Optimasi Parameter Otomatis Berbasis Neuro-Fuzzy Hibrida pada Pemodelan Adveksi–Difusi Sistem Reverse Osmosis

Muchsin Harahap; Dewi Sholeha; Hery Andi Sitompul

doi:10.59086/jti.v5i1.1357

Authors

Muchsin Harahap Program Studi Pendidikan Teknik Elektro Universitas Negeri Medan
Dewi Sholeha Program Studi Teknik Elektro Universitas Darma Agung
Hery Andi Sitompul Program Studi Teknik Elektro Universitas HKBP Nommensen Medan

DOI:

https://doi.org/10.59086/jti.v5i1.1357

Abstract

Pemodelan numerik berbasis persamaan diferensial parsial banyak digunakan pada sistem reverse osmosis (RO) untuk merepresentasikan fenomena transportasi massa di dalam membran. Persamaan adveksi–difusi umum digunakan, namun akurasi solusi numeriknya sangat bergantung pada pemilihan parameter fisik dan numerik yang umumnya ditentukan secara statis, sehingga kurang adaptif terhadap perubahan kondisi operasi. Penelitian ini mengusulkan kerangka kerja optimasi parameter otomatis berbasis Hybrid Neuro-Fuzzy untuk pemodelan adveksi–difusi pada sistem reverse osmosis. Sistem RO diperlakukan sebagai sistem dinamis, di mana parameter kecepatan aliran, koefisien difusi, dan batas integrasi berperan sebagai variabel yang sensitif terhadap pengendalian. Adaptive Neuro-Fuzzy Inference System (ANFIS) diterapkan sebagai pengendali adaptif dalam mekanisme loop tertutup, dengan galat solusi numerik digunakan sebagai sinyal umpan balik untuk memperbarui parameter secara dinamis. Hasil simulasi menunjukkan bahwa pendekatan yang diusulkan mampu menurunkan galat numerik, mempercepat konvergensi, dan meningkatkan kestabilan sistem dibandingkan metode statis. Pendekatan ini berkontribusi pada pengembangan kontrol adaptif dan rekayasa komputasi untuk sistem reverse osmosis yang cerdas dan otomatis.

Numerical modeling based on partial differential equations is widely used in reverse osmosis (RO) systems to represent the mass transport phenomenon in the membrane. The advection–diffusion equation is commonly used, but the accuracy of its numerical solution is highly dependent on the selection of physical and numerical parameters which are generally determined statically, making it less adaptive to changes in operating conditions. This study proposes a Hybrid Neuro-Fuzzy based automatic parameter optimization framework for advection–diffusion modeling in reverse osmosis systems. The RO system is treated as a dynamic system, where the flow rate, diffusion coefficient, and integration limits act as control-sensitive variables. Adaptive Neuro-Fuzzy Inference System (ANFIS) is implemented as an adaptive controller in a closed-loop mechanism, with the error of the numerical solution used as a feedback signal to dynamically update the parameters. Simulation results show that the proposed approach is able to reduce numerical errors, accelerate convergence, and improve system stability compared to static methods. This approach contributes to the development of adaptive control and computational engineering for intelligent and automated reverse osmosis systems.

References

Biesheuvel, P. M., Porada, S., Elimelech, M., & Dykstra, J. E. (2022). Tutorial review of reverse osmosis and electrodialysis. In Journal of Membrane Science (Vol. 647). https://doi.org/10.1016/j.memsci.2021.120221

Fatimah Zahra Al-Gadri, P., Kusumawati, W. I., Harianto, -, & Musayyanah, -. (2023). Penerapan Controlling Auto Light Dimmer Menggunakan Fuzzy Logic Pada Hidroponik Indoor. Techno.Com, 22(2). https://doi.org/10.33633/tc.v22i2.7799

Haefner, M. W., & Haji, M. N. (2023). Integrated Pumped Hydro Reverse Osmosis System optimization featuring surrogate model development in Reverse Osmosis modeling. Applied Energy, 352. https://doi.org/10.1016/j.apenergy.2023.121812

Kumari, S., Agarwal, S., Kumar, M., Sharma, P., Kumar, A., Hashem, A., Alotaibi, N. H., Abd-Allah, E. F., & Garg, M. C. (2025). An exploration of RSM, ANN, and ANFIS models for methylene blue dye adsorption using Oryza sativa straw biomass: a comparative approach. Scientific Reports, 15(1). https://doi.org/10.1038/s41598-025-87274-3

Maftouh, A., Fatni, O. El, Echchikhi, A., & Rhaleb, H. El. (2023). Experimental and Mathematical Modelling of Reverse Osmosis System to Investigate the Impact of Feed Water Pressure: a Case Study. International Review of Civil Engineering, 14(1). https://doi.org/10.15866/irece.v14i1.21621

Nugraha, Y. T., Cahyadi, C. I., Rida, R., Ningsih, M. S., Sholeha, D., & Roza, I. (2025). Application of the adaptive neuro-fuzzy inference system for prediction of the electrical energy production in Jakarta. IAES International Journal of Artificial Intelligence, 14(3), 1790–1798. https://doi.org/10.11591/ijai.v14.i3.pp1790-1798

Saeed, A., & Alhawaj, M. (2024). Mathematical modeling of reverse osmosis system design and performance. Water Practice and Technology, 19(7). https://doi.org/10.2166/wpt.2024.141

Sholeha, D., & Zambak, M. F. (2022). The Implementasi ANFIS Dalam Prakiraan Perkembangan Energi Baru dan Terbarukan di Indonesia Pada Tahun 2030. Jurnal Sistem Informasi …, 5(2).

Sitompul, H. A., & Sholeha, D. (2025). Integrasi Romberg dan Monte Carlo, Dua Pendekatan Berbeda Untuk Solusi Sistem Reverse Osmosis Pada Media Berpori. In Jurnal Teknologi dan Informas (Vol. 4, Number 3).

Sitompul, Hery Andi, Tarigan, K., Sholeha, D., & Sitompul, Hery Andy. (2025). Impression: Jurnal Teknologi dan Informasi Monte Carlo Implementation on Bisection and Regula Falsi Methods for Finding Multiple Roots of Polynomial and Exponential Equations. In Jurnal Teknologi dan Informas (Vol. 4, Number 2).

Yadav, G. K., Kirar, M. K., Gupta, S. C., & Rajender, J. (2025). Integrating ANN and ANFIS for effective fault detection and location in modern power grid. Science and Technology for Energy Transition (STET), 80. https://doi.org/10.2516/stet/2025013