Optimization for discrete balance weights based on hybrid genetic algorithm

The masses and phases of balance weights for rotor dynamic balance are discrete generally. Considering the fact, the optimization problem of balance weights with the constraints that the masses and phases of balance weights are discrete is discussed for the first time. With the constraint condition, the disadvantages of former optimization methods are got over, which depress the balance effect because the balance weights added actually are not equal to those optimized in theory. A hybrid genetic algorithm with a neighborhood search operator to solve the optimization problem is discussed to strengthen the local extremum search ability, and the defect easy to miss the right global optimum because of the attraction of local optimum is overcome. An example shows that the optimization method can reduce the test times and improve the balance efficiency.