Abstract: [Objectives] To accelerate the green and low-carbon transformation of ports, liquefied natural gas (LNG) cold energy demonstrates great potential in port energy utilization. Aiming at the problem of efficient operation control of the organic Rankine cycle cold energy power generation system, and considering the nonlinear and multivariable coupling characteristics of the system, a predictive control strategy for port LNG cold energy power generation system based on a data-driven surrogate model is proposed. [Methods] Firstly, a control-oriented dynamic model of the cold energy power generation system is established, and a rectified linear unit network is used to achieve adaptive piecewise linearization of the nonlinear dynamic model. Secondly, an optimization-based control model is constructed in a receding horizon framework. To enhance the efficiency of online optimization from arbitrary initial conditions, an optimization-based bound tightening method is utilized to determine the big-M parameters in the reformulated model. Finally, the effectiveness of the proposed control strategy is verified by the simulation analysis of system performance. [Results] The proposed control strategy achieves precise dynamic control of power output within the system’s safe operating range, effectively overcoming the limitations of conventional PID control and nonlinear model predictive control. [Conclusions] The proposed control strategy provides an intelligent control solution for LNG cold energy power generation systems, offering significant benefits in enhancing the stability and reliability of clean energy generation systems in ports.

Key words: liquefied natural gas (LNG), cold energy power generation, organic Rankine cycle, model predictive control, low-carbon transformation, port energy utilization, cascade utilization