Study on Mistuning Identification of Vehicle Turbocharger Turbine BLISK

Abstract

Blade vibration is one of the most critical items to be solved in turbocharger design. Turbocharger has much broader operating range than other turbo machinery and some unavoidable resonance may remain in operating range. In such a critical situation, predicting resonant vibration level is desirable for evaluating criticality. Additionally, knowing what amount of margin should be taken for the predicted nominal vibratory stress (known as “Magnification Factor”) is also important for the structural design. Blade to blade vibration response have significant scatter in real rotor. This kind of phenomena is called as “mistuned bladed disk effect” and discussed by many researchers in aero-engine field. Some research studies have appeared also in the automotive turbocharger field but its feature is not well understood especially on the item of mistuning identification and the cause of mistuning. Considering above situation, studies on mistuned vibration were performed for turbine BLISK of automotive turbocharger. Three type of mistuning (frequency mistuning, geometric mistuning, and mistuning of macroscopic property of directionally solidified material) were assumed and these effect on above items (mistuning identification and the cause of mistuning) were investigated. This paper consists of three parts. At first, frequency mistune model and geometry mistune models were prepared. To build a frequency mistuning model, FMMID proposed by Feiner and Griffin [6] is applied. After the basic function test by virtual BLISK model with known mistuning, FMMID was applied to the actual BLISK. Natural frequency mistuning of each blade was identified by FMMID based on the modal measurement result. Obtained frequency mistuning is reflected to the (geometry tuned) FE model by changing the Young’s modulus of each blade in corresponding rate. The 3D measurement was also performed to the same BLISK and dimensional information from this measurement was reflected to the geometrically mistuned (material property tuned) FE model. In the next step, vibration analysis (eigenvalue and frequency response) was performed and these results were compared to the measurement result. Vibration measurement in operating condition was performed at the resonant point of mode3 and nozzle count excitation frequency by utilizing the NSMS (Non-intrusive stress measurement system). Analytical result of the frequency mistuning model shows a good agreement with the experimental, while the analytical results of the geometrically mistuned model did not match to the experimental result. At the last part, cause of discrepancy between the analytical result of the geometry mistuning model and the measured result was investigated from the view point of the effect of the anisotropy of elastic constants on the vibration characteristics of the DS (Directionally Solidified) blades.