DualMS: Implicit Dual-Channel Minimal Surface Optimization for Heat Exchanger Design

Weizheng Zhang, Hao Pan, Lin Lü, Xiulan Duan, Xin Yan, Ruonan Wang, Qiang Du

Abstract

Heat exchangers are critical components in a wide range of engineering applications, from energy systems to chemical processing, where efficient thermal management is essential. The design objectives for heat exchangers include maximizing the heat exchange rate while minimizing the pressure drop, requiring both a large interface area and a smooth internal structure. State-of-the-art designs, such as triply periodic minimal surfaces (TPMS), have proven effective in optimizing heat exchange efficiency. However, TPMS designs are constrained by predefined mathematical equations, limiting their adaptability to freeform boundary shapes. Additionally, TPMS structures do not inherently control flow directions, which can lead to flow stagnation and undesirable pressure drops. This paper presents DualMS, a novel computational framework for optimizing dual-channel minimal surfaces specifically for heat exchanger designs in freeform shapes. To the best of our knowledge, this is the first attempt to directly optimize minimal surfaces for two-fluid heat exchangers, rather than relying on TPMS. Our approach formulates the heat exchange maximization problem as a constrained connected maximum cut problem on a graph, with flow constraints guiding the optimization process. To address undesirable pressure drops, we model the minimal surface as a classification boundary separating the two fluids, incorporating an additional regularization term for area minimization. We employ a neural network that maps spatial points to binary flow types, enabling it to classify flow skeletons and automatically determine the surface boundary. DualMS demonstrates greater flexibility in surface topology compared to TPMS and achieves superior thermal performance, with lower pressure drops while maintaining a similar heat exchange rate under the same material cost.

Introduction

In thermal engineering, the ideal heat exchanger maximizes surface area for heat transfer while minimizing the pressure drop that consumes pumping energy. For years, Triply Periodic Minimal Surfaces (TPMS)—mathematically elegant, sponge-like structures such as the Gyroid—have been the gold standard for achieving high surface-area-to-volume ratios. However, TPMS designs are inherently rigid; they are defined by global periodic equations that do not adapt well to irregular, freeform containers. Furthermore, they are flow-agnostic, often creating "dead zones" or sharp turns that lead to fluid stagnation and excessive pressure drops.

Methodology: The DualMS Framework

DualMS introduces a shift from fitting a shape to a formula to "growing" a shape from the intended flow. The framework employs a two-stage optimization process:

1. Dual Flow Skeleton Optimization

The target volume is first discretized into a dense graph. The optimization problem is then formulated as a constrained connected maximum cut problem. This partitions the graph into two intertwined "skeletons," one for each fluid. Crucially, flow constraints are applied to ensure these skeletons align with desired flow directions and avoid dead ends, making the design inherently aerodynamic.

Dual-channel minimal surface optimization within a multi-bend pipe shell, showing the internal structure (left) and the resulting fluid separation (right).

2. Neural Implicit Surface Optimization

With the skeletons established, a Multi-Layer Perceptron (MLP) is trained as a binary classifier. The network maps spatial points to binary flow types (Fluid A or Fluid B). The heat exchanger wall is defined as the decision boundary where the classification flips. To ensure a smooth, minimal surface, the authors incorporate a "smoothness loss" based on Total Variation from Geometric Measure Theory. This forces the neural network to find the smallest possible area that separates the two skeletons, resulting in a smooth, soap-film-like interface.

Results and Discussion

The DualMS framework was validated against several challenging geometric benchmarks, including U-shape domains, zigzag shells, and multi-bend pipes.

DualMS results across various geometric benchmarks: U-shape, zigzag, and multi-bend pipe shells, demonstrating adaptability to complex boundaries.

Key performance metrics include:

  • Pressure Drop Reduction: In U-shape tests, DualMS achieved a 35-40% lower pressure drop than standard Gyroid designs while maintaining comparable heat exchange rates.
  • Efficiency vs. Topology Optimization: Compared to state-of-the-art topology-optimized (TO) designs, DualMS achieved 92% of the heat transfer but with only 29% of the pressure drop.
  • Manufacturability: The feasibility of these complex designs was proven through the successful 3D printing of a U-shape prototype using Selective Laser Melting (SLM).

Conclusion

DualMS represents a significant advancement in heat exchanger design by moving beyond the constraints of periodic mathematical models. By combining graph-based flow optimization with neural implicit modeling, it enables the creation of custom, flow-aware minimal surfaces for any 3D geometry. This flexibility is particularly valuable for industries like aerospace and electric vehicle manufacturing, where optimized thermal management in constrained spaces is critical. Future work will focus on reducing computational overhead and extending the framework to multi-fluid systems.

Summary updated

Abstract

Heat exchangers are critical components in a wide range of engineering applications, from energy systems to chemical processing, where efficient thermal management is essential. The design objectives for heat exchangers include maximizing the heat exchange rate while minimizing the pressure drop, requiring both a large interface area and a smooth internal structure. State-of-the-art designs, such as triply periodic minimal surfaces (TPMS), have proven effective in optimizing heat exchange efficiency. However, TPMS designs are constrained by predefined mathematical equations, limiting their adaptability to freeform boundary shapes. Additionally, TPMS structures do not inherently control flow directions, which can lead to flow stagnation and undesirable pressure drops. This paper presents DualMS, a novel computational framework for optimizing dual-channel minimal surfaces specifically for heat exchanger designs in freeform shapes. To the best of our knowledge, this is the first attempt to directly optimize minimal surfaces for two-fluid heat exchangers, rather than relying on TPMS. Our approach formulates the heat exchange maximization problem as a constrained connected maximum cut problem on a graph, with flow constraints guiding the optimization process. To address undesirable pressure drops, we model the minimal surface as a classification boundary separating the two fluids, incorporating an additional regularization term for area minimization. We employ a neural network that maps spatial points to binary flow types, enabling it to classify flow skeletons and automatically determine the surface boundary. DualMS demonstrates greater flexibility in surface topology compared to TPMS and achieves superior thermal performance, with lower pressure drops while maintaining a similar heat exchange rate under the same material cost.
BibTeX
@misc{zhang2025dualms,
  author = {Weizheng Zhang and Hao Pan and Lin Lü and Xiulan Duan and Xin Yan and Ruonan Wang and Qiang Du},
  title = {{DualMS: Implicit Dual-Channel Minimal Surface Optimization for Heat Exchanger Design}},
  year = {2025},
  eprint = {2504.02830},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2504.02830},
}
Figure