| The Controllable Pitch Propeller(CPP),as the primary propulsion system of large ships,is a complex servo-control mechanism composed of machinery,electrical control devices,and hydraulic units.Its intricate structure and variable operating conditions often lead to frequent failures.These failures are typically characterized by their randomness,diversity,and nonlinear relationship between cause and effect.Given the critical role that the CPP plays in ship propulsion,it is of paramount importance to research its fault diagnosis to ensure the reliable long-term operation of the ship,and to support condition-based maintenance and the development of unmanned cabins.In this paper,the authors analyze the CPP’s fault modes and build a fault simulation platform using AEMSim/Simulink.The paper then studies the fault diagnosis models based on the Belief Rule Base(BRB)algorithm.To begin with,the authors provide an overview of the CPP’s operating principle and structure,and analyze the causes and locations of the two main fault modes,"pitching failure" and "pitching sluggish".This paper also identifies eight faults as the research items for fault diagnosis.To overcome the limitations of operating data deficiency and fault data shortages,the authors focus on the load properties of CPP’s,the leakage features of the oil distribution box,and the flow performance of the proportional directional valve.Based on this analysis,the authors build a system for CPP fault simulation using the AEMSim/Simulink co-simulation platform.Through fault simulation tests for eight main faults,the authors obtain simulation operation data for use in the subsequent fault diagnosis of the CPP.Next,the authors introduce the composition and reasoning mechanism of the traditional BRB algorithm and propose an improved algorithm based on joint structure and particle swarm parameter optimization that is better suited for CPP devices.Through nonlinear function fitting and oil supply circuit fault diagnosis examples,the authors demonstrate the improved algorithm’s high diagnostic accuracy,strong generalization ability,and low data dependence.Using the testing data from the simulation platform,the authors build a fault diagnosis model for the CPP based on the improved BRB algorithm and achieve a high diagnosis accuracy rate.Finally,the authors briefly introduce the CPP test bench and perform a simulation test for four main faults.By using the field data obtained from the tests,this paper verifies the effectiveness of the CPP fault diagnosis model and confirms its accuracy of 96%. |
