Abstract
Vibration predictions for rotating machinery with high-speed flexible rotors must account for the methods and limitations of the balance test process which determine the residual rotor unbalance. Vibration predictions based on finite element analysis (FEA) methods are highly dependent upon the assumed rotor unbalance amplitude and phase. The actual residual unbalance distribution depends upon the measured influence coefficients and the least-mean-square (LMS) algorithm used to calculate balance correction weights. Repeatability of the vibration measurements is a key factor in successful balancing. The vibration predictions described in this paper use estimates of final residual unbalance obtained by simulating the balance test process. The simulation uses FEA based influence coefficients, a test based measurement uncertainty (repeatability) model, and LMS balance weight calculations including the specified vibration target levels. The simulations can be used to predict the limit of balance performance of the machinery and to evaluate design options for impact on residual unbalance levels.
The Ultra High Speed LMS Algorithm Implemented on Parallel Architecture Suitable for Multidimensional Adaptive Filtering
