I explore artificial intelligence methods (metaheuristics, machine learning) for the shape and topology optimisation of lightweight mechanical structures especially sound-absorbing porous material structures.
I find happiness in helping students understand complex technical concepts deeply. I make interactive web demos to make learning fun. Check them out if you are interested.🙂
I am fascinated by science and nature. I love learning new and
interesting concepts. If you do too, you might like a compilation of
concepts that I found fascinating. If you think I
should check something out,
let me know
! 🙂
Topology optimisation, metaheuristics, acoustics, and additive manufacturing.
Historically, the geometric shapes of mechanical parts and structures, such as an aircraft wingbox, have been designed through a trial-and-error process followed by iterative improvements to existing designs. A breakthrough in geometric design unfolded in the 1990s through the introduction of structural topology optimisation techniques, which allow generation of optimal shapes and topologies for mechanical structures from scratch. Many parts and products have undergone significant innovations due to topology optimisation. A notable example is the reduction in weight of 1000 kilograms per aircraft in Airbus A380 (Krog et al. 2012).
Finding the most optimal shape and topology for a given problem in terms of loads and boundary conditions is computationally hard and there are no practicable algorithms available in many problem cases. Whereas finding near-optimal solutions that are good enough in a quick enough time can be achieved through adjoint methods. These gradient-based methods can get stuck at locally optimal solutions if the mathematical structure of the objective function is multimodal. The problem of local optimal solutions can be addressed through heuristics and metaheuristics strategies which constitute a subset of artificial intelligence.
The overall research aim is to improve topology optimisation techniques through the use of AI and extend the practicability of these techniques.
Ramamoorthy, V.T., Özcan, E., Parkes, A.J., Luc, J. and Bécot, F.X., 2019. Metaheuristic Optimisation of multilayered porous materials, proceedings of the 23rd International Congress on Acoustics, RWTH Aachen, Germany.
Ramamoorthy, V.T., Özcan, E., Parkes, A.J., Luc, J. and Bécot, F.X., 2020. On topology optimisation of acoustic porous materials, e-Forum Acusticum 2020, Lyon, France.
Ramamoorthy, V.T., Özcan, E., Parkes, A.J., Luc, J. and Bécot, F.X., 2021. The OR society conference-OR63.
Parkes, A.J., Ramamoorthy, V.T., Özcan, E., 2021. Bi-objective Search for Acoustic Topology Optimisation and Noise Reduction, EURO 2022 - Espoo.
Gradient-based approaches are quite efficient than metaheuristic strategies applied to the discrete topology optimisation problem. But metaheuristics used in conjunction with gradient-based methods have been able to overcome the problem of local optimality by outperforming gradient strategies for the same computational time (Ramamoorthy et al. 2022).
Manufacturability forms a roadblock in extending the use of topology-optimised shapes in certain applications. While advancements in additive manufacturing have ensured the feasibility of manufacturing topology-optimised structures, the fabrication time and tool costs involved vs. conventional manufacturing methods needs due consideration to assess which approach is indeed the truly optimal in engineering practice. Hence, it is of interest to develop approaches to deal with manufacturability within the algorithm.
If you find these topics interesting and are looking to discuss or collaborate, hesitate not to contact me. 🙂
Game perspective is the paradigm of learning through games or interactive demos. While this is not a replacement for conventional learning (books, lectures, exercises), it is a valuable alternative teaching method.
With the ever-growing forest of human knowledge, there are too many exciting things to learn, and life is too short.
With the help of games or interactive demos, knowledge could be assimilated faster and in a more fun way allowing us to learn more in a given lifetime. Check out some of my works available in the web projects section. Thanks to Luc Jaouen for inspiring me to make these. Check out his works at Acoustic Porous Material Recipes.
A shoutout to Dan Russell for his amazing work.
These are some of the concepts that I found intriguing. It is always interesting to ask people what concepts fascinate them the most, as I always learn something new. Every person has their list, and here is mine. If you think I need to include something, let me know. 🙂