In off-grid systems oftentimes thermal generators are operated in conjunction with renewable energy sources and energy storage. In these cases it becomes a non-trivial task to determine the optimal economic dispatch of the integrated system.

We will show that the assumption on the future economic dispatch during operation has a strong impact on the optimal system design and sizing of components and should already be carefully anticipated while planning the system.

Furthermore, we will present a novel technique using machine learning and artificial intelligence (AI) that can help to efficiently design off-grid systems that comprise renewable energy sources, energy storage and thermal generators while anticipating challenging operation regimes.

In a case study our algorithm is compared to conventional optimization methods that are used in the course of system design. We quantify the impact on the component sizing and the reduction of costs and emissions that can be achieved by using the AI-based algorithm. Finally, we will give an overview on how AI can generally leverage the global expansion of off-grid energy systems. Key Learning Points:

• Assumption on the economic dispatch has a strong impact on optimal system design in complex off-grid energy systems that integrate other energy technologies.
• In this case AI is beneficial for the design of the energy system.
• AI helps to precisely model the fluctuation and the stochastic nature of renewable energy sources as well as the dynamics of conventional generators and energy storage.
• The high resolution insight into short term dynamics and reserve requirements allows for a better system design and operation strategy as well as increased system stability.
• Costs and emissions can be significantly reduced by using AI-based algorithms as compared to conventional methods.

Speaker: