Smart Diagnostics For Marine Diesel Engines Using Torsional Vibrations Signals on the Ship Propulsion Shaft-Line

摘要

The diagnostics for a marine diesel engine (MDE) by torsional vibration signal (TVS) on the shaft-line permit classifying and recognizing the observed technical condition (TC) in every cylinder. To determine whether all cylinders are working normally or if any of them of the cylinders works abnormally or misfires. The diagnostics of the MDE during operation at sea has the goal to protect the main propulsion plant (MPP) from dangerous shaft-line TVS levels.This paper presents classified algorithms based on the optimization of functional goals with the Mahalanobis distance instead of the conventional distance Euclid for a series of diagnostic tasks, such as: confidence testing the input database; making standard (reference) characteristics of the technical condition classes; making classification of the observed TC based on the referenced database with one of the three vectors of diagnostic signs (DSV): The maxima (or minima) of TVS in real time relative to the cylinders firing order, and the characteristics of the TVS in the frequency domain.The proposed classification, diagnostic algorithms were used to build the diagnostic software in Lab- View and were tested for the marine vessel (MV), called by MV.HR.34000DWT. The input databases for classification were simulated by the authors using the Software for Automatic Calculation of the Torsional Vibrations (SATVC) that was made at the Vietnam Maritime University.