South China University of Technology and Xi’an Jiaotong University Introduce ‘HULU’ to Accelerate Clean Energy Material Discovery

August 2, 2026

EurekAlert! China

Overview

Researchers from South China University of Technology and Xi’an Jiaotong University have developed ‘HULU,’ a flexible Python framework integrating advanced Machine Learning Potentials (MLPs) with Monte Carlo adsorption simulations. This innovation significantly accelerates clean energy material discovery by natively enabling high-accuracy AI models within simulation software. HULU promises more precise predictions for materials in applications like atmospheric water harvesting and hydrogen storage, drastically boosting development efficiency.

In Depth

Background

Clean energy technologies, including atmospheric water harvesting, hydrogen storage, and CO2 capture, are vital for a sustainable future. However, their efficacy hinges on the adsorption properties of constituent materials. The challenge lies in discovering ideal adsorbent materials and accurately predicting their performance, a task complicated by an immense material search space and intricate physicochemical interactions. Traditional simulation methods have typically fallen short, either due to insufficient accuracy or prohibitive computational costs. The HULU framework addresses this critical bottleneck by fusing AI with high-performance computing, thereby accelerating innovation in the clean energy sector.

Key Findings

A collaborative research team from South China University of Technology and Xi’an Jiaotong University has unveiled ‘HULU,’ a flexible Python framework engineered for the seamless integration of advanced Machine Learning Potentials (MLPs) with Monte Carlo adsorption simulations. This groundbreaking framework dramatically speeds up the discovery process for clean energy materials by making high-accuracy AI models natively compatible with existing adsorption simulation software. HULU is poised to enable significantly more accurate predictions for material design in crucial applications such as atmospheric water harvesting and high-efficiency hydrogen storage, promising a substantial boost in development efficiency.

Technical Details

The core innovation of the HULU framework lies in its ability to leverage the full potential of MLPs, which deliver an accuracy comparable to first-principles calculations (Density Functional Theory – DFT) while achieving the computational speed characteristic of classical molecular dynamics (MD) simulations. Conventional adsorption simulations have long grappled with a trade-off between accuracy and computational expense; HULU resolves this by directly embedding the high accuracy of MLPs into Monte Carlo adsorption simulations. Specifically, HULU employs interatomic interaction energies, calculated by MLPs, to precisely predict adsorbate configurations, adsorption amounts, and adsorption isotherms across a range of temperatures and pressures. Its Python-based architecture ensures that researchers can readily integrate and customize HULU within their existing computational workflows. This integration dramatically curtails the iterative trial-and-error process in material exploration, thereby streamlining the optimization of adsorption properties, particularly in complex porous and composite materials.

Strategic Significance & Outlook

Beyond its immediate impact on clean energy material discovery, the HULU framework is envisioned as a versatile tool applicable to numerous other fields where adsorption phenomena are critical, including chemical engineering, environmental science, and pharmaceutical development. The research team has articulated plans to further expand HULU’s functionalities and strengthen its integration with diverse types of MLPs and simulation methodologies. This forward-looking approach will accelerate the realization of ‘AI-driven material discovery,’ where AI autonomously designs adsorbent materials and accurately predicts their performance. HULU is positioned to become a foundational technology, indispensable for achieving a truly sustainable society.

Source: https://www.eurekalert.org/news-releases/1137983

Get our weekly technology intelligence — free

Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.

Subscribe Free — Weekly Tech Intelligence

By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.