AI for Faulty Transformers: AI4SD project targets Ghana’s power reliability issues

Ghana faces ongoing challenges with electricity reliability. However, a new technology may help. Developers are working to use Artificial Intelligence (AI) for improvements.

The “AI for Energy” project is part of the AI4SD initiative. It collaborates with the Ghana Grid Company (GRIDCo). The goal is to predict transformer faults and schedule maintenance using AI.

Prof. Francis Kemausuor leads the AI4SD Energy theme. He stated that reducing downtime from transformer failures is crucial. Transformer failures significantly contribute to power outages.

“Transformers are essential for electricity distribution and industrial operations,” he noted. When they fail, it causes expensive downtime affecting many homes and businesses.

The project is in its early stages. Currently, it trains AI models using publicly available data from various sources. These models will be tested with Ghana’s grid data.

“We have requested data from GRIDCo,” Prof. Kemausuor said. “Once we receive it, we will test our models.”

The collaboration with GRIDCo provides access to years of transformer performance data. The initial plan is to pilot the system on a few transformers first. If successful, the technology will expand across the national grid.

“We need transformers with long-term operational data,” he explained. This history helps AI systems identify patterns leading to faults.

Prof. Kemausuor compared this system to a mobile phone overheating. “If your phone exceeds 50 degrees Celsius often, it may break down.” Transformers have similar parameters that can be monitored.

By tracking these parameters with sensors and AI, failures can be prevented before they occur.

The initiative also aims to learn from international examples. “Companies in France and Finland are already doing this,” he said. They plan to visit these companies and adapt their systems for Ghana.

Prof. Kemausuor thanked the French government for supporting the AI4SD initiative through its embassy in Ghana. He invited local AI experts to collaborate on this project as well.

“This work requires collective effort,” he emphasized. The team includes young engineers and computer scientists gaining valuable skills along the way. They are excited about improving electricity reliability in Ghana using AI.