Hayuningsih, Prihatsiwi (2017) Perancangan Adaptive Neuro-Fuzzy Inference System (ANFIS) untuk Pengendalian Level pada Kolom Distilasi Biner. Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
2413100089-Undergraduate_Theses.pdf Download (4MB) | Preview |
Abstract
Kolom distilasi biner berfungsi untuk memisahkan dua komponen penyusun campuran berdasarkan perbedaan volatilitasnya. Pada pengendalian kolom distilasi konfigurasi DV, yang dimanipulasi untuk mendapatkan komposisi produk yang diinginkan adalah flow refluks dan flow produk bawah. Pada tugas akhir ini dilakukan pengendalian pada level dasar kolom distilasi dengan variabel manipulasi flow produk bawah. Data yang digunakan diambil dari data operasi Ammonia Stripper 105-E PT Petrokimia Gresik. Pengendalian dilakukan dengan adaptive neuro fuzzy inference system atau ANFIS yang mampu menghasilkan sinyal pengendalian dari mekanisme pembelajaran adaptif untuk mendapatkan parameter premis dan konsekuen dari sistem penalaran pengendali fuzzy. Dilakukan variasi pada tipe dan jumlah fungsi keanggotaan fuzzy, yaitu jumlah fungsi keanggotaan 3, 5, dan 7 pada masing-masing tipe keanggotaan segitiga, trapesium, generalized bell, dan gaussian. Dari analisa didapatkan hasil yang paling baik adalah pengendali ANFIS generalized bell dengan 3 fungsi keanggotaan. Pengendali ini menghasilkan respon dengan maximum overshoot sebesar 2.358%, settling time 66.790 detik.
==============================================================================================================================
Binary distillation column is used to separate two liquid mixture component based on the difference of component’s volatility. On a DV configuration distillation column control, reflux flow and bottom product flow is manipulated to control the product composition. In this final project, a control system for base level column with bottom product flow as manipulated variable is designed. Data used on this final project is taken from operational data of Ammonia Stripper 105-E in PT Petrokimia Gresik. The control system is adaptive neuro fuzzy inference system or ANFIS which is capable to produce control signal from adaptive learning mechanism to obtain premise and consequent parameter of fuzzy inference system. Variation of type and number of fuzzy membership function is used, they are 3, 5, and 7 membership functions on each triangular, trapezoid, generalized bell, and gaussian type. It is found that the best ANFIS controller for the system is 3 generalized bell membership function. This controller produce a response with 2.358% maximum overshoot and 66.790 seconds of settling time.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | adaptive neuro fuzzy inference system, distillation column, fuzzy membership function |
| Subjects: | T Technology > TJ Mechanical engineering and machinery > TJ217.6 Predictive Control T Technology > TJ Mechanical engineering and machinery > TJ223 PID controllers |
| Divisions: | Faculty of Industrial Technology > Physics Engineering > 30201-(S1) Undergraduate Thesis |
| Depositing User: | Prihatsiwi Hayuningsih |
| Date Deposited: | 14 Nov 2024 05:59 |
| Last Modified: | 14 Nov 2024 05:59 |
| URI: | http://repository.its.ac.id/id/eprint/46362 |
Actions (login required)
 |
View Item |