Optimal Operation of Integrated Power and Oil Transmission Systems

This paper focuses on the optimal operation of integrated power and oil transmission systems where pumps along the oil pipeline have huge electricity consumption. Considering the contradiction between calculation speed and accuracy, a variable step mixed-integer linear programming (MILP) model is established. Additionally, to describe the diffusion effect of drag reducers on the hydraulic loss, the description based on long short-term memory recurrent neural network (RNN-LSTM) is proposed which can perform sequence-to-sequence nonlinear regression. Case studies show that the model in this paper outperforms the classical model in saving more energy and cost, and the RNN-LSTM can calculate the hydraulic loss with high accuracy. The solution is intuitive for the pipeline dispatcher and the established model is capable of industrial use, meeting the goal of energy conservation and emission reduction to relieve the energy crisis.