Deep reinforcement learning uncovers processes for separating azeotropic mixtures without prior knowledge

TL;DR

This paper introduces a deep reinforcement learning approach that autonomously learns to synthesize separation processes for azeotropic mixtures, demonstrating adaptability and near-optimal performance across multiple chemical systems without prior knowledge.

Contribution

A novel general deep reinforcement learning method for flowsheet synthesis that adapts to various chemical systems and learns fundamental engineering paradigms without prior knowledge.

Findings

Agent separates over 99% of materials into pure components

Agent learns fundamental process engineering paradigms

Method applies across multiple chemical systems

Abstract

Process synthesis in chemical engineering is a complex planning problem due to vast search spaces, continuous parameters and the need for generalization. Deep reinforcement learning agents, trained without prior knowledge, have shown to outperform humans in various complex planning problems in recent years. Existing work on reinforcement learning for flowsheet synthesis shows promising concepts, but focuses on narrow problems in a single chemical system, limiting its practicality. We present a general deep reinforcement learning approach for flowsheet synthesis. We demonstrate the adaptability of a single agent to the general task of separating binary azeotropic mixtures. Without prior knowledge, it learns to craft near-optimal flowsheets for multiple chemical systems, considering different feed compositions and conceptual approaches. On average, the agent can separate more than 99% of the involved materials into pure components, while autonomously learning fundamental process engineering paradigms. This highlights the agent's planning flexibility, an encouraging step toward true generality.