Authors: Michael Rybalko, Aeropropulsion Engineer, Boom Supersonic & Jean-Charles Bonaccorsi, Technical Director, NUMECA USA
Founded in 2014 in Denver, Colorado, Boom Supersonic is redefining what it means to fly by building Overture, history’s fastest commercial airliner. Overture will travel at a speed of Mach 2.2 and a cruising altitude of up to 60,000 feet.
The fastest airliner in history
With pre-orders and options for 30 Overture airliners from Japan Airlines and Virgin Group already booked, the race is on to design the next generation supersonic plane. Besides the challenges that supersonic flight inherently imposes, Boom’s Overture designers also need to consider important environmental and social factors. The UN’s CORSIA climate agreement of carbon-neutral growth requires the offset of all international aviation emissions, whether subsonic or supersonic. To support this, Boom Supersonic plans to accommodate drop-in sustainable alternative fuels that will reduce their carbon footprint by roughly 80% and is actively looking at ways to incorporate environmentally minded innovations into Overture's design, without causing technical risk to their development timeline. One such innovation is its partnership with Prometheus Fuels, a company using clean energy to make zero-net carbon fuels out of carbon dioxide that is already in the atmosphere. Mitigating the community's exposure to the noise of sonic booms is another priority. They will do this by limiting supersonic speeds solely to trans-oceanic flight segments and implementing the latest noise-reducing technologies to ensure no increase of existing noise contours during take-off and landing.
Addressing the challenges
Due to the complexity of designing for supersonic speeds, Boom Supersonic engineers need to be able to test multiple conditions and try out many different design ideas. They also work with a very short time to concept, which means they need a solution that is fast to set up and even faster to get results. In a pilot program with NUMECA, Boom managed to achieve results up to 14 times faster than with their previous design environment, according to Tim Conners, Lead Propulsion Engineer at Boom Supersonic. NUMECA solutions have not only advanced the development of the XB-1 subscale demonstrator by providing a dramatically streamlined and highly automated workflow, but the NUMECA partnership has given Boom the opportunity for significant savings in computational resources and a reduction in design cycle time.
Boom managed to achieve results up to 14 times faster than with their previous design environment.
The NUMECA partnership has given Boom the opportunity for significant savings in computational resources and a reduction in design cycle time.
Tim Conners, Lead Propulsion Engineer at Boom Supersonic
The unstructured hex-dominant meshing tool HEXPRESS™/Hybrid and the unstructured flow solver FINE™/Open with OpenLabs™ were selected for this task. Adding to the reduction of computational resources, Boom has been able to take advantage of CPUBooster™, NUMECA’s unique convergence acceleration technique, for the majority of their runs.
Throughout the past year, several design studies used NUMECA solutions for both XB-1 and Overture, including:
- Inlet bleed plenum outflow venturi sizing and secondary flow path performance validation. These small cases were initially run on a laptop and later transitioned to the Rescale cloud computing platform as some of the first tests of the workflow.
- Isolated ejector nozzle simulations, for solver verification, which when compared with the previous in-house approach yielded a percent difference of 0.1% for nozzle gross thrust coefficient at cruise. Simulations with meshes of up to 100 million cells were run on 360 cores on Rescale.
- Cooling door sizing study. Results of 24 cases with 100 million cells were used to characterize the flow pumping characteristics of the ejector nozzle and helped match the flow schedules of the inlet and nozzle.
- Overture wing/body and wing/body/nacelle simulations. Meshes of 200-250 million cells were generated for a full-span model, and simulations were performed at cruise conditions to compare viscous results with lower fidelity preliminary design tools and inviscid simulations.
- Ejector nozzle analysis spanning the flight envelope (developed beside).
Ejector nozzle analysis
This analysis was part of the development of the XB-1 demonstrator. The objective was to run an analysis of the ejector nozzle spanning the flight envelope.
To reduce engineering time and streamline the meshing process, the integration started at CAD level by naming each part/surface of the geometry. The CAD file was then transferred to HEXPRESS™/ Hybrid, which applies specific mesh refinements based on the name of each part/surface. The naming is automatically performed in the CAD system, so each new iteration does not require an additional adaptation of the input geometry, saving a significant amount of engineering time.
Specific parts of the mesh can be easily swapped and connected to the rest of the geometry, automatically and in batch mode. The variable engine nozzle was set to the appropriate area for each condition, resulting in 24 different geometries.
A single ASCII input file was used to generate a high-quality mesh for each of the configurations, ensuring mesh consistency and again reducing engineering time. HEXPRESS™/Hybrid automatically captured the details of the modified geometry and repaired tiny gaps which can appear when changing the design.
The latest developments for viscous layers insertion with smooth transition were used in the near-wall region, with the wall distance of the first cell in the viscous sublayer specified for use with a low Reynolds number turbulence model.
The meshing process was done in parallel with a single in-house workstation, generating 24 meshes of 150 million cells each over the course of a weekend.
The nozzle was simulated across the mission envelope from M0.02 at 5,000 feet to M2.2 at 40,000 feet. All cases were simulated at maximum dry thrust and with full afterburner. Five operating points for each of the 24 geometries were analyzed, resulting in 120 3D RANS simulations.
Analysis on the 150 million cell meshes was performed using 360 cores per job and was largely completed within a week and a half (3-4 hours per job). CPUBoosterTM, multi-grid acceleration, and OpenLabs™ were used in order to improve convergence and reduce CPU time.
The results of these simulations provided key information applicable to a variety of multi-disciplinary design aspects, from performance to structural and hardware components allowing for evaluation and optimization of key elements such as:
- Flow pumping characteristics of the ejector nozzle
- Nozzle gross thrust coefficient used in the engine cycle and aircraft performance models
- Nozzle axial loads for nozzle attachment hardware design
- Temperature ranges and pressure deltas for material selection and structural design validation
According to Boom’s engineers, this specific analysis provided the most impactful results generated to-date with NUMECA tools and highlights how well these NUMECA tools can work within a stable workflow.