Abstract
The paper introduces a new approach for designing a data-driven Soft Sensor for a plant, in the presence of an unknown measurement delay. More specifically, Deep Belief Networks are used for determining the Soft Sensor. The latent features, obtained after the unsupervised learning phase, are exploited for estimating the measurement delay. The procedure is applied to the design of a Soft Sensor for a debutanizer, which is a part of a refinery settled in Sicily. Both the procedure, required for the Soft Sensor design and the obtained results are reported in the paper.

This publication has 13 references indexed in Scilit: