Publications Scientifiques

Permanent URI for this communityhttps://dspace.univ-boumerdes.dz/handle/123456789/10

Browse

Subject: search.filters.subject.Neural Networks

Search Results

Now showing 1 - 7 of 7
  • No Thumbnail Available
    Item
    Rigorous Explainable Artificial Intelligence Models for Predicting CO2-Brine Interfacial Tension: Implications for CO2 Sequestration in Saline Aquifers
    (American Chemical Society, 2025) Nait Amar, Menad; Youcefi, Mohamed Riad; Alqahtani, Fahd Mohamad; Djema, Hakim; Ghasemi, Mohammad
    Carbon capture and sequestration (CCS) is an attractive approach for reducing carbon dioxide (CO2) emissions, with saline aquifers offering promising sites for long-term sequestration. Interfacial tension (IFT) between CO2 and brine plays a crucial role in the trapping efficiency. This study develops explainable artificial intelligence (XAI) models to accurately predict the IFT in CO2–brine systems. Three advanced machine learning models, namely, Super Learner (SL), Elman Neural Network (ENN), and Power Law Ensemble Model, were implemented based on a data set comprising 2616 measurements. Among the established paradigms, SL achieved the highest accuracy (RMSE = 0.7813 and R2 = 0.9953) across diverse conditions. To enhance model transparency, Local Interpretable Model-agnostic Explanations and SHAP (SHapley Additive Explanations) interpretability techniques were employed, confirming strong alignment with experimental trends. Comparative analysis further demonstrated that the SL scheme surpasses existing literature models. Overall, this study highlights the effectiveness of XAI-based predictive modeling for accurately estimating the CO2–brine IFT under diverse operational conditions. Future implementation in real CCS projects can offer valuable insights into injection strategies, trapping mechanisms, and long-term formation stability
  • No Thumbnail Available
    Item
    Rigorous Explainable Artificial Intelligence Models for Predicting CO2-Brine Interfacial Tension: Implications for CO2 Sequestration in Saline Aquifers
    (American Chemical Society, 2025) Nait Amar, Menad; Youcefi, Mohamed Riad; Alqahtani, Fahd Mohamad; Djema, Hakim; Ghasemi, Mohammad
    Carbon capture and sequestration (CCS) is an attractive approach for reducing carbon dioxide (CO2) emissions, with saline aquifers offering promising sites for long-term sequestration. Interfacial tension (IFT) between CO2 and brine plays a crucial role in the trapping efficiency. This study develops explainable artificial intelligence (XAI) models to accurately predict the IFT in CO2–brine systems. Three advanced machine learning models, namely, Super Learner (SL), Elman Neural Network (ENN), and Power Law Ensemble Model, were implemented based on a data set comprising 2616 measurements. Among the established paradigms, SL achieved the highest accuracy (RMSE = 0.7813 and R2 = 0.9953) across diverse conditions. To enhance model transparency, Local Interpretable Model-agnostic Explanations and SHAP (SHapley Additive Explanations) interpretability techniques were employed, confirming strong alignment with experimental trends. Comparative analysis further demonstrated that the SL scheme surpasses existing literature models. Overall, this study highlights the effectiveness of XAI-based predictive modeling for accurately estimating the CO2–brine IFT under diverse operational conditions. Future implementation in real CCS projects can offer valuable insights into injection strategies, trapping mechanisms, and long-term formation stability
  • No Thumbnail Available
    Item
  • No Thumbnail Available
    Item
    The proposed hybrid intelligent system for path planning of Intelligent Autonomous Systems
    (2009) Hachour, Ouarda
    In this paper, , we discuss the ability to deal with a Hybrid Intelligent Systems (HIS) for Intelligent Autonomous Vehicles IAV in unknown environment. The aim of this work is to develop HIS combining Genetic Algorithms (GA), Fuzzy Logic (FL), Neural Networks (NN) and Expert Systems (ES). This project deals with a simulation program that allows a robot to identify a path to reach a specified target avoiding obstacles. The combination of (ES FL, NN, GA) offers design flexibility and robust integration and has the benefits of reduced communications overhead and improved runtime performance. This integration provides the robot the possibility to move from the initial position to the final position (target) without collisions. The robot moves within the unknown environment by sensing and avoiding the obstacles coming across its way towards the target. The algorithm permits the robot to move from the initial position to the desired position following an estimated trajectory. The proposed hybrid navigation strategy is designed in unknown environment with static unknown obstacles. This approach must make the robot able to achieve these tasks: to avoid obstacles, and to make ones way toward its target by ES_FL_GA_NN system capturing the behavior of a human expert. The integration of these technologies (FL, NN, ES, and GA) has proven to be a way to develop useful real-world applications, and hybrid systems involving robust adaptive control. The proposed approach has the advantage of being generic and can be changed at the user demand. The results are satisfactory to see the great number of environments treated. The results are satisfactory and promising. the proposed method is computationally efficient and is suitable for more integration of hybrid intelligent systems
