Periodic Reporting for period 1 - Training42Phase (Next generation turbomachinery with two-phase flow)
Okres sprawozdawczy: 2023-12-01 do 2025-11-30
Training42Phase addresses critical technological challenges in the energy transition by advancing two-phase turbomachinery for next-generation energy-conversion systems. Two-phase turbomachines are flow devices operating with fluids that are partly liquid and partly vapor, and lie at the core of key emerging technologies, such as partial-evaporation organic Rankine cycles, supercritical carbon dioxide (CO2) Brayton cycles, liquid-air energy storage, and transcritical CO2 heat pumps.
Currently, a significant portion of energy in these systems is lost through valves or inefficient heat transfer processes. Two-phase turbo-expanders, pumps, and ejectors offer the potential to recover part of this energy, improving system efficiency and flexibility. However, operating under two-phase conditions introduces challenges such as losses, component wear by erosion, and reduced reliability. Understanding and managing these effects is therefore essential to ensure safe, durable, and cost-effective energy systems.
The main objective of Training42Phase is to develop improved design methods, modelling tools, and maintenance strategies for future two-phase turbomachinery, while simultaneously training a new generation of highly skilled researchers. The project supports 12 doctoral candidates (DCs), providing them with interdisciplinary expertise spanning thermodynamics, fluid mechanics, structural mechanics, reliability analysis, and data-driven methods, ultimately preparing them for careers in both academia and industry. Traning42Phase includes 4 academic partners each employing 2-4 DCs and one industrial beneficiary employing 1 DC, supported by 6 associated industrial partners.
Currently, a significant portion of energy in these systems is lost through valves or inefficient heat transfer processes. Two-phase turbo-expanders, pumps, and ejectors offer the potential to recover part of this energy, improving system efficiency and flexibility. However, operating under two-phase conditions introduces challenges such as losses, component wear by erosion, and reduced reliability. Understanding and managing these effects is therefore essential to ensure safe, durable, and cost-effective energy systems.
The main objective of Training42Phase is to develop improved design methods, modelling tools, and maintenance strategies for future two-phase turbomachinery, while simultaneously training a new generation of highly skilled researchers. The project supports 12 doctoral candidates (DCs), providing them with interdisciplinary expertise spanning thermodynamics, fluid mechanics, structural mechanics, reliability analysis, and data-driven methods, ultimately preparing them for careers in both academia and industry. Traning42Phase includes 4 academic partners each employing 2-4 DCs and one industrial beneficiary employing 1 DC, supported by 6 associated industrial partners.
The research activities are organised into three technical work packages (WPs), each closely linked to the individual research projects of the doctoral candidates.
WP4 – Aero-thermodynamics
Doctoral candidates working in WP4 developed new modelling tools to predict two-phase flows through turbines, compressors, and nozzles. Their work includes the development of low-fidelity simulation tools to predict performance and losses, as well as high-fidelity computer models to accurately describe phenomena such as cavitation, condensation, and vaporisation occurring within high-speed flows. These results support the design of more efficient components for power generation, heat pumps, and energy storage applications.
WP5 – Structural mechanics
In WP5, doctoral candidates focused on analysing how droplets and two-phase flows affect solid components, such as turbine blades. They developed and validated new simulation approaches that are much faster than existing methods, while remaining accurate. Their work also showed how erosion changes the dynamic behaviour of blades, providing valuable insights for safer designs and longer component lifetimes. Experimental validation activities are currently ongoing.
WP6 – Reliability and maintenance
WP6 doctoral candidates worked on predicting and reducing failure risks in two-phase turbomachinery. They developed new data-driven reliability and optimisation methods that account for uncertainty and rare but critical failure events. To overcome the limited availability of experimental data, they have applied physics-informed machine learning techniques to support more reliable design decisions.
The DCs have played a central role across all WPs, producing new scientific knowledge, developing tools relevant to both industry and academia, and building strong interdisciplinary skills during the first 24 months of the project. To date, their work has been disseminated through two peer‑reviewed conference proceedings and two peer‑reviewed journal publications, with several additional manuscripts underway. Project activities were disseminated on social media (e.g. LinkedIn) and on the project website.
In addition to scientific work, they have participated in four extended training programs, covering transferable skills, advanced thermo-fluid-dynamics, optimisation methods, artificial intelligence applied to energy systems, as well as including visits to industrial plants and laboratories across Europe.
