Cilj završnog rada jest prikazati utjecaj transportnog kašnjenja (tzv. mrtvog vremena) procesa u sustavima upravljanja, ući u problematiku upravljanja tih sustava, te pronaći rješenje kvalitetnog upravljanja sustava s transportnim kašnjenjem. U radu su korištena dva realna industrijska procesa s transportnim kašnjenjem. Izvršena je analiza tih procesa, primjenjena je jedna od klasičnih metoda upravljanja (PID regulacija), te interpretacija rezultata kroz odzive sustava na skokovitu funkciju i kroz neposredne parametre kakvoće upravljanja. Iskazane su negativnosti klasičnih metoda regulacije nad procesima s transportnim kašnjenjem, te prikazano je rješenje kvalitetnijeg upravljanja sustavom idejom Smithovog prediktora. Realiziran je prediktivni PI regulator (tzv. PPI) na konceptu upravljanja Smithovim prediktorom, također projektirana je njegova varijacija FPPI dodavanjem filtra prvog reda u vanjsku povratnu vezu. Dodavanje filtra PPI regulatoru povećava robusnost sustava. Prikazano je ponašanje različitih vrsta regulatora u slučaju promjenjivog mrtvog vremena procesa. Kao rezultat završni rad kroz detaljnu analizu prikazuje i definira rješenje problematike upravljanja sustavima s transportnim kašnjenjem nad stvarnim industrijskim procesima.

The goals of this dissertation are to show the influence of a transport delay (often referred to as dead time) in control systems, define the issues of controlling such systems and find a quality solution of controlling systems with long dead time. Two actual industrial processes with dead time were used in this research. One of the classic control system methods (PID control) was used for the purpose of analyzing the systems and the results were interpreted through a step response and direct parameters of the quality of regulation. The disadvantages of classic control methods over dead time processes were defined and a better solution of the quality regulation, the idea of the Smith prediction, was shown. The predictive PI controller (so called PPI) based on the concept of the Smith predictor was designed. Additionally, its variation FPPI was designed by adding a first order filter into the outer loop. Adding a filter to the PPI controller improves the system robustness. The behavior of different types of controllers in the case of the variable process dead time was shown. Conducting a detailed analysis, a solution to the problematic actual industrial dead time process system control was shown and defined.