Supercomputers help ensure safety in power plants and chemical plants

🎧 Listen to this successful collaboration in audio form!

Partner: G-Team a.s.

Field: mechanical engineering

When safety matters

In the energy sector, you cannot afford to make mistakes. Every machine and component must function reliably. And when it comes to key equipment in nuclear power plants or large chemical plants, safety is an absolute priority.

This was precisely the challenge faced by the Czech engineering G-Team company, which supplies large technological units for nuclear power plants such as Dukovany and Temelín and for ORLEN Unipetrol RPA, when it approached the National Competence Centre in HPC (NCC).

Challenge: Prevent resonance that can destroy the entire device

Imagine a bridge with a group of soldiers marching across it. If they all march at the same pace, they may excite one of the bridge’s eigenfrequencies, causing it to vibrate dangerously and potentially collapse.

The same principle applies to piping systems in the energy sector. If excitation frequencies from drives or steam flow, for example, excite one of the piping’s eigenfrequencies, serious damage could occur.

The key task for experts from NCC and IT4Innovations National Supercomputing Center was to assess whether the excitation frequency in the modified pipeline designs was the same or close to any eigenfrequency.

Solution: Calculations and simulations that save time and money

The objective of the first phase of cooperation was to verify the new configuration of the G-Team piping system and propose possible adjustments to key components.

“Using numerical models that employ the finite element method, we used a supercomputer to verify how various design modifications would behave in terms of their eigenfrequency spectrum,” explained Tomáš Karásek, head of the NCK. Supercomputer simulations confirmed that the eigenfrequency spectrum of the designed pipeline will lie outside the excitation frequency.

Thanks to supercomputers, many variants of calculations can be run at once, allowing researchers to quickly identify key parameters affecting eigenfrequencies. Verifying design modifications through experiments is time-consuming and financially inefficient.

“Our cooperation with G-Team a.s. continues, and we are focusing on state-of-the-art optimisation of the models created for specific projects,” summarised Tomáš Karásek.