Introducing AI applications in engineering education (PBL): An implementation of power generation at minimum wind velocity and turbine faults classification using AI

Abstract

AbstractThis article explores the integration of artificial intelligence (AI) applications into project‐based learning (PBL) education as a means to enhance students' education. Specifically, the implementation of AI in the context of power generation is addressed, focusing on achieving power generation at minimum wind velocity and classifying turbine faults using AI techniques. The researchers have proposed a novel generating unit which is going to generate 1 KW of electric power at a specific flow rate of air and the generated power will be stored in the battery bank through the charge controller and then the load is driven from the battery through an inverter. Iron or core losses (Hysteresis, Eddy Current losses) can be acknowledged as one of the major reasons for the inefficiency of conventional generators, therefore anovel coreless model generator was proposed which also improved efficiency and reduces drag. Wind Turbine prototype was fabricated and deployed for the testing and validation of the proposed novel design. The design produced outstanding power ratings and electrical generation characteristics compared with other existing strategies at minimal air flow. Results proved that the proposed coreless axial flux generator has the capability to produce a better power rating compared with the existing wind turbine generators. Proposed Axial flux achieved 10.73 watt power at wind velocity at around 80 rpm. At a wind velocity of 10 m/s and around 800 rpm 313–330 kwh was produced by the proposed generator while the conventional generator produced around 300 kwh. The proposed generator design performed 35% better in terms of production efficiency under load and no load conditions. Moreover, faults in turbines are very common due to the various temperatures, therefore the faults have been classified using state‐of‐the‐art AI‐based classifiers. A comparison of space vector modulation (SVM) and Naive Bayes classifiers was performed in the study to classify wind turbine faults. It was found that both classifiers performed well in achieving high accuracy. SVM achieved a slightly higher accuracy of 0.9861 compared with Naive Bayes, which achieved an accuracy of 0.967. Based on the results, it can be inferred that SVM may be a more suitable classifier for wind turbine fault classification. The case study results demonstrated the potential of AI applications in PBL education, offering students a multidisciplinary learning experience that enhances their technical knowledge, problem‐solving skills, and teamwork abilities.