Reflux provides an AI‑driven interface for chemical process simulation tools like Aspen Plus, automating the setup, editing, convergence checking, and result summarization of simulation runs. Engineers can trigger workflows via a desktop UI, CLI, or API to generate optimized designs—such as low‑energy operating points—while the platform handles file management and detailed stream analysis. This accelerates process development by offloading routine simulation tasks and delivering clear trade‑off summaries.
Funding
Funding not disclosed
Founders
Product
Problem
Process engineers spend significant time manually setting up and editing Aspen Plus simulations, checking convergence, and extracting results, which slows development and limits the ability to run iterative studies efficiently.
Solution
Reflux provides an AI‑driven workflow that automates the creation, modification, and execution of Aspen Plus cases while preserving native simulator files. The platform builds flowsheets step‑by‑step in a visible UI, runs convergence checks, and generates stream summaries and sensitivity analyses. Engineers can trigger runs via a control panel, CLI, or external tools, and receive reviewed artifacts—including native archives, convergence logs, and CSV summaries—for audit and decision‑making. By keeping the simulator central and exposing only controlled actions, Reflux ensures transparency and repeatability while freeing engineers to focus on interpreting results and optimizing operating conditions.
Target Audience
Primary users are chemical process engineers and simulation specialists who develop and optimize distillation trains and other unit operations using Aspen Plus.
Features
- Visible, step‑by‑step flowsheet construction that records each edit and spec change
- Automated convergence monitoring and native Aspen Plus case saving (APWZ files)
- Generation of stream‑summary CSVs and sensitivity reports with dominant constraint identification
- Adapter layer that isolates read‑only data gathering from write actions behind an explicit review boundary
- CLI and API triggers for integration with control panels, MCP tools, or custom scripts
- Parallel execution of multiple studies with artifact logging for auditability