<![CDATA[Artificial intelligence (AI) can extend practice-oriented vocational learning when digital guidance remains anchored to physical evidence. This study developed and evaluated an AI-assisted off-grid photovoltaic troubleshooting trainer for Grade XI students in the Solar, Hydro, and Wind Energy Engineering specialization at an Indonesian vocational school. The ADDIE framework guided analysis, design, development, implementation, and evaluation. The system integrated a physical off-grid photovoltaic trainer, current- and voltage-sensing through an Internet-of-Things controller, deliberately introduced fault conditions, and a web-based generative-AI interface deployed through Ollama. Evidence was collected through observation, structured interviews, documentation, expert review sheets, and a four-point student response questionnaire. Two subject-matter experts, one instructional-media expert, and 34 students participated in the evaluation. Post-revision expert scores were 95.8%, 97.9%, and 97.9%, while the overall student response reached 93.0%. The trainer was therefore highly feasible and well accepted for practicum activities involving circuit assembly, measurement, fault identification, corrective-action reasoning, and AI-supported troubleshooting. Because no pretest-posttest comparison, control group, or independent benchmark of AI diagnostic accuracy was used, the findings establish product feasibility rather than learning effectiveness. The developed trainer offers a practical model for integrating generative AI with hands-on vocational practice by supporting evidence-based troubleshooting and preparing students for technology-enhanced renewable-energy learning environments.]]>
Copyrights © 2026
