Optimization of crude oil blending with neural networks

Crude oil blending is an important unit in petroleum refining industry. Most of blend automation system is a real-time optimizer (RTO). RTO is a model-based optimization approach that uses current process information to update the model and predict the optimal operating policy. But in many oil fields, they hope to make decision and do supervision control based on the history data, i.e., they want to know the optimal inlet flow rates without on-line analyzers. To overcome the drawkack of the conventional RTO, in this paper we use neural networks to model the blending process by the history data. Then the optimization is carried out via the neural model. The contributions of this paper are: (1) we propose a new approach to solve the problem of blending optimization based on history data. (2) Sensitivity analysis of the neural optimization is given. (3) Real data off a oil field is used to show effectiveness of the proposed method.