Acid gas removal (AGR) unit is used to remove CO2 and H2S from the sour gas in LNG, refinery and other gas processing plants. Foaming often or suddenly occurs in AGR unit and the operators have to deal with it immediately. Late measures against the foaming increases risks of product off-spec, production turn down, or unscheduled shut down.

In one LNG plant, the operators suffer from foaming hundreds of times yearly. To solve such situation, Chiyoda developed an AI model from plant operation data, which predicts occurrence of foaming. The AI can predict more than half of all foaming cases, at least one hour prior to the occurrence. Foaming Prediction AI System equipped with the foaming prediction AI enables mitigation of production opportunity loss of hundreds of thousands of US$.