The growing need for more accurate renewables power forecasting

Ioannis Papadopoulos, Business Manager, Forecaster, DNV

Renewable energy sources are the main building block for a decarbonized energy system. At the same time, wind and solar will be the primary sources of renewable energy being deployed at scale for the foreseeable future and these are variable technologies meaning that the output of a renewable power plant is as changeable as the weather. Therefore, an energy system and more specifically an electricity system based predominantly or exclusively on renewables will need to be able to balance out that variability to ensure grid stability and continuous supply of energy.

The key solution being considered at this time is energy storage, either at project or grid level. The storage solutions themselves also take many forms with conventional hydro plants, chemical batteries, flywheel, gravity, or compressed fluid storage all being evaluated, while generating and storing hydrogen at a grid level is also being explored as a possible way forward.

However, most grids globally still have significant amounts of non-renewable generation online, based predominantly on fossil fuel generation. This in turn means that any fluctuation in the availability/price of the fossil fuel resources can also have significant impacts on the performance of the system in which they are embedded.

Whichever solution or combination of solutions proves to be more appropriate to combine with renewable generation, there will still be the need to ensure that the supply of electricity from renewable energy projects is accurately predicted, to ensure the smooth running of the system. This is where short-term power forecasting becomes a key piece of the puzzle. This will become increasingly true as wind and solar generation starts to play an ever bigger role in the energy system and as the continuous need to drive down the costs of generation mean that higher efficiencies will be sought in every aspect of the development and operating profile of a renewable energy project.

Although forecasting has been happening for a long time in some markets, it is also true that not all forecasts are created equal. While sufficient computing power is a prerequisite, the main defining characteristics of an accurate service are the weather input datasets, namely numerical weather prediction models (NWPs), and the fine tuning of well-designed models. The weather inputs are available to the wider market on an ongoing basis, but without sufficient historical datasets built up over years of work in the field, significant long-term insights can be missed.

Similarly, the rise of machine learning means that many providers from non-renewable-related backgrounds are now trying to capitalize on their data analytics skills gained in other fields. However, the key to good machine learning results is a strong understanding of the data, so again without the datasets and experience to train these machine learning models, success could be limited.

Finally, forecasting is just one aspect of securing renewable energy production and goes hand in hand with accurate operational monitoring, planning, and a whole lifecycle approach to maintenance. The more integrated these aspects are, the more secure the supply. However, the ability and knowledge to combine all these elements is something that few operators (and even fewer advisors) can do successfully.

Making sure your assets are online and that the resources they need are available are the only way to secure financial and system prosperity – and I am not just referring to conventional generation!

Author biography:

Ioannis is a Principal Engineer in DNV with 14 years of professional experience in the field of renewable energy. Ioannis received an aerospace engineering degree and a PhD in the energy analysis of power systems, before joining the company in 2009. In DNV he was involved in wind energy assessments in Greece, the Middle East and Africa, before moving to the Strategy and Policy division of the company. There, he worked for large private investors and governments on innovative business projects, including providing strategy advice to offshore marine manufacturers, policy advice to international development banks and undertaking financial assessments for developers. Following this period, Ioannis was a Director of the financial consultancy division of PEP, which provided advice to the renewables investors, as well as Project Director of the developer and advisory division, undertaking Owner's Engineering activities in wind power, globally. In his time in those roles he was involved in over 2 GW of wind projects in Europe and the US for large equity investors and pension funds. As part of this work, Ioannis was also responsible for assessing the potential markets for innovative business plans and products, both for entrepreneurs and potential investors. In late 2018, Ioannis re-joined DNV as the Business Lead for the Renewable Energy Analytics department. In this role, Ioannis is responsible for developing the department's preconstruction energy modelling activities to meet customer needs as well as driving innovations in the services and products under his remit.