Study On Speed Control Strategy For Marine Low-speed Engine Based On Reinforcement Learning

In recent years,with the rapid development of electronic control and intelligent technology,how to use new intelligent technology to improve the safety and energy efficiency of ship navigation has become an important research topic.As the core of a ship’s propulsion system,its reliable and efficient operation is essential for ship navigation.As we all know,the primary problem of marine diesel engines is the stable control of the rotational speed.However,when the operating conditions of the diesel engine change or the components are aging,the original fixed speed controller parameters are liable to fail,which leads to abnormal fluctuations in the rotational speed and even speeding And other dangerous phenomena.Therefore,the effective control of the speed instability phenomenon has important practical significance for improving the navigation safety of ships.This paper takes W?rtsil? RT-flex35 low-speed diesel engine as the research object,and designs a speed control algorithm based on reinforcement learning for the problem of speed instability caused by the failure of control parameters under steady-state conditions.The algorithm uses reinforcement learning theory to manage PI control parameters online.When rotational speed instability occurs,the algorithm automatically adjusts the control parameters based on the action strategy library.At the same time,the action strategy library can learn online to achieve self-adaptation to changes in diesel engine characteristics.This paper uses a model-based design method in the design and development process.First,a real-time simulation model of the research object and a prototype of the control algorithm are established,and the simulation verification of the speed control algorithm is initially completed.Then based on the algorithm,a complete speed control strategy is designed according to the controlled needs of the research object.Finally,relying on the product-level rapid control prototype and automatic code generation toolbox,a hardware-in-the-loop test of the speed control strategy based on reinforcement learning was completed.The simulation and experimental results show that the marine low-speed diesel engine speed control strategy designed in this paper can stably control the low-speed diesel engine speed in the full range of operating conditions.Compared with the traditional PI controller,the PI controller based on reinforcement learning has more Good robustness,can realize theinstability and self-rescue of diesel engine speed.