IDENTIFICATION OF MRAC MODEL AND CONTROL DESIGN FOR DEVELOPMENT OF A DIGITAL TWIN ADSORPTION SYSTEM FOR LABORATORY SCALE BATIK WASTE TREATMENT

• Main Author: Cahyadi, Hansen
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/81388
• Institution: Institut Teknologi Bandung

Batik industry in Indonesia has great potential from their economic perspective, but the problem of batik liquid waste is still a serious problem. Processing batik liquid waste is necessary to reduce its impact on the environment. One process that can be used to process batik liquid waste is the adsorption process. So far, a laboratory scale waste processing adsorption system is available at the Materials Processing Laboratory, Bandung Institute of Technology. However, a control system and Digital Twin development efforts have not yet been developed for the batik liquid waste adsorption system. Therefore, this research will focus on model identification and control design, as well as the development of a Digital Twin of an adsorption system for processing batik liquid waste at the Material Processing Laboratory, Bandung Institute of Technology using blustru and chitosan adsorbents, as well as a fixed-bed method on the adsorption column. Data acquisition for determining the system model was carried out from ESP32 to ThingSpeak on two channels with ID 2424327 and 2415656 using a turbidity sensor at various variations in pump percentage and adsorbent parameters. This data is used to determine the system model using system identification in MATLAB. The identified model dynamics are in the form of a 1st order system equation with gain (K) and time constant (T1) which change based on differences in input changes in pump percentage and parameters in the form of the amount of adsorbent with a model accuracy of approximately 80.22%. The range of K values obtained is between 1.38 to 3.31, while the T1 value is between 1413.40 to 3293.20 s. MRAC (Model Reference Adaptive Control) control simulation design to follow the reference model with gain and time constant has been carried out using parameters ? and ????0 and the results show that the most consistent parameters ? and ????0 occur in adsorbent parameter 12. A Digital Twin has been built for the system adsorption processing of batik waste on a laboratory scale which is capable of carrying out data acquisition via the internet network to ThingSpeak, determining models, displaying system open loops and reference models, as well as carrying out system control simulations with changes in parameters ? and ????0. Keywords: batik liquid waste adsorption system, system identifiction, MRAC, Digital Twin