The vibration behaviour of a rotor system supported by hydrodynamic bearing is significantly influenced by the dynamic characteristics of the bearings. An identification procedure for stimating bearings oil-film coefficients is developed in this work. Eight linearized oil-film coefficients are identified by fitting the measured complex functions of frequency response to those calculated theoretically. The theoretical model and reliability are examined by computer simulation. From error analysis, some interesting conclusions are reached and a new identification method is suggested in order to reduce the influence of phase measurement errors on estimated results.

The vibration behaviour of rotors with journal bearings is significantly influenced by the dynamic characteristics of the bearing oil-film. To validate analytical modes of oil-film dynamics, numerous experimental investigations have been undertaken. The authors have previously carried out a project to improve the quality of the estimation of oil-film coefficients. In the companion paper (part I), a system model and an algorithm with simulation results have been presented for a flexible rotor-bearing system under operating conditions. This second paper gives details of an experimental investigation designed to demonstrate the feasibility of the method.

在回顾传统的设计方法及状态监测的传统任务的基础上，展望现代设计技术与状态监测诊断内涵及共相互联系，提出状态监测的研究方向。

利用信息理论建立了描述多传感器之间关联特性的数学模型，并用相关和相干函数的概率分布导出度量两传感器在时域和频域关联程度的计算公式，从理论上研究了多传感器间关联特性的定量分析和计算问题。

A new strategy for the techniques of monitoring and diagnosing hydrodynamic bearing-rotor systems is recommended in this paper. The strategy integrates the traditional technique of spectrum and statistics analysis(backward inference) with the application of knowledge and information of rotor-bearing dynamics and tribology (forward inference).This strategy has been used to develop a monitoring and diagnosing system for rotating machinery and examples of the successfuI application of the system are given.