Browsing by Author "Kheldoun, Aissa(supervisor)"

Now showing 1 - 5 of 5

Design and Iimplementation of a PLC based automated machine vision system for industrial quality inspection
(Université M'hamed Bougara Boumerdès: Institue de génie electronic et electric, 2024) Melouk, Mohamed Redouan; Ichira, Sofiane; Kheldoun, Aissa(supervisor)
This work is aimed to develop an industrial quality control system that employs image processing techniques to enhance the quality of manufacturing processes. The thesis delves into the benefit so fimplementin ga nautomate dsyste mtha tca ndetec tproduc tfla wsand defects in real-time. In this regard, a machine vision system has been designed to inspect plastic bottles quality on a conveyor belt through various developed image-processing algorithms. The system incorporates a web-application interface that communicates with a PLC to record history of the inspected products. The results of the study indicate that the system is effective in enhancing product quality and reducing defects. The thesis is written in English and is sixty two pages long, including four chapters, forty nine figure san don etable.
Machine Learning for the Classification of Natural Events and Cyber Attacks in Power System
(Université M'hamed Bougara Boumerdès: Institue de génie electronic et electric, 2024) MENOUER, Loubna; BENNAI, Zohra; Kheldoun, Aissa(supervisor)
Migration from conventional power system to smart power system paradigm is expected to solve many problems related to reliability and environments. However, new kinds of vulnerability, such as cyber attacks can affec tit sstability .These vulnerabilities pose a significan tris ka shacker sca nmanipulat eth eoperational network by injecting false data. Such malicious activities can go undetected for a prolonged period, leading to severe consequences. The impact can range from infrastructure damage, financia llosses ,an dt opotentia lfatalities In this project, machine learning techniques are investigated to identify these faults in addition to the classical faults. Using a publicly available dataset produced in Mississippi State University’s Oak Ridge National Laboratory, simulations are run on Kaggle. Results show that the Extra-Trees algorithm produced in average superior results, with an accuracy of 95.31% for binary classificatio nan d96.90 %fo rthree-class classification ,an dRandom-Fores talgorith mwit h92.28 %accurac yfo rmulti-class classification .Thereb youtperformin git scounterpar talgorithm si nterm so faccu- racy, precision, recall, and F1-score.
Parameters Identification of bifacial pv cellsu nder different irradiations
(Université M'hamed Bougara Boumerdès: Institue de génie electronic et electric, 2024) Lagab, Sarra Oumnia; Kheldoun, Aissa(supervisor)
Given the non-linear characteristics of photovoltaic modules (PV) and their dependency on operating conditions, it becomes crucial to accurately forecast their behavior under varying temperatures and irradiance. As the demand for this technology continues to grow, it is imperative to develop effectiv emethod sfo rprecisel yextractin gth eintrinsi cparameter so fthese modules. To help in design and assess the performance of PV panels, a de-veloped model is used. The model is none other than an equivalent electrical circuit with basic components (a source, resistors, and one diode or more). Single-diode and double-diode models are the most popular in the literature. In this project, these models are used to model the bifacial photovoltaic cell. Equivalent circuit parameters must be obtained, from either a set of exper-imental data or a manufacturer’s data sheet, in order to construct a model. The aim is to obtain values that yield an accurate model. The problem is tackled as an optimization one, where the sum of Root mean square error (RMSE), between the experimental and the calculated data, and the power error around the maximum power point (MPP) is the function to be optimized. Optimization is achieved using two differen tmeta-heuristi calgorithms :Ma-rine Predators Algorithm (MPA) and Snake Optimizer Algorithm(SOA). The aforementioned algorithms are adapted to extract the bifacial PV parameters (af , I0f , Iphf , Rpf , Rsf , ar, I0r, Iphr, Rpr, Rsr) using MATLAB
PV module parameters identification using global search algorithms
(2017) Agoudjil, Mohamed Abdellah; Bouredji, Khalid; Kheldoun, Aissa(supervisor)
The nonlinear behavior of photovoltaic modules (PV) and their dependence on the working conditions imply a reliable forecasting of its behavior at different temperatures and irradiance. It is therefore essential with the growing demand of this technology to develop effective methods for accurate extraction of the module intrinsic parameters. This project presents a novel approach to extract automatically and accurately the photovoltaic module single-diode model five parameters Iph, Isat, A, Rs, and Rp using the obtained Data from the acquisition circuit. The main contribution of this project is the use of different algorithms with different PV modules in order to extract highly accurate results. Six algorithms are used to estimate the five parameters of ten modules. The local algorithm uses “fsolve” to solve resulting nonlinear equations, while the six other global algorithms use their global search feature to extract parameters and draw the I-V curve.
PV power forecasting using machine learning with hyperparameter optimization
(Université M'hamed Bougara Boumerdès: Institue de génie electronic et electric, 2024) Zouaoui, Samir; Beribeche, Abdessalem; Kheldoun, Aissa(supervisor)
The growing prominence of renewable energy, particularly photovoltaic (PV) power, neces-sitates accurate forecasting of PV power for both short and long-term horizons. Reliable forecasts are vital for effective decision-making and ensuring the stability of the electric grid. This thesis endeavors to analyze and compare a range of machine learning-based forecasting methods as alternatives to classical statistical time series forecasting techniques. Furthermore, the thesis presents a novel approach to hyperparameter tuning using meta-heuristic algorithms with Differential Evolution and Particle Swarm Optimization. The models are evaluated based on their characteristics and performance, employing multiple metrics from existing literature. Additionally, a comparative study between different hy-perparameter tuning algorithms is conducted. The investigation encompasses two distinct datasets and encompasses single-step and multi-step forecasting horizons. This thesis ap-proach involves employing models to forecast the global irradiance, which is subsequently used to predict the output power of PV systems. By decoupling the prediction process into two stages, this method offers potential advantages in terms of accuracy and reliability.