Data driven accelerated design of high-efficiency corrosion inhibitor for magnetic refrigeration materials

Water-based heat exchange fluid causes severe corrosion of magnetic refrigeration materials. Here we successfully predict a high-efficiency corrosion inhibitor for magnetic refrigeration by using machine learning trained on a small database. The inhibitor has a significant corrosion inhibition effect, e.g., the La(Fe, Si) ₁₃ alloy maintains perfect surface after 60 days of immersion, whereas Na ₂MoO ₄ and Na ₂HPO ₄ from an inhibitor form the protective film to prevent corrosion. In comparison with the traditional trial-and-error method, the machine learning method could reduce data collection by approximately 26 times, while the prediction accuracy can be improved by approximately 2–3 orders of magnitude.