Planning hydrogen network infrastructure
Hydrogen is one of the key components of the envisioned future fossil-free economy. It can be produced from surplus renewable energy, is easier to store over long time periods than electricity, and may replace fossil fuels and chemical products, either by direct usage or as part of other substituting compounds. Currently, there is an ongoing design process for the European network infrastructure for transporting hydrogen on shore. The final network will consist of a mix of repurposed existing natural gas pipelines and newly built connections.
An important aspect of the evaluation of a given hydrogen network design is the expected wear of the planned pipelines. To a large extend, this is a result of pressure level changes within the pipe, which cause hydrogen embrittlement and thereby limit the possibilities network control. To determine the expected number of pressure load cycles over a given period of time, the traditional usage of steady-state balanced planning scenarios is no longer a feasible option. Furthermore, these extreme scenario only capture the capacity requirements for a given network, and do not take into account the need to react with flexibility on volatile supply and demand requests. To properly incorporate all of these effects, a dynamic simulation of the network state evolving over time is required.
Photo: Jarle Vines (Creative Commons Attribution Sharealike 3.0)
Control optimisation in the context of network planning
The dynamic control of complex gas transport networks is still a manual process involving expert dispatchers who decide on control changes of individual network elements based on vast experience and knowledge of the network. However, it is unreasonable to expect human experts to determine the best network control for every potential network design and supply and demand scenarios. Instead, an automated tool is required that can determine control recommendations for each network element, while taking into account various objectives. These objectives include the above-mentioned reduction of pressure load cycles, meeting all future supply and demand requirements as best as possible, and reducing the overall number of control recommendations.
The CORGI tool
As there is currently no tool that can dynamically control hydrogen (or natural gas) networks of realistic size, the German transport system operator Open Grid Europe (OGE) asked us to join a competition to create a prototype that would solve the given task for hydrogen networks on a set of test cases. As a result, we created CORGI, a tool for Control Optimisation for Resilient Gas Infrastructure. While other companies declared the problem unsolvable using current state-of-the-art methods, CORGI found a simulator-validated solution for every test case. In fact, the majority of cases could be solved by CORGI alone. In addition, CORGI was often faster than human experts in solving the test cases. In a subsequent evaluation, we found that CORGI was even able to solve all test cases of a newly defined set of problems based on actual hydrogen planning networks.
Going productive
With the development of CORGI, we have proven that it is possible to create an automated system for determining control decisions for hydrogen planning networks. Thereby, we achieved what even experts in the field claimed to be impossible. The next step is to turn the current prototype version of CORGI into a robust application suitable for the day-to-day evaluation of dynamic hydrogen planning scenarios. In the long term, the technology even has potential to be used as a real-time support system in network dispatching.