| An effective methodology for dynamically analyzing and diagnostically monitoring spindle-bearing structures is necessary for the fast development of automated manufacturing processes. In this work, attempts have been made to improve high-speed spindle performance based on both the original design and the operational signature, which is comprised of three parts.;Second, a new time domain strategy has been proposed and tested for diagnosing and monitoring rotating machinery using random vibration signatures. This strategy includes two steps: (1) monitoring and (2) diagnosis. In the first step, the concept of forming the indices from prediction error analysis is explored. The second step starts only when abnormal conditions occur in the first step. For this case, a classification of prospective faults is carried out according to the "nearest neighbor rule". A new discriminating scheme is developed based on the cross-entropy between two individual time series representing different conditions. A practical algorithm for calculating the discriminant functions via ARMA modeling is given.;Finally, the strategies and algorithms have been experimentally verified using an experimental spindle-bearing structure. In dealing with some practical problems, attempts are also made to cancel the noise during measurement through optimal filtering. In addition, for the simulation study, an algorithm for calculating the response of a structure is formulated in the state space domain.;First, a new mathematical model for a rotating spindle-bearing structure has been established. This model employs Timoshenko beam theory, with the gyroscopic moment due to rotating motion taken into account. A Finite Element Method (FEM) formulation is carried out to seek the numerical solution. However, FEM alone can not adequately obtain an accurate result unless the unknown bearing parameters have been identified. Therefore, the Dynamic Data System (DDS) methodology is used, together with a FEM model after condensation, to estimate the bearing parameters. Then, the dynamic behavior of the spindle-bearing structure is investigated theoretically based on this model and experimentally via DDS modal analysis. Some dynamic phenomena of the structure are explored, which are useful in high speed spindle design. |