  • No Thumbnail Available
    Item
    Neural path planning for mobile robots
    (2011) Hachour, Ouarda
    Navigation is a major challenge for autonomous, mobile robots. The problem can basically be divided into positioning and path planning. The proposed path finding strategy is designed in a known static environments. The proposed method starts from an initial point to a target point establishing a control nodes neural networks for which connections are made to determine the form of the path. This algorithm provides the robot the possibility to move from the initial position to the final position (target). The robot moves within the unknown environment by sensing and avoiding the obstacles coming across its way towards the target. The proposed algorithm can deal with any shape obstacles even if it is the case of circular obstacles. This case is the hardest one in any navigation problem. The problem is solved by proposing neural networks navigation systems. Indeed, NNs are well adapted in appropriate form when knowledge based systems are involved. Since the network is able to take into account and respond to new constraints and data related to the external environments, the adaptation here is largely related to the learning capacity. Besides, Networks of neurons can achieve complex classification based on the elementary capability of each neuron to distinguish classes its activation function. Some useful solutions are proposed for each situation. For any proposed environment, the robot succeeds to reach its target without collisions. The results are satisfactory to see the great number of environments treated The simulation results display the ability of the neural networks based approach providing autonomous mobile robots with capability to intelligently navigate in several environments
  • No Thumbnail Available
    Item
    Cognitive tasks behavior of intelligent autonomous mobile robots
    (2011) Hachour, Ouarda
    In this paper we propose a neural network based navigation for intelligent autonomous mobile robots. The proposed neural networks algorithm deals with unknown static obstacles. Neural Networks deal with cognitive tasks such as learning, adaptation generalization and they are well appropriate when knowledge based systems are involved. To solve navigation problems, neural networks prove interesting to deal with the behaviour of autonomous mobile robots near the human being in reasoning. This paper deals with an algorithm for two dimensional (2D) path planning to a target for mobile robot in unknown environment. A complete path planning algorithm should guarantee that the robot can reach the target if possible, or prove that the target can not be reached. Just as human being, a neural network relies on previously solved examples to build a system of “neurons” that makes new decisions, classification and forecasts. Networks of neurons can achieve complex classification based on the elementary capability of each neuron to distinguish classes its activation function. In designing a Neural Networks navigation approach, the ability of learning must provide robots with capacities to successfully navigate in the environments like our proposed maze environment. The simulation results display the ability of the neural networks based approach providing autonomous mobile robots with capability to intelligently navigate in several environments
  • No Thumbnail Available
    Item
    Artificial neuron network usage for asynchronous motor malfunction diagnostics in real-time operation mode
    (2003) Chetate, Boukhmis; Khodja, Djalal Eddine
    In the article it is told about the device of automatic diagnostics of electromechanical systems, which consists of two subsystems: a subsystem of acting data transformation and a subsystem of data processing. The first carries out data reception and their processing (distribution of data, estimation of parameters and their representation) while the second finds out failures (under the Artificial Neural Network help) which can occur in an electromechanical system and gives the recommendations for their elimination. However, the investigation of three Neural Networks have been proceeded to choose the most effective diagnostic failure Neural Network. In addition, to give the improve diagnostic, it is important to do the correct choice of parameters. According to made analysis stator current, rotation speed and acting signals are the most important parameters to be considered describing failures influence (their changes are essentially more in the defect occurrence case) and their physical values can be measured easily with the sensor