Hybrid Predictive- Neural Network Based On Proton Exchange Membrane Fuel Cell Maximum Power Point Tracking

Muhammad Arif Hafizuddin, Ahmad (2023) Hybrid Predictive- Neural Network Based On Proton Exchange Membrane Fuel Cell Maximum Power Point Tracking. [Final Year Project Report] (Unpublished)

	PDF Muhammad Arif Hafizuddin Bin Ahmad (24pgs).pdf Download (403kB)
	PDF (Please get the password by email to repository@unimas.my , or call ext: 3914 / 3942 / 3933) Muhammad Arif Hafizuddin Bin Ahmad (Fulltext).pdf Restricted to Registered users only Download (6MB)

Abstract

Renewable energy is the most efficient and dependable energy source that can be utilised in a variety of sectors, particularly in daily and commercial applications, and can reduce the continuous consumption of fossil fuel. Due to its various benefits, such as being environmentally friendly, highly efficient, noiseless, and safe to run, fuel cell power generation technology is gaining popularity in the present global energy generation, distribution, and consumer demand scene. In this study, PEMFC is chosen due to its promising characteristics, which can be widely applied in daily life and aid in all aspects with the expectation of larger results. Non-linear and greatly impacted by these parameters, the output of this fuel cell is extremely sensitive to cell temperature, oxygen partial pressure, and membrane water content. Each input has a unique operational output that maximises output, and it is best to determine the voltage or current that maximises the fuel cell's efficiency. MPPT methods, including SMC, P&O, INC, and PSO are introduced to achieve maximum production. This project's objective is to simulate a fully functional PEMFC system that is based on mathematical models and combines them with an artificial neural network. This is to improve accuracy and reduce the high computational. This project will implement a PEMFC system coupled with an ANN system. In this project, specific features such as model predictive control (MPC) are being added to provide reliable dataset, and DC-DC boost converters are being utilised to manage the output voltage of fuel cells to extract the maximum output power, with the switches being controlled by MPC. This project's primary objective is to demonstrate that ANN may achieve maximum power output when parameters are emphasised, whereas PEMFC systems will become more efficient in a shorter amount of time. Since the proposed method and other existing MPPT algorithms are compared, the performance of PEMFC is determined., and an additional ANN feature is introduced to support system efficiency. As a result, the ANN system will display rapid MPP locus tracking with exceptional precision and robustness.

Actions (For repository members only: login required)

View Item