Napredna pretraga

Pregled bibliografske jedinice broj: 1024030

Estimation of combined cycle power plant power output using multilayer perceptron variations


Lorencin, Ivan; Car, Zlatan; Kudláček, Jan; Mrzljak, Vedran; Anđelić, Nikola; Blažević, Sebastijan
Estimation of combined cycle power plant power output using multilayer perceptron variations // 10th International Technical Conference - Technological Forum 2019 - Proceedings / Kudláček, Jan (ur.).
Prague: Czech Technical University in Prague, 2019. str. 94-98 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)


Naslov
Estimation of combined cycle power plant power output using multilayer perceptron variations

Autori
Lorencin, Ivan ; Car, Zlatan ; Kudláček, Jan ; Mrzljak, Vedran ; Anđelić, Nikola ; Blažević, Sebastijan

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
10th International Technical Conference - Technological Forum 2019 - Proceedings / Kudláček, Jan - Prague : Czech Technical University in Prague, 2019, 94-98

ISBN
978-80-87583-30-2

Skup
10th International Technical Conference - Technological Forum 2019

Mjesto i datum
Hlinsko, Republika Češka, 18-20.06.2019

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
Activation function ; Combined cycle power plant ; Gas turbine ; Multilayer perceptron ; Steam turbine

Sažetak
In this paper an artificial intelligence­based method for output electric power of combined cycle power plant (CCPP) is presented. Method is based on utilizing various designs of multilayer perceptron (MLP). Different MLPs are designed by varying activation functions, solvers and configurations of hidden layers. MLPs are trained and tested by using training dataset of 7500 samples and test dataset of 2068 samples. From obtained results it can be concluded that by using appropriate activation function and solver, low output errors are achieved, regardless of hidden layers. Furthermore, it can be concluded that minimal error is achieved if MLPs designed with Tanh activation function are used.

Izvorni jezik
Engleski

Znanstvena područja
Strojarstvo