Engine state monitoring and fault diagnosis of large marine diesel engines

Abstract

The reliable detection of engine malfunctions in order to predict and to plan maintenance intervals is of major importance in various fields of industry. For instance, occurring faults of marine diesel engines which are on the high seas for months may lead to expensive holding times. In this context, condition monitoring systems (CMS) should be able to assess engine health, to predict developing failures, i.e. engine state degradation, and to diagnose failure modes at a low price. In this article, two different thermodynamical model-based approaches to detect two common failure modes -increased blow-by and compression ratio failures -of large diesel engines given cylinder pressure traces with low sampling rate are discussed and compared. Special focus is put on estimation robustness and reliability by excluding the combustion phase and signal parts with high noise level. The proposed algorithms are validated with experimental data. Zustands€ u uberwachung und Fehlerdiagnose von Schiffsdieselmotoren. In vielen Industriezweigen ist die zuverl€ a assige Erkennung von Fehlfunktionen im Motor f€ u ur die Vorhersage und Planung von Wartungsintervallen unabk€ o ommlich. Auf Hochseeschiffen, die sich oft mehrere Monate auf offener See befinden, kann ein Ausfall des Motors zu teuren Standzeiten f€ u uhren. Zustandsdiagnosesysteme (ZDS) sollten daher in der Lage sein, kosteng€ u unstig sowohl den Gesundheitszustand des Motors abzusch€ a atzen als auch Abn€ u utzungserscheinungen rechtzeitig erkennen und identifizieren zu k€ o onnen. In diesem Artikel werden zwei verschiedene auf thermodynamischen Modellen basierende Ans€ a atze f€ u ur die Erkennung von zwei h€ a aufig in Großmotoren auftretenden Fehlerursachen -erh€ o ohtes blow-by und Kompressionsverluste -unter Verwendung der Zylinderinnendruckverl€ a aufe mit geringer Abtastrate diskutiert. Besonderes Augenmerk liegt dabei auf der Robustheit und Zuverl€ a assigkeit der Parametersch€ a atzung durch Ausblenden der Verbrennungsphase und von Signalteilen mit hohem Rauschpegel. Die vorgestellten Algorithmen werden anhand von realen Messdaten validiert. Schl€ u usselw€ o orter: Parameteridentifikation; blow-by; Kompressionsfehler; Eliminierung von Offsetgr€ o oßen

Optimized SVM-Driven Multi-Class Approach by Improved ABC to Estimating Ship Systems State

“…Finally, by merging the two types of nectar sources, calculating and sorting their fitness, the best N nectar sources as the initial populations can be selected via (5).…”

“…Set the maximum number of iterations N I =100, the maximum number of searches for each nectar source N R =50, the number of nectar sources N=40, the optimization range of parameters C, σ is 0-100, and generate employed bees and onlookers corresponding to the number of nectar sources. Adopt the unlearning improvement strategy via (3), (4) and (5) to generate 40 initial nectar sources as shown in Fig. 6.…”

“…Chandroth [4] proposed a method combined vibration data with cylinder pressures data for the condition monitoring of the main engine. Accordingly, the thermodynamic model of marine engine has been developed and used to achieve condition monitoring [5], [6]. Accordingly, Li et al [7] developed a self-learning algorithm to perform fault diagnosis in the combustion system of a marine diesel engine.…”

Investigating an SVM-driven, one-class approach to estimating ship systems condition

“…For example, a methodology where vibration data are combined with performance data (cylinder pressures) for the condition monitoring of a main engine has been suggested (Chandroth 2004). Accordingly, thermodynamic model of main and auxiliary engines have been developed and used to perform condition monitoring (Watzenig et al 2009;Lamaris and Hountalas 2010). Accordingly, Hountalas (2000) developed a diesel engine performance model that can account for both normal and faulty conditions.…”

An Enhanced Data Visualization Method for Diesel Engine Malfunction Classification Using Multi-Sensor Signals

“…In recent years, different measurement methods were used for condition monitoring and diagnosis of diesel engines. These measurement methods, including vibrations [ 4 , 5 , 6 , 7 ], instantaneous speed [ 8 , 9 , 10 ], oil analysis [ 11 ], acoustic emission [ 12 , 13 ], cylinder pressure [ 14 , 15 , 16 ], etc. , have shown good performance in condition monitoring of the diesel engine.…”