Publications Internationales

Permanent URI for this collectionhttps://dspace.univ-boumerdes.dz/handle/123456789/13

Browse

Author: search.filters.author.Kouadri, AbdelmalekHas files: No

Search Results

Now showing 1 - 10 of 11
  • Thumbnail Image
    Item
    Enhancing Fault Detection in Stochastic Environments Using Interval-Valued KPCA: A Cement Rotary Kiln Case Study
    (Institute of Electrical and Electronics, 2025) Louifi, Abdelhalim; Kouadri, Abdelmalek; Harkat, Mohamed-Faouzi; Bensmail, Abderazak; Mansouri, Majdi
    Fault detection in industrial processes is challenging due to significant data uncertainty, which complicates the accurate modeling of interval-valued data and the quantification of errors necessary for reliable detection. Existing approaches, such as kernel principal component analysis (KPCA), struggle with these challenges because they rely on single-valued data representations and are unable to effectively handle interval-based variability. To address these limitations, this paper introduces the interval-valued model KPCA (IV-KPCA), which extends KPCA by redefining similarity measures and kernel functions to accommodate interval-valued uncertainty. IV-KPCA preserves the interval structure throughout the modeling process, enhancing robustness to dynamic uncertainties and improving fault detection in complex nonlinear systems. Within this framework, fault detection statistics (T 2 , Q, and 8) are developed to enable precise error quantification. The proposed method is validated on a cement rotary kiln process, a highly stochastic industrial system characterized by significant uncertainties. Experimental results demonstrate that IV-KPCA reduces false alarms, missed detections, and detection delays by over 100%, 90%, and 95%, respectively, compared to traditional methods. These findings underscore the potential of IV-KPCA in enhancing fault detection performance in complex, uncertain environments
  • No Thumbnail Available
    Item
    Adaptive Polynomial Kolmogorov–Arnold Networks with Multi-Modal Sensor Data Fusion for High-Precision Fault Diagnosis in Wind Energy Systems
    (Institute of Electrical and Electronics Engineers, 2025) Attouri, Khadija; Mansouri, Majdi; Kouadri, Abdelmalek
    This paper presents a robust multi-modal sensor-based fault diagnosis system for a variable-speed Wind Energy Conversion System (WECS) equipped with a Squirrel Cage Induction Generator (SCIG). The system’s intricate dynamics and the non-linear nature of faults pose significant challenges to accurate and timely detection. To address these challenges, we introduce a novel classifier based on the Adaptive Polynomial Kolmogorov–Arnold Network (AdaptpolyKAN), which employs trainable polynomial basis functions to capture complex signal patterns collected from heterogeneous sensors. A comprehensive dataset was generated, covering three fault types open-circuit, short-circuit, and wear-out artificially injected into the system’s dual-converter topology. The proposed AdaptpolyKAN model is evaluated using two sensor data fusion strategies: Early Fusion, which concatenates features from three distinct sensor modalities (Generator, DC Bus, and Grid), and Late Fusion, which aggregates predictions from separate classifiers. The results demonstrate that the AdaptpolyKAN model, particularly with the Early Fusion strategy, achieves a near-perfect accuracy of 99.97%, outperforming all other benchmark classifiers. In particular, it clearly outperforms the baseline PolyKAN. With Early Fusion, AdaptpolyKAN achieved a near-perfect accuracy of 99.97%, compared to PolyKAN’s 95.03%. Even under Late Fusion, AdaptpolyKAN (96.24%) maintained a substantial margin over PolyKAN (91.99%). This performance gap highlights the superiority of AdaptpolyKAN’s adaptive polynomial basis functions in capturing nonlinear, multimodal interactions, whereas the fixed structure of PolyKAN remains limited. Furthermore, the study confirms that Early Fusion consistently delivers superior performance compared to Late Fusion for this application. These findings underscore the effectiveness of AdaptpolyKAN’s adaptive architecture and the advantages of a holistic multi-sensor fusion approach for high-precision fault diagnosis in WECS
  • Thumbnail Image
    Item
    Deep Learning-Based Fish Health Monitoring and Diagnosis: A Review
    (IEEE, 2025) Kheriji, Lazhar; Kouadri, Abdelmalek; Mansouri,Majdi
    Fish in aquaculture systems face health challenges influenced by aging, water quality, and environmental conditions. These issues affect critical components like feeding and filtration, potentially reducing efficiency and causing system failure. Effective Health Monitoring and Diagnosis (HMD) relies on high-quality features such as behavior, physical condition, feeding habits, and water parameters. However, traditional hand-crafted approaches often fail to capture the complex and nonlinear interactions between biological and environmental factors, limiting their adaptability to sudden changes in water conditions or disease outbreaks. This gap motivates the use of intelligent, multimodal learning strategies that integrate diverse data sources for more robust and reliable analysis. Advances in computing power, large datasets, and sophisticated algorithms have made deep learning (DL) a transformative tool in this field. By combining DL with multimodal data integration, it becomes possible to learn high-level representations directly from heterogeneous inputs such as water quality measures, behavioral signals, and visual observations, thereby overcoming the limitations of conventional feature-based methods. This paper reviews DL-based multimodal approaches in aquaculture HMD, comparing recent techniques, their strengths, and limitations. We also discuss future directions, emphasizing multimodal data fusion to enhance DL-driven health monitoring. This review provides a concise resource for researchers and practitioners aiming to advance aquaculture health monitoring.
