Research On Fault Detection And Reconstruction Method For Vessel Attitude And Position Measurment System

In recent years, with the increasing requirements for vessels operating at sea, theDynamic Positioning System (DPS) as an important maritime operations support system, itspositioning accuracy and stability requirement is getting higher and higher. The ship controlrequires high reliability on motion state and environmental information, to access the ship’sposition and attitude information, the DP-3system is configured redundant measurementsensor. Dynamic Positioning System often engages in long-term special operations underadverse sea conditions, so the sensor measurement information is vulnerable to fail becauseof the external environment such as wind and waves flow interference. Once the sensor fails,it will seriously impact the control system’s follow-up monitoring and control. So how tomake the dynamic positioning ships achieve or close to the desired control objectives on thesituation of sensors fault or failure has become a significant research. To improve theaccuracy of the measurement information and the performance of the ship dynamicpositioning control system, This paper carried fault detection on sensor subsystem and builtdifferent redundant measurement system. The subsystem passed fault detection canparticipate in multi-level hierarchical data fusion of state estimation. Multi-sensor measuringsystem reconstruction is realized and the DP-3system’s safe operation is ensured.This paper introduced ship position and attitude measuring system of DPS. The shipkinematics model and heading and the position sensor measurement equation is built. Thesimulation experiment indicates the effectiveness of the proposed model.For the fault detection of ship dynamic positioning system sensor measurement system,this paper mainly use the state estimation method based on the analytical model. Theperformance of detection algorithm is determined by residual. In order to construct moreaccurate residuals, state estimation filter method based on Kalman filtering and unscentedKalman filtering is used for estimating heading and position.Considering the disadvantage of residual chi-square detection algorithm has lowsensitivity to gradient fault, this paper adopted a improved fault detection algorithm based oneigenvalue detection. The result of MATLAB simulation shows that the algorithm had highsensitivity to sudden failure and gradient fault. The method could further improve theaccuracy and reliability of ship dynamic positioning system.To take full advantage of the measurement information provided by sensor subsystem,scientific and rational allocation of multi-sensor resources is needed. A hierarchicalmulti-level integration algorithm is established. Hierarchical fusions are obtained using some different state parameters of filters. The next level integration is done for the hierarchicalfusions, and this improves the fusion accuracy. Combined switching control with faultdetection algorithm, subsystem can be isolated or recovered from fusion system. Systemreorganization is achieved ultimately. In order to build a complete ship dynamic positioningpose and position information processing system and achieve the optimal sensor structuremodel, fault detection and system reconfiguration method when there are outliers is studied.Through system reconfiguration, simulation indicates that the system exists fault still has ahigh positioning accuracy. In summary, the fault detection and reconstruction methodproposed in this paper is workable, it enhances ship control systems’ performance in faulttolerant and ensures the safety, reliability, and overall accuracy of the dynamic positioningsystem. The research has both theoretical and practical significance.