Naslov
A Comparison of Several Modeling Methods in Soft Sensor Development

Autori
Grbić, Ratko ; Nyarko, Emmanuel Karlo ; Slišković, Dražen

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Skup
32nd Science in Practice 2014 (SiP 2014)

Mjesto i datum
Osijek, Hrvatska, 15.10.2014. - 17.10.2014

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Nije recenziran

Ključne riječi
soft sensor; difficult-to-measure process variable; process modeling

Sažetak
Many different process variables which are in close relationship with the final product quality cannot always be measured online. In order to enable continuous process monitoring and efficient process control, soft sensors are used to estimate these difficult-to-measure process variables. The heart of the soft sensor is the mathematical model of the process which approximates unknown relationship between easy- to-measure process variables and the difficult- to-measure process variable of interest. Process model building can be based on plant data taken from the process database. In this paper, several well known data based modeling methods are presented. Advantages and disadvantages of each modeling approach are outlined and discussed. The efficiency of the presented methods is shown and compared on a real industrial problem.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo

POVEZANOST RADA