  • No Thumbnail Available
    Item
    Deep Learning for Sustainable Aquaculture: Opportunities and Challenges
    (Institute of Electrical and Electronics Engineers Inc., 2025) Kheriji, Lazhar; Kouadri, Abdelmalek; Mansouri, Majdi
    With the rising global demand for aquatic products, aquaculture has become a cornerstone of food security and sustainability. This review comprehensively analyzes the application of deep learning in sustainable aquaculture, covering key areas such as fish detection and counting, growth prediction and health monitoring, intelligent feeding systems, water quality forecasting, and behavioral and stress analysis. The study discusses the suitability of deep learning architectures, including CNNs, RNNs, GANs, Transformers, and MobileNet, under complex aquatic environments characterized by poor image quality and severe occlusion. It highlights ongoing challenges related to data scarcity, real-time performance, model generalization, and cross-domain adaptability. Looking forward, the paper outlines future research directions including multimodal data fusion, edge computing, lightweight model design, synthetic data generation, and digital twin-based virtual farming platforms. Deep learning is poised to drive aquaculture toward greater intelligence, efficiency, and sustainability
  • No Thumbnail Available
    Item
    Sensor Fault Detection in Uncertain Large-Scale Systems Using Interval-Valued PCA Technique
    (IEEE, 2025) Louifi, Abdelhalim; Kouadri, Abdelmalek; Harkat, Mohamed-Faouzi
    Principal component analysis (PCA)-based fault detection and diagnosis (FDD) is a well-established, data- driven method that has shown remarkable performance. Despite the excellent reputation of the PCA, it is not an opti- mal solution, mainly due to the effect of system parameters’ uncertainties and imprecise measurements. These drasti- cally affect the decision-making concerning the operating state of the process. In this article, the data collected by different sensors are transformed from a single value to an interval value form by which errors and uncertainties in the measurements are quantified satisfactorily. Then, the process modeling based on the PCA technique has been duly performed for interval-valued. Afterward, the well-known fault detection statistics T 2 , Q, and 8 are obtained under an interval-valued representation. The developed technique is tested in the cement rotary kiln process. Its performance in terms of false and missed alarms and detection delay is compared with that of other techniques through an actual involuntary system fault and other different types of sensor faults. The obtained results show high superiority in detecting accurately and quickly distinct faults in a stochastic environment, including unknown and uncontrolled uncertainties. Consequently, the results have been reduced by more than 33%, 85%, and 45% for T 2 , Q, and 8, respectively, compared with the best results of the studied methods.
  • No Thumbnail Available
    Item
    Medium-term wind power forecasting using reduced principal component analysis based random forest model
    (SAGE Publications Inc, 2024) Jamii, Jannet; Trabelsi, Mohamed; Mansouri, Majdi; Kouadri, Abdelmalek; Mimouni, Mohamed Faouzi; Nounou, Mohamed
    Due to its dependence on weather conditions, wind power (WP) forecasting has become a challenge for grid operators. Indeed, the dispatcher needs to predict the WP generation to apply the appropriate energy management strategies. To achieve an accurate WP forecasting, it is important to choose the appropriate input data (weather data). To this end, a medium-term wind power forecasting using reduced principal component analysis (RKPCA) based Random Forest Model is proposed in this paper. Two-stage WP forecasting model is developed. In the first stage, a Kernel Principal Component Analysis (KPCA) and reduced KPCA (RKPCA)-based data pre-processing techniques are applied to select and extract the important input data features (wind speed, wind direction, temperature, pressure, and relative humidity). The main idea behind the RKPCA technique is to use Euclidean distance for reducing the number of observations in the training data set to overcome the problem of computation time and storage costs of the conventional KPCA in the feature extraction phase. In the second stage, a Random Forest (RF) algorithm is proposed to predict the WP for medium-term. To evaluate the performance of the proposed RKPCA-RF technique it has been applied to data extracted from NOAA’S Surface Radiation (SURFRAD) network at Bondville station, located in USA. The presented results show that the proposed RKPCA-RF technique achieved more accurate results than the state-of-the-art methodologies in terms of RMSE (0.09), MAE (0.23), and R2 (0.85). In addition, the proposed technique achieved the lowest overall computation time (CPU).
  • No Thumbnail Available
    Item
    RKPCA-based approach for fault detection in large scale systems using variogram method
    (Elsevier, 2022) Kaib, Mohammed Tahar Habib; Kouadri, Abdelmalek; Harkat, Mohamed Faouzi; Bensmail, Abderazak
