arXiv：cs.LG· Arthur Jessop, Mohammed Alsubeihi, Ben Moseley, Ashwin Kumar Rajagopalan·· 13 小时前
可微分混合建模：从实验数据学习并优化化学传输过程
Differentiable Hybrid Modelling for Learning and Optimising Chemical Transport Processes from Experimental Data
AI 导读
研究者提出一种可微分混合建模框架，将神经网络嵌入 GPU 加速的种群平衡方程求解器中，直接从结晶实验数据中发现本构定律并同化初始条件。基于 13 组真实多维尺寸与形状测量实验，同化初始条件可恢复物理一致的非线性速率定律，轨迹误差较传统建模降低七倍。端到端可微分性还支持对 300 变量温度曲线进行梯度优化，在质量不变条件下使晶体长宽比降低 25%。
正文
Abstract:Transport models underpin the design and optimisation of chemical processes, yet their constitutive laws are typically hand-picked, their experimental initial conditions are assumed exact, and their computational cost restricts optimisation to a handful of process variables. Here we introduce a differentiable hybrid modelling framework that embeds neural networks within a GPU-accelerated population balance equation solver to discover constitutive laws and assimilate initial conditions directly from experimental crystallisation data. Using thirteen real experiments with multi-dimensional size and shape measurements, we show that treating measured initial conditions as exact yields a physically inconsistent rate law despite improved predictive accuracy, whereas assimilating them recovers a physically consistent, non-linear law and reduces trajectory error sevenfold relative to conventional modelling. End-to-end differentiability further enables gradient-based optimisation of high-dimensional operating profiles, identifying a 300-variable temperature profile that reduces crystal aspect ratio by 25% at constant mass, facilitating superior product attributes. Our framework augments, rather than replaces, mechanistic modelling, offering a general route to learning transport models from data and optimising processes with them.
| Subjects: | Computational Engineering, Finance, and Science (cs.CE); Machine Learning (cs.LG); Numerical Analysis (math.NA) |
| Cite as: | arXiv:2609.04011 [cs.CE] |
| (or arXiv:2609.04011v2 [cs.CE] for this version) | |
| https://doi.org/10.48550/arXiv.2609.04011 arXiv-issued DOI via DataCite |
Submission history
From: Ashwin Kumar Rajagopalan [view email]
[v1]
Thu, 3 Sep 2026 15:50:56 UTC (2,639 KB)
[v2]
Fri, 9 Oct 2026 14:43:00 UTC (2,911 KB)
来源：arXiv：cs.LG · arxiv.org