Arc fault is one of the reasons which cause the equipment or lines electrical fire. In this paper

a novel diagnosis method of arc fault in series based on mathematical morphology is proposed in order to diagnose arc fault in series in distribution system rapidly. The noise interferences in the current signal can be filtrated effectively using the combination of morphological opening and closing operation

and the abrupt change features of current waveform from the pre and post during the arc fault in series could be exacted via the multi-resolution morphological gradient（MMG） computing. And then

the morphological gradient values of currentduring the normal working and arc fault conditions can be obtained. It can be judged accurately whether arc fault has been occurred or not by comparing the size of current morphology gradient values pre and post of the arc fault. The proposed method of arc fault in series diagnosis has advantages of simple logic operation

small computation amount and easy to realize.