Naufa Falah, Faris (2026) Estimasi Remaining Useful Life (Rul) Pada Proton Exchange Membrane Fuel Cell (Pemfc) Open Cathode Berbasis Model Degradasi State Of Health (SOH). Other thesis, Institut Teknologi Sepuluh Nopember.
|
Text
5009211118-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) |
Abstract
Proton Exchange Membrane Fuel Cell (PEMFC) konfigurasi open-cathode banyak digunakan pada sistem daya skala kecil–menengah karena sederhana (tanpa humidifier dan kompresor tekanan tinggi), namun kondisi ini membuat performa stack sangat sensitif terhadap dinamika temperatur dan kelembapan sehingga rentan mengalami gangguan pengeringan membran (drying) dan penumpukan air cair (flooding) yang menurunkan tegangan serta mempercepat degradasi. Tugas akhir ini membahas pengembangan model PEMFC open-cathode berkapasitas 1 kW dengan 60 sel untuk merepresentasikan kondisi baseline, memodelkan dampak drying dan flooding pada rugi tegangan, serta mengestimasi parameter degradasi dan Remaining Useful Life (RUL) menggunakan Extended Kalman Filter (EKF). Tegangan stack dimodelkan dari tegangan reversible (Nernst) dengan rugi aktivasi, ohmik, dan konsentrasi. Drying direpresentasikan menggunakan dryness index (ζ) yang dipetakan menjadi kadar air membran (λ) untuk menghitung konduktivitas proton, sehingga resistansi ohmik meningkat seiring membran mengering. Flooding dimodelkan melalui flooding index (s_L) yang menurunkan limiting current (i_L) akibat hambatan transport massa oksigen, sehingga rugi konsentrasi membesar ketika sistem mendekati batas transport massa. State of Health (SoH) didefinisikan secara eksplisit melalui kenaikan resistansi ohmik dan penurunan limiting current yang dipadatkan dalam indikator degradasi α, sedangkan β merepresentasikan laju degradasi; keduanya diestimasi EKF secara diskrit dengan waktu sampling 0,1 detik. Estimasi RUL dihitung dari proyeksi α menuju ambang akhir umur (αmax) dan dievaluasi pada tiga scenario yaitu normal, drying, dan flooding. Pada kondisi normal didapatkan RMSE 29,04 jam atau sebesar 1.9%, pada kondisi drying didapatkan RMSE sebesar 7,51 jam atau 0.7%, dan pada kondisi flooding dengan pembatasan i_L sebesar 65% didapatkan RMSE sebesar 24,431 jam atau 4,87%. Hasil menunjukkan EKF mampu memberikan estimasi RUL yang stabil pada kondisi normal dan terutama drying, sedangkan akurasi menurun pada flooding akibat sensitivitas rugi konsentrasi.
======================================================================================================================================
Proton Exchange Membrane Fuel Cell (PEMFC) stacks with an open-cathode configuration are widely used in small-to-medium power systems due to their simplicity (no humidifier and no high-pressure compressor). However, this configuration makes stack performance highly sensitive to temperature and humidity dynamics, increasing susceptibility to membrane drying and liquid-water accumulation (flooding), which reduce voltage and accelerate degradation. This thesis develops a 1 kW open-cathode PEMFC model with a 60- cell series stack to represent the baseline condition, model the impacts of drying and flooding on voltage losses, and estimate degradation parameters and Remaining Useful Life (RUL) using a two-state Extended Kalman Filter (EKF). The stack voltage is modeled from the reversible (Nernst) voltage with activation, ohmic, and concentration losses. Drying is represented using a dryness index (ζ) mapped to membrane water content (λ) to compute proton conductivity, such that the ohmic resistance increases as the membrane dries. Flooding is modeled via a flooding index (sL) that reduces the limiting current (iL) due to oxygen mass-transport limitations, thereby increasing concentration losses as the system approaches the mass-transport limit. The State of Health (SoH) is explicitly defined through the increase in ohmic resistance and the decrease in limiting current, condensed into a degradation indicator α, while β represents the degradation rate; both are estimated by the EKF in discrete time with a sampling period of 0.1 s. The RUL is predicted by projecting α toward an End-of-Life threshold (αmax) and evaluated under three scenarios: normal, drying, and flooding. In the normal condition, the RUL estimation yields an RMSE of 29.04 h (1.93%); under drying, the RMSE is 7.51 h (0.7%); and under flooding with iL limitation of 65%, the RMSE is 24.431 h (4.87%). The results indicate that the EKF provides stable RUL estimates under normal operation and, in particular, under drying conditions, whereas accuracy deteriorates under flooding due to the strong nonlinearity of concentration losses.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | drying, EKF, flooding, PEMFC open-cathode , RUL |
| Subjects: | T Technology > T Technology (General) > T57.62 Simulation T Technology > TA Engineering (General). Civil engineering (General) > TA169 Reliability (Engineering) |
| Divisions: | Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis |
| Depositing User: | Faris Naufal Falah |
| Date Deposited: | 03 Feb 2026 05:38 |
| Last Modified: | 03 Feb 2026 05:38 |
| URI: | http://repository.its.ac.id/id/eprint/131884 |
Actions (login required)
![]() |
View Item |
