Application of nonlinear programming for balancing rotor systems

A systematic procedure for balancing flexible rotor dynamic systems using nonlinear programming techniques is presented. The procedure does not require the use of trial runs, but a valid mathematical model of the system dynamics is required. The rotor response to balance corrections obtained from an optimal search process is then simulated using this model. The objective is to determine the particular correction distribution which will minimize the difference between simulated and experimental results. The effort and expense of trial runs are eliminated, however, this is replaced by the time and expense of generating a valid mathematical model. An example is presented to explore the feasibility and potential of this semi-automated balancing procedure. The results are encouraging and the procedure appears to be a viable alternative to conventional balancing procedures for applications where accurate rotordynamic mathematical models are available.