Based on the characteristics and performance of the predictive maintenance system of turbine generator and the requirement of data,the collection and analysis of algorithm of data is researched and designed.

With the predictive maintenance (PM) being popularized in enterprise, the fault prediction technique is getting more and more attention.

Combining with the research in fault diagnosis and predictive maintenance system for the mechanical and electrical equipment in a wanship, this paper presents the constitution of predictive maintenance data base, the determination of alarm parameters and in practical application in naval vessels.

The paper examines the development existing circumstance and future about condition-based maintenance internally compared with externally .

In addition to prospective maintenance and detectivemaintenance, it describes reliability-centerial maintenance and cost-effective maintenance in detail, the latter is based onquantified analyses.

The thesis includes the transportation process of corrosive substances in the concrete, rebar corrosion model, structural time-varied reliability analysis and repair demand forecast mode.