Numerical And Experimental Study On Dynamic Characteristics Of Blades Of An Axial Compressor

Axial compressors play an important role in petrochemical, metallurgical and other industrial sectors and raises higher requirements than before on the design and manufacturing technology of the machines. The safety and reliability of blades as the core component of axial compressor are particularly important. The evaluation of the natural frequency of is an indispensable step in the design process, which means the identification of natural frequency method must be highly accurate.This thesis focuses on the first and second stage blades of the axial compressor Nitric Acid4-in-1. Two ways of identifying natural frequencies of the blade are carried out:the finite element simulation technology and testing techniques of dynamic characteristics test. Comparison of the results by these two methods provides accuracy of identification of natural frequency of the blades.First, the process of three-dimensional solid modeling of the blade with software SolidWorks is introduced, and the model of the three-dimensional twisted structure of leaf blade is discussed in detail. Second, the process of converting the three-dimensional solid modeling into a finite element model is shown. Based on the accurate three-dimensional finite element model, the first ten natural frequencies and mode shapes are computed using the Block Lanczos approach with the ANSYS Modal analysis. Identification of the mode shapes shows that the first vibration mode is the first-order bending, the second mode is the first-order coupling, the third mode is the first-order torsion, the fourth mode is the second-order bending, the fifth mode is the third-order bending, and the sixth to tenth modes are the second-to the tenth-order couplings.As for the technology of dynamic characteristics tests, a brief introduction about the rig and hardware for the dynamic characteristics test are provided. The procedure of blade dynamic characteristics test is elaborated. Finally, the file of frequency response function of each measuring point is imported into the modal analysis software. The mode shapes at the excitation point are normalized by the peak pickup method. The measured values and mode shapes of the first ten natural frequencies are obtained. Subsequently, the natural frequencies and vibration modes obtained by the simulation and the test are compared and found in good agreement.