Improving efficiency and maintenance schedules while extending the plant’s useful life expectancy

About the customer

Assystem is an international engineering and digital services group focused on low-carbon projects that accelerate the transition to clean energy. Assystem is committed to the development of decarbonized electricity (fusion energy, renewables and electricity grids) and clean hydrogen. The group is also helping drive the use of decarbonized electricity in industrial sectors such as transportation. Assystem supports the fusion energy sector in the UK, including this project to develop a physics-based Digital Twin demonstrator for future fusion energy projects.

These programs include support to the development of the Spherical Tokamak for Energy Production (STEP), the U.K.’s prototype fusion power plant that will demonstrate fusion energy’s commercial viability.

Their challenge

Fusion power plants require complex digital simulation models during the design assessment phase. Their inspection and maintenance intervals and total life are defined based on the expected loading on the as-designed model, factoring in significant risks. Meanwhile, the real-world loads the plant is subjected to may differ. This gap provides a scope for programs directed at either improving the plant’s lifetime value or quantifying the effects of higher-than-expected usage.

The Assystem team wanted to leverage the expensive design models to create a Digital Twin by inputting the sensor data that was livestreamed from the plant, which helps engineers understand the plant’s structural integrity and further optimize inspection and maintenance schedules. The simulation models can be the original full fidelity models or they can be converted via reduced order models (ROMs) and functional mock-up units (FMUs). The finished system had to accommodate virtual sensors for variables that cannot be physically measured on real-world systems and had to interface seamlessly with the Client’s existing communication systems.

Our solution

The hosted (cloud agnostic) version of Rapidminer® IoT Studio software, which is part of the Siemens Xcelerator business platform of software, hardware and services, formed the solution’s backbone and analytic frontend. It allows automation using serverless backend functions, data storage, edge connectivity to the operational systems and advanced dashboarding capabilities that give users immediate insight. The pilot phase used synthetic data to simulate the real-time data stream from operating plants into the internet of things (IoT) platform. Defined Python scripts trigger related physics models, which then process the data and return the results back to the IoT platform. A configured frontend provides visual analytics highlighting event repercussions for better decision-making. Finite element analysis (FEA) models from multiple vendors, a commercial-grade fatigue analysis solver and real-time weather data from OpenWeather were all connected to the IoT platform, and the simulation process data management (SPDM) system at the Client’s was linked to store the data models. Email for calendar requests for maintenance and short message service (SMS) for alerts were also integrated to alert the operators to exceptional events (like seismic events).

“We could set up the interfaces to connect structural integrity models with Rapidminer IoT Studio right out of the box,” says Dr. Adam Towse, head of discipline, simulation and assessments, at Assystem. “This empowers operators to leverage the powerful models originally used to design their plant, feeding them the live operational data from the plant throughout its lifecycle to gather critical insights and make decisions.”

Results

Leveraging Rapidminer IoT Studio helped Assystem create a seamless data flow between the plant and the physics-based models to create a true Digital Twin. Incorporating virtual sensors allowed engineers make better decisions regarding support on fatigue damage and helped calculate the remaining useful life (RUL). Insight into the impact of real-world loading conditions improved efficiency and maintenance schedules, further extending the plant’s useful life. Lastly, using the original design models in the Digital Twin solution resulted in a higher return on Assystem’s development investments.