Naslov
Development of Soft Sensor for Diesel Fuel Quality Estimation

Autori
Bolf, Nenad ; Galinec, Goran ; Baksa, Tomislav

Izvornik
Chemical Engineering & Technology (0930-7516) 33 (2010), 3; 405-413

Vrsta, podvrsta i kategorija rada
Radovi u časopisima, članak, znanstveni

Ključne riječi
soft sensor; neural network; process modelling; crude distillation unit; diesel fuel

Sažetak
Virtual software sensor is developed for quality estimation of diesel fuel as the crude distillation column product. Due to the growing norms for the fuel quality and needs to produce various gradations of quality with different feeds, often laboratory testing and quality controls of the products are necessary. On the basis of available continuous temperature measuring and flows of appropriate process streams, software sensors for estimating cold filter plugging point of diesel fuel have been developed. Nonlinear software sensor models have been developed with the best results achieved using neural networks. Statistical data analysis has been carried out and the results were critically examined. The results of the research show possibilities of applying software sensors for refinery product quality estimation of diesel fuel as an alternative for laboratory testing. In that way it is possible to estimate the quality of fuel continuously, as well as to apply the methods of inferential control for optimization of the plant operation.

Izvorni jezik
Engleski

Znanstvena područja
Kemijsko inženjerstvo

POVEZANOST RADA