By Matei Ignuta-Ciuncanu and Ricardo Martinez Botas
Dept of Mechanical Engineering, SETTL Group

Heavy-duty transport (e.g., trucks, ships, off-highway machinery) sits far from the headlines made by electric hypercars and robotaxis. Yet large carriers are the arteries of the world economy, and they will not shed heat engines overnight. In sectors like long-haul freight and port logistics, fast decarbonisation demands a systematic reduction of losses: lowering resistance, improving flow access, and recovering energy that would otherwise dissipate into the ambient.

Internal combustion engines will remain embedded within hybrid powertrains for decades. In that context, waste heat recovery (WHR) is not a legacy idea; it is a critical lever for cutting CO₂ per kilometre. A modern heavy-duty engine rejects more than half of its fuel energy as heat. The challenge is designing components that make constructive use of what flows.

Flow Physics First: Where Hidden Efficiency Resides

For years, however, WHR development has been limited by idealised testing. Heat exchangers are typically validated against sinusoidal pressure waves: smooth, convenient, and entirely unlike the real dynamics of an engine. At Imperial College London, the SETTL group set out to close this gap, striving to understand how to best make use of pulsating flows along a 30-year long journey started by Prof. Martinez-Botas. Using the Transient Air System Rig (TASR), we reproduced true valve lift-off profiles rather than simplified sine waves. The results revealed what industry models had missed. The flow reversal produced by real valve events, long treated as an inconvenience, actually enhances heat transfer. Under real engine conditions, we measured up to a 35% increase from steady-state flow.

For manufacturers of hybrid trucks, this matters. WHR systems sized using sinusoidal assumptions risk leaving measurable efficiency gains on the table. Our findings show that realistic transient flows are not a complication, but an opportunity. This strengthens the case for WHR in heavy-duty hybrids, where even a few percentage points of recovered energy translate into substantial lifetime fuel and CO₂ savings.

DendrON: Learning Design Logic from Nature

Understanding the flow leads to a second question: how should the hardware evolve to carry it? To answer this, we drew from nature's carbon wisdom. From rivers to lungs to snowflakes, the animate and inanimate flow systems follow a universal pattern: they evolve to provide better access to what moves through them by branching into dendritic (tree-like) structures. This is the essence of the constructal law discovered by Adrian Bejan in 1996 [1].

In late 2022, we began a computational experiment that grew into DendrON, a generative design framework that uses AI to emulate nature's adaptive intelligence. In our workflow, each configuration is free to morph, and physics determines what persists [3]. Some structures thicken into trunks to carry large thermal loads; others refine into slender, high-svelteness branches that reach distant regions with minimal resistance [4]. Yet, they all follow the same adaptive intelligence predicted by the constructal law, improving access to what flows (heat, charge, coolant). Examples of design problems solved by DendrON include:

  • Spreading (Leaves): Capillary leaf networks inspired generative bi-material heat-sink designs that improved multi-objective performance by 13% over gradient-based optimisation.
    • Insulating (Mushrooms): Porous mushroom gill structures informed rapid generation of thermal-cloaking metamaterials, produced in under 5 minutes on a standard computer.
    • Cooling (Vascular Systems): Vascular flow logic guided the evolution of battery-cooling networks, cutting hotspot temperature non-uniformity by 37% and pumping power by 50% relative to straight-channel plates.

This predictive capability allows us to design the next generation of WHR and thermal-management components for vehicles: lighter, more effective, and tailored to real operating conditions. At the Vehicle Futures Hub, we invite industry partners to work with us on applying carbon wisdom and silicon intelligence to solve the most pressing societal issue of our time: sustainable transport. In the race to net zero, every Joule counts—and none should be wasted.

[1] M. C. Ignuta-Ciuncanu, J. Michael, S. Qian, C. Noon, and R. F. Martinez-Botas, “Experimental characterization of heat transfer and fluid dynamics in pulsating exhaust flows,” Journal of Turbomachinery, 2025.

[2] A. Bejan, “Constructal-theory network of conducting paths for cooling a heat generating volume,” International Journal of Heat and Mass Transfer, 1997.

[3] M. C. Ignuta-Ciuncanu, H. Stärk, and R. F. Martinez-Botas, “Evolutionary design of conductive pathways using a generative autoencoder,” International Communications in Heat and Mass Transfer, 2025.

[4] M. C. Ignuta-Ciuncanu and R. F. Martinez-Botas, “Discrete svelteness: Evaluating flow structures in generative constructal design,” BioSystems, 2025.

[5] M. C. Ignuta-Ciuncanu, P.Tabor, and R. F. Martinez-Botas, “A generative design framework for passive thermal control with macroscopic metamaterials,” Thermal Science and Engineering Progress, 2024.