No more searching for days or weeks: maintenance dates in one overview

Asset Manager Senne develops python script

Imagine a customer asking you how many times a specific component in a wind turbine has failed in the past three years. Or how much oil was consumed in a year. The answer exists, of course, but is hidden in hundreds of service reports. Each stored as a separate pdf, with technical descriptions, tables and signatures. In theory, everything is documented. In practice, finding the answer is not so easy.

‘In one year, you can receive up to 900 documents. If you want to know how many times a particular part has been replaced or how many hours were spent on repairs, you have to open them one by one. That takes hours, sometimes even days of work,” says Senne Koetsier of GreenTrust’s Asset Management department.

The data is there. Every intervention is carefully recorded: what was replaced, how much material was used, how long the technicians were on site. But without a clear overview, these remain isolated records. The real questions are more specific: how much oil is consumed annually? Which components are most vulnerable? How long, on average, does a particular type of repair take? And what is the operational environmental impact in the context of ESG reporting?

That’s where the idea arose. If everything is already in the documents, why not let them speak for themselves?

Senne developed a Python script that automatically analyzes service reports and extracts the most important information, then converts it into a clear and filterable Excel summary. AI also helped set up the code. “I asked how to structure the script and how to extract specific data. Once I defined exactly what we needed, it went relatively quickly,” he explains.

What previously took days of manual work now takes only minutes. The tool lets you filter data by component, turbine or period. This gives you quick insight into how often a specific component was replaced, how much material was used and how many working hours were recorded. Based on such an overview, patterns quickly become visible.

‘You immediately see where the weak points are. If a component fails several times a year, that’s a signal that we need to rethink our spare parts strategy,’ says Senne.

This changes the way we monitor maintenance. We can enter into discussions with the maintenance partner to stock more specific components to avoid unnecessary downtime. Negotiations for new or extended maintenance contracts are based on actual data rather than estimates. Decisions become more accurate and risks decrease. The tool is also valuable in communicating with banks and investors. ‘For ESG reports, you need a lot of different data. Now we can retrieve those almost instantly: material consumption, frequency of interventions, working hours,’ Senne explains.

The tool is now deployed at several larger wind parks and is being further developed. ‘This is just the beginning. We are still investigating which additional analyses are possible and how we can make the tool even more useful,” he adds.

At its core, it’s about faster and more accurate understanding of data. And in maintenance, every faster response means fewer lost megawatt hours.