Some observations on the CLNA model in fretting fatigue