Sustania, Sinta (2026) Rancang Bangun Sistem Kontrol Suhu, PH, Electroconductivity, dan Dissolved Oxygen pada Smart Farming System untuk Tanaman Hidroponik Cabai Rawit Berbasis Kontrol Interval Tipe-2 ANFIS. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5009221076-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5009221076-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy |
Abstract
Fluktuasi harga cabai rawit (Capsicum frutescens) di Indonesia yang dipicu ketidakstabilan produksi mendorong kebutuhan sistem budidaya presisi yang mampu menjaga parameter larutan nutrisi tetap stabil. Penelitian ini merancang dan mengimplementasikan sistem kontrol interval tipe-2 Adaptive Neuro-Fuzzy Inference System (IT2-ANFIS) pada mikrokontroler ESP32 untuk mengendalikan suhu, pH yang dipisah menjadi sub-kontroler pH Up dan pH Down, electrical conductivity (EC), dan dissolved oxygen (DO) pada sistem hidroponik Nutrient film technique (NFT), dengan ANFIS Tipe-1 sebagai baseline pembanding. Parameter perancangan ditentukan melalui karakterisasi sensor, aktuator, dan plant, menghasilkan R² di atas 0,998 pada seluruh kurva kalibrasi sensor dengan error maksimum 3,29% pada sensor DO, serta persamaan regresi linear untuk keempat pompa peristaltik, aerator, dan kipas dengan R² di atas 0,95. Parameter IT2-ANFIS dioptimasi melalui nested cross-validation, menghasilkan konfigurasi Footprint of Uncertainty (FOU) yang bervariasi antar sub-sistem, tersempit pada EC dan DO (0,1) dan terlebar pada pH Up dan suhu (0,7). Dibandingkan ANFIS Tipe-1, IT2-ANFIS mencatat penurunan RMSE data uji paling signifikan pada DO, 2,39% menjadi 1,23%, dengan pH Up menjadi satu-satunya sub-kontroler yang performanya sedikit lebih baik pada ANFIS Tipe-1. Pengujian closed-loop menunjukkan ANFIS Tipe-1 gagal mencapai set point pada DO akibat pola PWM hunting, dengan error steady state 9,48%, sementara IT2-ANFIS konvergen tepat pada set point DO dengan error steady state 0,00%. Pada suhu, kedua arsitektur sama-sama tidak mencapai set point 27,5°C secara tepat, dengan error steady state 3,64% pada ANFIS Tipe-1 dan 1,82% pada IT2-ANFIS. Pada pH Down dan EC, IT2-ANFIS mencatat error steady state dan RMSE closed-loop yang lebih baik, dengan settling time pH Down kurang dari separuh dan penyimpangan pasca-settling dua puluh kali lebih kecil dibanding ANFIS Tipe-1. Pengecualian ditemukan pada pH Up: ANFIS Tipe-1 mencatat delay time, settling time, dan RMSE closed-loop yang lebih baik, sementara IT2-ANFIS unggul pada rise time dan error steady state. Pengujian lanjutan pada pertumbuhan tanaman menunjukkan tren pertambahan tinggi dan jumlah daun yang lebih baik pada kelompok dengan kontrol IT2-ANFIS dibanding kelompok pembanding.
=====================================================================================================================================
Price fluctuations of cayenne pepper (Capsicum frutescens) in Indonesia, driven by unstable production, have created a need for precision cultivation systems capable of maintaining stable nutrient solution parameters. This research designs and implements an interval type-2 Adaptive Neuro-Fuzzy Inference System (IT2-ANFIS) control system on an ESP32 microcontroller to regulate temperature, pH (split into pH Up and pH Down sub-controllers), electrical conductivity (EC), and dissolved oxygen (DO) in a Nutrient film technique (NFT) hydroponic system, with Type-1 ANFIS as the comparison baseline. Design parameters were determined through sensor, actuator, and plant characterization, yielding R² above 0.998 across all sensor calibration curves with a maximum error of 3.29% on the DO sensor, along with linear regression equations for all four peristaltic pumps, the aerator, and the fan, each with R² above 0.95. IT2-ANFIS parameters were optimized through nested cross-validation, producing a Footprint of Uncertainty (FOU) configuration that varied across sub-systems, narrowest for EC and DO at 0.1 and widest for pH Up and temperature at 0.7. Compared to ANFIS Tipe-1, IT2-ANFIS recorded the most significant test RMSE reduction on DO, from 2.39% to 1.23%, with pH Up being the only sub-controller where ANFIS Tipe-1 performed slightly better. Closed-loop testing showed ANFIS Tipe-1 failed to reach set point on DO due to PWM hunting, with a steady-state error of 9.48%, while IT2-ANFIS converged precisely on the DO set point with a steady-state error of 0.00%. On temperature, neither architecture precisely reached the 27.5°C set point, recording steady-state errors of 3.64% for ANFIS Tipe-1 and 1.82% for IT2-ANFIS. On pH Down and EC, IT2-ANFIS recorded better steady-state error and closed-loop RMSE, with pH Down settling time less than half and post-settling deviation twenty times smaller than ANFIS Tipe-1. An exception was found on pH Up, where ANFIS Tipe-1 recorded better delay time, settling time, and closed-loop RMSE, while IT2-ANFIS was superior in rise time and steady-state error. Follow-up testing on plant growth showed a favorable trend in height and leaf count gain for the IT2-ANFIS-controlled group compared to the comparison group.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Cabai rawit, hidroponik NFT, interval tipe-2 ANFIS, sinyal kendali, Cayenne pepper, NFT hydroponics, interval type-2 ANFIS, control signal |
| Subjects: | S Agriculture > SB Plant culture T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK3070 Automatic control |
| Divisions: | Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis |
| Depositing User: | Sinta Sustania |
| Date Deposited: | 01 Aug 2026 06:44 |
| Last Modified: | 01 Aug 2026 06:44 |
| URI: | http://repository.its.ac.id/id/eprint/141063 |
Actions (login required)
 |
View Item |