Comparison of PSO-based PI and FOPI performances in PEM fuel cells under dynamic and static load

Fuel cells are clean, highly efficient energy technologies that do not produce waste products that pollute the environment. Many kinds of fuel cells are being researched, and hydrogen is frequently used as the fuel. One of these is Proton Exchange Membrane Fuel Cells (PEMFC) and its important features are; It is compact and lightweight, has high output power density at low temperatures, low environmental impact and good start-up/shutdown performance. However, problems such as substandard performance, uncontrolled voltage and high fuel consumption are parameters that prevent the widespread use of PEMFCs in commercial applications. Therefore, it seems that an appropriate control design is required in PEMFCs. With this motivation, in this thesis study, it is aimed to improve the system performance by using traditional PI / PSO based PI / PSO based FOPI controllers in DC-DC converters within the scope of two different PEMFC designs: dynamic and constant load. In the PEMFC system, which works under both dynamic and static load, a 6 kW fuel cell was used and Hydrogen consumption [L/min] and oxygen air flow [m3/s] were taken as the input parameters of the system. In order to improve the performance of the PEMFC model under static load, three different applications were implemented: traditional PI, PSO-based PI and PSO-based FOPI. PSO application: 50 particles were used, and after ten repetitions, the desired results were attained. Based on this, it was found that Kp = 0.0084 and Ki = 0.5 were the optimal values produced in the application with the lowest cost. Looking at the variance and standard deviation values, it was observed that the lowest oscillation was obtained with the values of Kp = 0.00815 and Ki = 0.4623, and the optimum values determined by the PSO method were observed to be more stable than the reference change. Three distinct applications—traditional PI, PSO-based PI, and PSO-based FOPI—were implemented in order to enhance the PEMFC model's performance under static load. In comparing the performances of the methods; The largest overshoot, rise time and settling time were taken as criteria. In the traditional PI model, PI values were obtained as Kp = 0.001 and Ki = 0.15 with the empirical approach. In the PSO-based PI model, the coefficients are determined as (Kp = 0.2042, Ki = 1). For PSO, the model was tested with different particle numbers and it was determined that the best result was obtained with 20 particles. Finally, in the PSO-based FOPI application, PIλ controller coefficients were determined as (Kp = 0.7753, Ki = 0.3650 and λ = 0.2050). According to the findings, it has been observed that the performance obtained with the proposed PSO-based PI and PSO-based FOPI approaches is better than the performance obtained with the traditional PI approach. Keywords: Hydrogen energy, PEM fuel cell, Dynamic/Static load; PSO, PI controller,FOPI