Abstract:For large steam turbine generator units with high parameters, high efficiency and complex shaft system, the start-stop cost is too high.Improving the precision and efficiency of shaft system dynamic balance becomes the important factor to reduce the cost of the shaft system dynamic balancing. The mechanism and vibration characteristics of the rotor unbalance fault were introduced, the identification method of the shaft system unbalance based on vibration signal processing and deep learning was sorted out, and several calculation methods of the shaft system dynamic unbalance were listed, the dynamic balance management framework of the on-site shaft system was established. This framework integrates rotor unbalance identification, dynamic balance correction and influence coefficient database of the shaft system dynamic balance to improve the accuracy of on-site dynamic unbalance identification. The analysis shows that the efficient dynamic balance can effectively reduce the vibration of shafting, and has been successfully applied to prevent a variety of fault prevention like the thermal bending of rotor and rubbing to improve the safety and reliability of the steam turbine generator units. At the same time, if the weight is large, the high density balance block should be used, which can reduce the dispersion of the unbalanced block and further improve the dynamic balance effect. Meanwhile if the weight is larger, the high-density balance weight should be used, which can reduce the dispersion of the unbalanced weight and improve the effect of the dynamic balance.