cascade of neural networks. The obtained model is currently being implemented in the refinery in order to replace the measurement device during maintenance and guarantee continuity in the monitoring and control of the plant.

Important Links:

Go Back