WP4 – Aero-thermodynamics
Doctoral candidates working in WP4 developed new modelling tools to predict two-phase flows through turbines, compressors, and nozzles. Their work includes the development of low-fidelity simulation tools to predict performance and losses, as well as high-fidelity computer models to accurately describe phenomena such as cavitation, condensation, and vaporisation occurring within high-speed flows. These results support the design of more efficient components for power generation, heat pumps, and energy storage applications.
WP5 – Structural mechanics
In WP5, doctoral candidates focused on analysing how droplets and two-phase flows affect solid components, such as turbine blades. They developed and validated new simulation approaches that are much faster than existing methods, while remaining accurate. Their work also showed how erosion changes the dynamic behaviour of blades, providing valuable insights for safer designs and longer component lifetimes. Experimental validation activities are currently ongoing.
WP6 – Reliability and maintenance
WP6 doctoral candidates worked on predicting and reducing failure risks in two-phase turbomachinery. They developed new data-driven reliability and optimisation methods that account for uncertainty and rare but critical failure events. To overcome the limited availability of experimental data, they have applied physics-informed machine learning techniques to support more reliable design decisions.
The DCs have played a central role across all WPs, producing new scientific knowledge, developing tools relevant to both industry and academia, and building strong interdisciplinary skills during the first 24 months of the project. To date, their work has been disseminated through two peer‑reviewed conference proceedings and two peer‑reviewed journal publications, with several additional manuscripts underway. Project activities were disseminated on social media (e.g. LinkedIn) and on the project website.
In addition to scientific work, they have participated in four extended training programs, covering transferable skills, advanced thermo-fluid-dynamics, optimisation methods, artificial intelligence applied to energy systems, as well as including visits to industrial plants and laboratories across Europe.
During the first reporting period, Training42Phase developed new tools and methods that advance two-phase turbomachinery beyond current capabilities.
In WP4, the project developed a new meanline approach enabling faster and more accurate design optimization than conventional methods. The project also extended the Method of Characteristics to two-phase flows for the first time, supporting the design of components such as supersonic nozzles and turbine blades. Moreover, a new one-dimensional solver was developed for two-phase nozzles, enabling the prediction of how shock waves impact efficiency. Finally, newly developed high-fidelity computational models predict non-equilibrium phase changes in compressors and turbines, providing more reliable performance predictions.
In WP5, the project has introduced methods that improve both computational accuracy and efficiency. A new simulation approach for droplet impacts reduces computing time by up to 96% while avoiding physical inconsistencies present in previous techniques. Further, new reduced-order blade models predict how erosion affects mass, stiffness, and dynamic behaviour, allowing designers to explore more options without running expensive full-scale simulations.
In WP6, the project has developed tools to make two-phase turbomachinery more robust. These include a method to estimate rare failure events, the first application of Bayesian robust optimization to turbine blades, and the use of machine learning to improve design predictions even when experimental data are limited.
Taken together, these results represent significant progress beyond the state of the art of two-phase turbomachinery. They provide new ways to design and operate two-phase turbomachinery that are more efficient and reliable, helping to accelerate the development of sustainable and high-performance novel energy technologies.
In WP4, the project developed a new meanline approach enabling faster and more accurate design optimization than conventional methods. The project also extended the Method of Characteristics to two-phase flows for the first time, supporting the design of components such as supersonic nozzles and turbine blades. Moreover, a new one-dimensional solver was developed for two-phase nozzles, enabling the prediction of how shock waves impact efficiency. Finally, newly developed high-fidelity computational models predict non-equilibrium phase changes in compressors and turbines, providing more reliable performance predictions.
In WP5, the project has introduced methods that improve both computational accuracy and efficiency. A new simulation approach for droplet impacts reduces computing time by up to 96% while avoiding physical inconsistencies present in previous techniques. Further, new reduced-order blade models predict how erosion affects mass, stiffness, and dynamic behaviour, allowing designers to explore more options without running expensive full-scale simulations.
In WP6, the project has developed tools to make two-phase turbomachinery more robust. These include a method to estimate rare failure events, the first application of Bayesian robust optimization to turbine blades, and the use of machine learning to improve design predictions even when experimental data are limited.
Taken together, these results represent significant progress beyond the state of the art of two-phase turbomachinery. They provide new ways to design and operate two-phase turbomachinery that are more efficient and reliable, helping to accelerate the development of sustainable and high-performance novel energy technologies.