Research On Method And System Of Fault Diagnosis Of Ship's Turbocharger

As one of the critical system of marine diesel engine, turbocharger plays an important role in ensuring the power of ship's sailing provided sustainably and stably by diesel engine and reducing the impact of emissions on the environment. The turbocharger system of ships is complicated and usually works under bad conditions, making it easier to failure. In order to avoid terrible accident, fault diagnosis technologies are applied in tracing its working condition and making health assessment in real time. However, fault diagnosis of ship's turbocharger system has not yet attracted enough attention, in most cases, measurements are taken only after failure. Less advanced technologies have been in practical use, making its low efficiency. When analyzing the cause of fault and the phenomena and nonlinear parameters and correspondence between the operating characteristics and the type of fault, traditional methods and technologies appear powerless. New model of fault diagnosis which is about to be used in practice by adopting new methods based on current methods and technologies has to be established.Thus, a new kind of fault diagnosis model is put forward in this paper. Which adopted artificial neural networks and expert system. In connection with practical requirements of a turbocharger factory in Chongqing, a management system integrating fault diagnosis and operating and maintenance after sales of ship's turbocharger products is designed and developed in this paper, improving fault diagnosis' s efficiency of ship's turbocharger products and service quality of the company.First, structures and working principles of ship's turbocharger were introduced, especially the main types of fault, including their phenomenon, characteristic, causes and advice of maintenance. Based on that, characteristics of faults of ship's turbocharger product was summarized. Framework of the management system who integrated fault diagnosis and operating and maintenance after sale of turbocharger product and the mode of fault diagnosis and operating and maintenance of ship's turbocharger product are discussed in the next.Then, expert system based on fault tree of turbocharger's fault diagnosis was studied. The fault diagnosis' s model based on artificial neural networks was studied also. BP artificial neural network and BP-GA network of fault diagnosis of ship's turbocharger product were introduced respectively, including their structure, process of design, and the application situation.At last, the database of turbocharger's faults of ships was established and system was designed and developed. Functions such as data management of product's fault diagnosis, fault prediction and diagnosis of turbocharger, operating and maintenance after sale of product management were realized. Expert based on fault tree and artificial neural networks of diagnosis introduced were integrated in the management system. And a good effect was achieved thanks to the system.