Data-Driven Dynamic Modeling and Fault Diagnosis of Dimethyl Oxalate Industrial Production Process

Ethylene glycol (EG) is an indispensable substance in the chemical industry and polyester fiber supply chain. The synthesis of dimethyl oxalate (DMO) by carbon monoxide gas-phase catalytic coupling is a key step in the coal-based syngas to ethylene glycol process route. The strong coupling between the reaction unit and the feedstock regeneration unit, as well as the risk of feedstock gas explosion, poses a great challenge to the stability control and safety of the dimethyl oxalate production process. In this paper, the dimethyl oxalate production process is studied from three aspects: steady-state modeling, dynamic modeling and fault simulation. First, the dimethyl oxalate pro-duction process was comprehensively modeled using Aspen Plus software. Second, a dynamic model was constructed on the basis of the steady-state model to fit the actual production process. Finally, deep learning algorithms were combined with dynamic simulation techniques. Using the fault scenarios and data in the dynamic model, the T-DOAE algorithm is used to study the fault detection in the production process of dimethyl oxalate, which is of great significance to ensure the safe and stable operation of the gas-phase coupling process of dimethyl oxalate production in the process of coal chemical industry.