Safran Helicopter Engines, previously known as Turbomeca, is a French manufacturer of low- and medium-power gas turbine turboshaft engines for helicopters. The company also produces gas turbine engines for aircraft and missiles, as well as turbines for land, industrial and marine applications.Since its founding as Turbomeca during 1938, Safran Helicopter Engines has produced over 72,000 turbines. In its early years, it benefitted greatly from a rearmament programme conducted by the French state; operations were disrupted by the occupation of France during the Second World War, but the company survived and rebuilt quickly during the immediate postwar years. Prominent successes during the Cold War include the use of its Artouste II turboshaft engine to power the new Sud Aviation Alouette II helicopter (the first production turbine-powered helicopter in the world) as well as its involvement in Rolls-Royce Turbomeca Limited (a joint venture with British engine manufacturer Rolls Royce Ltd that produced turbojet and turboshaft engines).During September 2001, the French aerospace specialist SNECMA Group acquired the company, after which it was rebranded as Safran Helicopter Engines. The company states that it has more than 2,500 customers in 155 countries. Safran Helicopter Engines has 15 sites and operates on each continent, providing its customers with a proximity service through 44 distributors and certified maintenance centers, 18 Repair & Overhaul Centers, and 90 Field Representatives and Field Technicians. Safran Helicopter Engines subsidiary Safran Power Units is the leading European manufacturer of turbojet engines for missiles, drones and auxiliary power units. Safran Helicopter Engines has 6,300 employees worldwide, with 5000 based in France. In 2015, the company reportedly produced and delivered 718 new engines, and repaired around 1,700 engines..
In transonic turbine stages, complex interactions between trailing edge shocks from nozzle guide vanes and rotor blades generate unsteady wall pressure fields, impacting rotor aerodynamic performance and structural integrity. While shock-related phenomena are prominent, unsteady pressure fluctuations can also arise in subsonic regimes from wake interactions. Traditional methods like Unsteady Reynolds-Averaged Navier-Stokes (URANS) simulations are accurate but computationally expensive. To address this, a novel deep learning-based Reduced Order Model (ROM) is proposed, built on a database of URANS simulations, to predict unsteady pressure fields on turbine rotor blades at a fraction of the cost. The model consists of a Variational Auto-Encoder (VAE) integrated with a Gated Recurrent Unit (GRU) to capture time-series data, overcoming the limitations of traditional linear ROMs in capturing nonlinear phenomena, such as moving shocks. The goal is to develop a ROM that accurately reproduces unsteady pressure fields from URANS simulations while reducing computational costs. The ROM is applied to the Turbine Aero-Thermal External Flows (TATEF2) project configuration, a representative test case in turbomachinery research. Model performance is evaluated using machine learning quality metrics and design-oriented criteria, including the accuracy of the first harmonic in the Fourier transform of the unsteady pressure field. The impact of the simulation database size on model accuracy is also analyzed, considering the number of training simulations required for task-specific accuracy as a key factor in industrial applicability.
The Rich burn-Quick mix-Lean burn (RQL) concept is a staged combustion technology that guarantees flame stability at all operating conditions while significantly reducing the concentration of pollutants at the outlet of a combustion chamber. A lab-scale RQL combustion module equipped with a new-generation double-circuit fuel injection system and large optical accesses is designed to be integrated into the visualization module of a highpressure combustion facility fed with aeronautic multi-component liquid fuels. This new optical RQL module will be devoted to perform a simultaneous study into the three RQL sections of the physico-chemical processes involved in soot production and oxidation and NOx formation under realistic high-pressure / high-temperature operating conditions encountered in helicopter combustors by means of advanced coupled laser-based diagnostics. To this end, the architecture of the RQL module has been designed to satisfy various constraints such as a reactive zone height comparable to that of a helicopter combustor sector, an overall pressure drop of similar to 3 %, realistic fuel / air mass flowrates, an efficient heat transfer between the walls and the flame and the ability to introduce laser sheets through the RQL areas for optical measurements. The optimized geometry of the RQL module was achieved by performing iterative Large-Eddy Simulations (LES) of the reactive flowfield of a kerosene vapor / air mixture with the AVBP numerical solver at a nominal operating condition (13 bar), then by LES simulations performed with a liquid kerosene / air mixture. Numerical LES results obtained from both singlephase and two-phase reactive flows are discussed. They highlight the ability to provide distinct RQL areas, a Vshaped jet opening at the injector outlet, a swirling flame topology, a predominant premixed / partially-premixed combustion regime in the primary rich area, as well as a high combustion efficiency of similar to 100 % at the outlet of the combustion chamber. Experimental results of kerosene / air flame emission at 8 bar and 11 bar confirmed the potential to produce well-distinct RQL combustion areas accessible with laser-based diagnostics while ensuring the combustor's thermal resistance. Finally, the effectiveness of the iterative LES methodology adopted to design the lab-scale optical RQL combustion chamber was also highlighted.
