Consumers also keep track of how much energy they consume, and if someone has made an error when they declare a solar power plant’s yield, the supplier or aggregator or energy trading company can be fined. To prevent this, Inion Software has developed a tool that uses artificial intelligence to predict how a particular solar power plant will perform a few days ahead and how much electricity will be consumed.
The tool evaluates a specific solar power plant based on the capacity of its solar modules, its location, the angle of its construction and more, while also analysing the weather, the model of the solar power plant and technical parameters, and it uses this data to predict how much energy will be produced. Weather parameters include temperature, precipitation, cloud cover, and sun position. The AI consults several different sources of meteorological data to determine the most accurate forecast.
See also: Inion Software and Volter enter into a strategic partnership for Europe
The company's artificial intelligence tool is able to provide accurate forecasts for the next day. It can forecast production over more extended periods. However, the longer the period, the less accurate the forecast. The primary influence on the forecast is cloud cover, which is difficult for meteorologists to predict over a longer period. Clouds change the air temperature and brightness at a particular solar power plant location, and this means that its production output will change.
From the point of the concept of the tool, it took a year to get it launched. Still, the tool needs to be more comprehensive – it needs to be constantly updated, trained, refined and continuously monitored to see how it can be improved.
Anticipate future failures
In the future, Inion Software says the company will be bringing even more innovations using artificial intelligence to market. A project with the Norwegians is currently underway; once completed, this tool will be able to predict future solar power plant failures. Another benefit of it is that it will not need to check each panel individually when a fault occurs – this is particularly important with large solar farms. Even before a fault occurs, the new tool will acknowledge the discrepancy in the productivity schedule compared to neighbouring solar power plants and existing historical production figures.
Also interesting: Otovo receives financing of around 40 million euros
Inion Software is planning to develop a feature to assess when cleaning solar panels is financially beneficial, based on production volume and cost-effectiveness compared to energy loss. (mfo)
