Rancang Bangun Adaptive Pid Control System Berbasis Artificial Neural Network Sebagai Sistem Kendali Temperatur Dalam Proses Elektrolisis Air

Febriansyah, Ardy Diva (2026) Rancang Bangun Adaptive Pid Control System Berbasis Artificial Neural Network Sebagai Sistem Kendali Temperatur Dalam Proses Elektrolisis Air. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2042221042-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Penelitian ini membahas rancang bangun sistem kendali temperatur pada proses elektrolisis air menggunakan metode Adaptive PID berbasis Artificial Neural Network (ANN) untuk meningkatkan kestabilan dan kinerja sistem dibandingkan PID konvensional. Sistem dikembangkan menggunakan ESP32-S3 sebagai pengendali utama, sensor DS18B20 sebagai umpan balik temperatur, serta modul peltier sebagai aktuator pendingin. Metode PID konvensional dituning menggunakan pendekatan Ziegler–Nichols dan digunakan sebagai pembanding terhadap Adaptive PID berbasis ANN yang melakukan penyesuaian parameter Kp, Ki, dan Kd secara real-time berdasarkan error, set point, dan temperatur aktual. adaptive PID berbasis ANN memberikan respon dinamik yang lebih baik dibandingkan PID konvensional, dengan penurunan settling time dari 638 menit menjadi 408 menit dan penurunan rise time dari 611 menit menjadi 388 menit, serta mampu menekan maksimum overshoot menjadi 2,65%, meskipun error steady-state sedikit meningkat dari 0,1125% menjadi 0,1375%, namun masih berada dalam batas toleransi yang sangat kecil.
=============================================================================================================================================
This study discusses the design of a temperature control system in the water electrolysis process using an Artificial Neural Network (ANN)-based Adaptive PID method to improve system stability and performance compared to conventional PID. The system was developed using ESP32-S3 as the main controller, DS18B20 sensor as temperature feedback, and a Peltier module as a cooling actuator. The conventional PID method was tuned using the Ziegler–Nichols approach and used as a comparison to the ANN-based Adaptive PID, which adjusts the Kp, Ki, and Kd parameters in real-time based on error, set point, and actual temperature. The ANN-based adaptive PID provides a better dynamic response than conventional PID, with a decrease in settling time from 638 minutes s to 408 minutes and a decrease in rise time from 611 minutes to 388 minutes, as well as being able to suppress the maximum overshoot to 2.65%. although the steady-state error increased slightly from 0.1125% to 0.1375%, it is still within a very small tolerance limit.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Elektrolisis air, Adaptive PID, Articial Neural Network, Temperatur, Water electrolysis, Adaptive PID, Articial Neural Network, Temperature
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK3070 Automatic control
Divisions:	Faculty of Vocational > Instrumentation Engineering
Depositing User:	Ardy Diva Febriansyah
Date Deposited:	29 Jul 2026 07:55
Last Modified:	29 Jul 2026 07:55
URI:	http://repository.its.ac.id/id/eprint/139719

Actions (login required)

View Item