Instrumentation dans l'industrie pétrochimique

Permanent URI for this collectionhttps://dspace.univ-boumerdes.dz/handle/123456789/10962

Browse

Filters

20242

Settings

Author: search.filters.author.Habbi, Hacene (Promoteur)Subject: search.filters.subject.Pétrochimie : Instrumentations

Search Results

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Evolving cloud-based fault monitoring system for an industrial gas turbine
    (Université M’Hamed Bougara Boumerdès : Faculté des Hydrocarbures et de la Chimie, 2024) Fadel, Mohamed Saout El Hak; Messaoudi, Mohammed Rida; Habbi, Hacene (Promoteur)
    The Thesis focuses on the development and implementation of an evolving fault monitoring system for an industrial gas turbine. Building on our study of turbine during our internship, this work is dedicated to creating an evolving cloud-based system of AnYa-type to detect and identify different fault scenarios in the industrial gas turbine. This model leverages real-time data and non-parametric methods to adapt to dynamic environments, enhancing efficiency and reliability in complex systems. Our efforts have been directed towards building this monitoring system to ensure continuous learning and adaptation, ultimately improving operational performance under normal and fault modes.
  • No Thumbnail Available
    Item
    Intelligent calibration of level sensors using functional link neural networks with piecewise linear interpolation
    (Université M’Hamed Bougara Boumerdès : Faculté des Hydrocarbures et de la Chimie, 2024) Benhacine, Souheil Nadjmeddine; Naili, Mohamed Chakib; Habbi, Hacene (Promoteur)
    This Study focuses on designing an intelligent calibration model for level sensors by using the Functional Link Artificial Neural Network (FLANN). The FLANN has simple architecture and requires less computational effort compared to other neural networks models. This made it good enough in extending the linearity of many sensors and transducers as reported in recent literature. Despite its advantages, the standard FLANN model has limitations in terms of generalization and accuracy. To overcome this shortcoming, we propose in this study an approach that relies on integrating Piecewise Linear (PWL) interpolation with the FLANN model. This approach aims to improve the overall performance of the calibration process, offering better generalization capabilities. Building on an extensive experimental investigation of the intelligent calibration model on the typical problem of level sensors calibration, this manuscript outlines the methodology, implementation steps, and findings of our study.