    Principal Component Analysis (PCA)-based approach for fault detection is a simple and accurate data-driven technique for feature extraction and selection. However, PCA performs poorly if the data used has nonlinear characteristics where this type of data is widely present in most industrial processes. To overcome this drawback, Kernel PCA (KPCA) is an alternative technique used to work on this type of data but it requires more computation time and memory storage space for large-sized data sets. Many size reduction techniques have been developed to select the most relevant observations that will be employed by KPCA. This, known as Reduced KPCA (RKPCA), consequently requires less computation time and memory storage space than KPCA. Besides, it possesses the advantages of both KPCA and standard PCA. In this paper, a reduction in the size of a data set based on a multivariate variogram is proposed. According to its conventional formalism, the uncorrelated observations are selected and kept to form a reduced training data set. Afterward, the KPCA model is built through this data set for faults detection purposes. The proposed RKPCA scheme is tested using an actual involuntary process fault and various simulated sensor faults in a cement plant. Compared to other RKPCA techniques, the developed one yields better results
  • No Thumbnail Available
    Item
    Multivariate nuisance alarm management in chemical processes
    (Elsevier, 2021) Kaced, Radhia; Kouadri, Abdelmalek; Baiche, Karim; Bensmail, Abderazak
    Alarm systems are of vital importance in the safe and effective functioning of industrial plants, yet they frequently suffer from too many nuisance alarms (alarm overloading). It is necessary to intelligently enhance existing alarm systems and supply accurate information for the operators. Nowadays, process variables are more correlated and complicated. This correlation structure can be used as a basis to manage alarms efficiently. Hence, multivariate approaches are more appropriate. Designing a system aimed at reducing nuisance alarms is an essential phase to guarantee the reliable operation of a plant. Due to the definition of alarm limits, the problem of false alarms is inevitable in multivariate methods. In this paper, the conventional Principal Component Analysis (PCA) is applied to extract the sum of squared prediction error (SPE) known as the statistic and the Hotelling statistic. These statistics are used separately as alarm indicators where their control limits are duly modified. Consequently, for each statistic, a nonlinear combination of alarm duration and alarm deviation, is additionally exploited as a new requirement to activate an alarm or not. The resulting new index is fed to a delay timer with a defined parameter . The implementation of this technique resulted in a significant reduction in the severity of alarm overloading. Historical data collected from the cement rotary kiln operating under healthy conditions are employed to adequately build the PCA model and extract the proposed alarming indexes. Then, various testing data sets, covering different types of faults occurring in the cement process, are used to assess the performance of the developed method. In comparison with the conventional PCA technique, alarms are better managed nd almost nuisance alarms are suppressed. The proposed method is more robust to false alarms and more sensitive to fault detection
  • No Thumbnail Available
    Item
    Kernelized relative entropy for direct fault detection in industrial rotary kilns
    (John Wiley and Sons Ltd, 2018) Hamadouche, Anis; Kouadri, Abdelmalek; Bensmail, Abderazak
    The objective of this work is to use a 1-dimensional signal that reflects the dissimilarity between multidimensional probability densities for detection. With the modified Kullback-Leibler divergence, faults can be directly detected without any normality assumption or joint monitoring of related test statistics in different subspaces such as the T2 and SPE in principal component analysis–based methods. To relieve the difficulty associated with asymptotic high-dimensional density estimates, we have estimated the density ratio rather than the densities themselves. This can be done by approximating the density ratio with kernel basis functions and learn the weights from the available data. The developed algorithm is generic and can be applied to any industrial system as long as process historical data is available. As a case study, we apply this algorithm to a real rotary kiln in operation, which is an integral part of the cement manufacturing plant of Ain El Kebira, Algeria.
  • No Thumbnail Available
    Item
    Designing alarm system using modified generalized delay-timer
    (Elsevier, 2019) Kaced, Radhia; Kouadri, Abdelmalek; Baiche, Karim
    Alarm systems are of crucial importance in ensuring safety and efficiency of industrial installations. In practice, alarm systems are not properly designed or given the attention they deserve; their performance is unsatisfactory. The main role of alarms is to inform the operator of any incident in the plant. Regrettably, most occurred alarms are false and nuisance. To avoid this, industrial community developed techniques like deadbands, filters and delay-timers. Delay-timer is largely exercised in industry. This article presents a new concept for designing a Generalized delay-timer. out of n consecutive samples is not the only condition to activate an alarm, we will use additional setpoints to rise or clear an alarm. Three performance indicators namely, False Alarm Rate (), Missed Alarm Rate () and Average Alarm Delay () are computed for the proposed method. At the end, the modified Generalized delay-timer method is examined and compared with the Generalized delay-timer using a simulation and industrial case studies The obtained results show that alarm system performance is improved and even optimized using the modified Generalized delay-timer