International Research Journal of Education and Technology

 In recent years, researchers have increasingly focused on fault diagnosis in electrical machines. Both users and manufacturers are emphasizing the integration of diagnostic capabilities within software to enhance reliability and scalability. The rapid growth of the Internet of Things (IoT) has significantly contributed to technological advancements across industries, medical fields, and environmental applications. By leveraging cloud computing, IoT offers efficient processing power and presents new opportunities for industrial automation through the implementation of IoT and industrial wireless sensor networks (IWSN). Regular monitoring facilitates early detection of machine faults, which is advantageous for industrial automation by ensuring more efficient process control. The performance of fault detection and classification using machine learning algorithms depends heavily on the number of features considered. However, an increase in feature dimensionality can negatively impact classification accuracy. To address this challenge, the proposed work introduces a feature extraction method based on the oriented support vector machine (FO-SVM). This algorithm focuses on identifying the most relevant feature set, resulting in more accurate fault classification. Extracting relevant features prior to classification enhances accuracy and reduces the computational complexity. Additionally, the lower dimensionality of the proposed approach reduces processing time, making it more suitable for cloud-based platforms. Experimental results from implementing the proposed method demonstrate an effective solution for monitoring machine conditions and accurately predicting faults. This is achieved through the integration of cloud computing, industrial wireless sensor networks, and IoT services.

Key Words: industrial wireless sensor networks (IWSNs); internet of things (IoT); support vector machine; fault diagnosis