Accurate prediction of unsteady aerodynamic loads remains a major challenge in turbomachinery design. High-fidelity Computational Fluid Dynamics (CFD) simulations are expensive, while aeroelastic Quantities of Interest (QoI) depend sensitively on the temporal evolution of the pressure field. This work introduces periodic Fourier Neural Mapping (p-FNM), a neural-operator framework for predicting unsteady pressure distributions on turbine rotor blades simulated using the chorochronic numerical hypothesis. The architecture embeds temporal periodicity into the model and learns a continuous mapping from operating conditions and time to pressure fields. Unlike sequential latent-space approaches, p-FNM predicts pressure fields independently at any time, avoiding error accumulation while preserving temporal continuity. The model is evaluated on a database of unsteady rotor-blade simulations and compared with a reduced-order baseline based on a variational autoencoder and recurrent neural network, refered as the Temporal Prediction Model (TPM). Performance is assessed for pressure fields and Generalized Aerodynamic Forces (GAFs), the primary aeroelastic QoI. Across all training datasets, p-FNM consistently outperforms TPM. On the largest dataset, p-FNM achieves a pressure-field mean absolute percentage error of 0.46
Abstract Ignition and light-around of lean hydrogen-methane flames was investigated in an atmospheric annular combustor which consists of six equally spaced injectors oriented tangentially to the annulus. This produces flames and flows that are predominantly in the azimuthal direction known as the Spinning Combustion Technology (SCT). Ignition limits and characteristic light-around times were investigated across a range of operating conditions to understand the main mechanisms governing the light-around times under fully premixed injection conditions. The effect of bulk velocity, equivalence ratio, H2 volume fraction from 0 to 100% and effusion cooling were measured and characterised using highspeed flame imaging and an array of photomultipliers with OH* filters. For each operating condition, 25 runs were performed to ensure repeatability resulting in a large dataset that includes 2450 ignition sequences. Results show that both the mean azimuthal velocity Uθ as well as changes in flame speed SL and thermal expansion ratio ρu/ρb of the fuel mixture affect the light-around time. However, overlapping measurements with experiments conducted by replacing the injectors with premixed bluff body axially oriented injectors shows that the light-around time for the SCT configuration is almost twice as fast which is attributed to the bulk swirl induced by the SCT burner orientation. Finally, it is demonstrated that the level of effusion cooling increases light-around times and reduces the ignition limits. These results validate the SCT with reduced light-around times and extended lean ignition limits.
This paper presents SOMA, a multi-agent digital twin designed to support collaborative decision-making in helicopter engine maintenance. The system addresses challenges related to information loss and coordination among experts by integrating multi-agent modeling and stigmergic mechanisms. SOMA enables structured interactions between human and artificial agents, improving situational awareness and organizational learning. A prototype was developed and tested with maintenance professionals from Safran Helicopter Engine company. Results show that the system is relevant and that stigmergic indicators, i.e., visual signals based on past experiences, can influence decision-making. These findings suggest that SOMA can enhance maintenance processes by supporting effective collaboration and decision support.
