Repository logo
 

Faculty of Engineering and Built Environment

Permanent URI for this communityhttp://ir-dev.dut.ac.za/handle/10321/9

Browse

Has files: YesSubject: search.filters.subject.Computational intelligence

Search Results

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    The role of computational intelligence techniques in the advancements of solar photovoltaic systems for sustainable development : a review
    (Arab Academy for Science, Technology, and Maritime Transport (AASTMT), 2022) Moyo, Ranganai Tawanda; Dewa, Mendon
    The use of computational intelligence (CI) in solar photovoltaic (SPV) systems has been on the rise due to the increasing computational power, advancements in power electronics and the availability of data generation tools. CI techniques play an important role in modelling, sizing, forecasting, optimizing, analysing and predicting the performance and control of SPV systems. Thus, CI techniques have become an essential technology as the energy sector seeks to meet the rapidly increasing demand for clean, cheap, and reliable energy. In this context, this review paper aims to investigate the role of CI techniques in the advancements of SPV systems. The study includes the involvement of CI techniques for parameter identification of solar cells, PV system sizing, maximum power point tracking (MPPT), forecasting, fault detection and diagnosis, inverter control and solar tracking of SPV systems. A performance comparison between CI techniques and conventional methods is also carried out to prove the importance of CI in SPV systems. The findings confirmed the superiority of CI techniques over conventional methods for every application studied and it can be concluded that the continuous improvements and involvement of these techniques can revolutionize the SPV industry and significantly increase the adoption of solar energy.
  • Thumbnail Image
    Item
    Comparative analysis of different computational intelligence techniques for maximum power point tracking of PV systems
    (University of Oradea, 2022-10-01) Moyo, Ranganai Tawanda; Tabakov, Pavel Y.; Moyo, Sibusiso
    The performance of a photovoltaic (PV) module can be improved by employing maximum power point tracking (MPPT) controllers. MPPT controllers are algorithms that are included in PV battery charge controllers or inverters to extract the maximum available power from PV modules for any given temperature and irradiance. Several studies report that the use of PV modules without MPPT controllers results in power losses, which ultimately results in the need to install more solar panels for the same power requirement. Numerous techniques of varying complexities have been proposed in the literature to solve the MPPT objective function. This paper presents a comparative analysis of three computational intelligence (CI) based MPPT techniques namely, the fuzzy logic (FL) based controller, artificial neural networks (ANN) based controller, adaptive neuro-fuzzy inference system (ANFIS) based controller and one conventional technique, the perturbation and observation (P&O) controller. These MPPT controllers are designed, simulated and analysed in the MATLAB/Simulink environment. The performance of the studied MPPT techniques is evaluated under steady-state weather conditions, rapidly changing weather conditions and varying load conditions. CI-based MPPT controllers are found to be more efficient than the P&O controller. Moreover, the ANFIS-based MPPT controller shows an outstanding MPPT performance for all the scenarios studied.
  • Thumbnail Image
    Item
    Evaluation and optimisation of biogas production in municipal wastewater treatment plant using computational intelligence approach : potential to generate electricity
    (2022-05-13) Ramrathan, Zesha; Enitan, Abimbola Motunrayo; Han, Khin Aung
    Development and optimisation of valued added derivatives from wastewater represent the future sustainability paradigm. Among the various challenges in the management of wastewater treatment is the energy consumption for the treatment process that could render this process inefficient in terms of cost and energy consumption. This study focusses on the evaluation of egg-shaped digesters treating municipal wastes, and the optimisation of biogas production using computational intelligence approach (CIA) for sustainable energy production and policy implementation. The study further estimates the amount of electricity that could be generated from the optimised biogas produced from the anaerobic digesters. Historical 5-year (2010-2015) data of the anaerobic digesters were obtained from the Darvill Wastewater Treatment Plant in the KwaZulu-Natal Province of South Africa. The raw data were pre-processed for data cleaning, integration, reduction and data transformations using a rigorous scientific method to test their accuracy, reliability, consistency, and localisation gaps with different multivariate statistical tools. Computation intelligence methods using partial least square (PLS), principal component analysis (PCA) and Fuzzy Logic algorithms were used in this study for simulating the best operational condition and predicting the biogas production. The study further created a contextual framework against the assessment of biogas to energy potential and uses an excel-based tool to determine the bio-economy of energy recovery from an anaerobic egg-shaped digester per cubic meter of treated sludge. In average, the actual methane production was 59.60% while, predicted by Fuzzy-Logic was 65.4%. This shows that the model employed in the improvement of methane production from biogas plants by varying the operational parameters at; Inflow = 590m3 /day, Temp = 32.3°C, pH = 7.12, TS = 3.47%, VS = 43.4% and COD = 510 mg O2/L. The obtained total biogas production was 802.80 m3 /day based on status quo conditions and process configurations. The biogas production translates to electrical energy of 4580.5 KWh/day with an estimated saving (at R1.90 per kWh electricity) of approximately R3.1 million per annum.