Automatic optimization of real CFD engineering problems.
HELYX-Adjoint is a continuous adjoint-based CFD tool for topology and shape optimization. It was designed to solve real engineering optimization problems, complex geometries, handling mesh sizes bigger than 200 million cells.
STREAMLINED WORKFLOW
Designed to deliver surface and volume sensitivities within a single solver execution.
1
Primal solution setup
Refers to the “standard” CFD solution. You generate the mesh, define material properties, boundary conditions, numerical settings and monitoring functions as usual.
2
Adjoint setup
You define the adjoint solution parameters, such as number of iterations, objective functions (minimize power loss, maximize uniformity, etc.), geometry surfaces allowed to change, among other parameters.
3
Single solver execution
HELYX-Adjoint runs both Primal and Adjoint solutions at once, calculating the surface or volume sensitivities based on the objective functions selected.
4
Results
Besides the surface or volume sensitivities fields, HELYX-Adjoint automatically changes the baseline geometry to accommodate the selected objective function target. The result is an optimized STL surface.
UNIQUE FEATURES
Unique features of HELYX-Adjoint.
Single solver execution
HELYX-Adjoint calculates the surface or volume sensitivities within a single solver execution, without the need of an iterative workflow from the user.
Low RAM requirements
Unlike the discrete adjoint method, which requires high memory capacity, the continuous adjoint method requires similar hardware capcity than usual CFD simulations.
Streamlined workflow
HELYX-Adjoint was designed for simplified ease-of-use. With a few additional parameters to a standard CFD simulation, you can successfully run an adjoint optimization case.
Multi-objective
HELYX-Adjoint has a wide set of pre-defined objectives and constraints, that can be combined and used at the same time.
Robust and accurate
HELYX-Adjoint provides 2nd order approximation for the adjoint momentum equation for improved accuracy.
Steady and transient
Sensitivities are calculated from RANS, uRANS, DES, and LES primal calculations to take advantage of steady-state and transient flow solutions.
what HELYX-ADJOINT can do for you
Single solver execution.
Real engineering problems. Streamlined workflow.
HELYX-ADJOINT FAQ
Frequently Asked Questions about HELYX-Adjoint.
What are the benefits of continuous over discrete adjoint formulation?
- Lower Memory Consumption: Continuous adjoint formulation requires significantly less memory compared to the discrete approach. This makes it more feasible for large-scale engineering problems.
- Flexibility: Continuous adjoint methods are derived from the continuous governing equations, providing greater flexibility to adapt to various mesh types and flow regimes.
- Handling of Complex Physics: The continuous approach can better accommodate complex physical phenomena and boundary conditions due to its derivation from the original continuous equations.
Overall, the continuous adjoint formulation is better suited for practical engineering applications, offering a balance of flexibility, efficiency, and lower memory consumption.
As a customer, can I access HELYX-Adjoint source code?
Yes, all our HELYX-Adjoint customers have full access to it’s source code.
Does HELYX-Adjoint only solve simpler cases, like other software?
No. HELYX-Adjoint was specifically designed using the continuous adjoint formulation to address real-world engineering challenges effectively. Our continuous adjoint approach ensures lower memory consumption, making it feasible to handle large-scale problems efficiently. Over the past 10 years, our customers have successfully used HELYX-Adjoint to run cases with more than 200 million cells, demonstrating its robustness and reliability for complex engineering applications.
Can I combine objective functions that yields opposite results (e.g. lift and drag)?
Yes, you can combine multiple objective functions in HELYX-Adjoint and assign different weights to each one. The adjoint solver will then find the best approximate optimal solution that satisfies the combined objective functions. This flexibility allows you to tailor the optimization process to meet specific engineering requirements and achieve balanced results.
References
De Villiers, E. et.al. “Adjoint Optimisation for Vehicle External Aerodynamics”. JSAE 2015.
Othmer, C. “The Adjoint Method Hits the Road: Applications in Car Aerodynamics”. Stanford University 2014.
Othmer, C. “Adjoint Methods for Cars Aerodynamics”. Journal of Mathematics in Industry 2014.
GET STARTED TODAY
Book a discovery meeting to learn more about HELYX-Adjoint.
Consent Notice
By submitting your request, you agree to be contacted by email to receive important announcements from ENGYS, including new software releases, product updates, sponsored events and company news.