In order to extract fault features of the rotortostator rubimpact for compressor,a method based on local mean decomposition(LMD) spectrum analysis of elastic supporter strain signals was proposed.In this method,the vibrational strain signals of the rotor system during aeroengine operating was obtained by sticking strain gages on the strips of the elastic supporter in the vertical and horizontal directions,the vibrational strain signals were decomposed by LMD method,and the signals obtained by LMD were analyzed through the analysis of cycle frequency spectrum and slice spectrum to extract the characteristics of rotortostator rubimpact fault.The results demonstrate that LMD can effectively decompose the vibration strain signals of elastic supports.Both the cycle frequency spectrum of the first production function(PF) component and the slice spectrum of each PF component for the elastic supporter vibration strain signals can extract rotortostator rubimpact fault characteristic frequencies.The validity of the proposed method in the identification and diagnosis of the rotortostator rubimpact fault was verified by the disassembly and inspection results of the impeller blades.