#TITLE_ALTERNATIVE#

• Main Author: MEIRNADIAS (NIM : 13311003) - SUGENG SANTOSO (NIM : 13311008), TYFFANI
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/31393
• Institution: Institut Teknologi Bandung

<br /> <p align="justify"> In industrial process, it often meets a condition of fluid mixing with different temperature which interfering in more than one tank, like miniplant in Instrumentation and Control Management System Laboratory, Engineering Physics ITB, which has a miniplant flow series connected to industrial process control device, Distributed Control System (DCS). This research creates a condition of fluids mixing with different temperature in order to make it the same with the desired condition based on manipulated variable, in this case, the manipulated variable is the main valve opening an outlet of hot water tank (CV-01). In order to improve the response of the process performance, a control by using PID algorithm is conducted. <br /> <br /> However, PID algorithm’s response has a dissatisfying results such as its slow response, and long delay time. Therefore, an alternate design of compensator with Smith-Predictor algorithm is conducted. Based on controller without compensator and controller using compensator implementations, the IAE plant without compensator value is gotten in sequences, 32,03; 30,90; 22,18 and the IAE plant with compensator is 18,80. The best response result is shown by Smith-Predictor controller, with the lowest IAE value, fastest response without overshoot with 250 seconds settling time. On the set point tracking test, Smith-Predictor has a better performance than Lambda Tuning, it can approach a better set point in every input change. Moreover, in disturbance rejection on Smith-Predictor, the fastest relief time after disturbance as fast as 210 seconds is gotten. Meanwhile, in ±50% Delay Time Error Model Test, it decreases settling time become 32% lower than the optimum settling time, and it shows a higher overshoot on the 0,5 L delay time error test compared with the 1,5 L delay time error test. <br /> <p align="justify">