Predict Early Warning Protection, Arc Fault Arc Sound

As a frequent disaster in medium-low switchgear equipment, the fault electric arc makesgreat danger to the safety of personnel and equipment. Once happened, it would release highenergy through sound, light, heat, electricity and so on. It would burn cables, switchgearapparatus and transformers. Moreover, it would arouse explosion, threat people's life, andcause a great economical and social problem. It is very important to research early arcing faultdetecting methods to avoid great loss and occurrence probability.In this paper, actuality of research & apply about fault electric arc in internal & overseasand various detection methods are introduced. It is pointed out that early detecting of arcingfaults and intelligent detection algorithm is the trend of future study in this field. A specificexperiment system is set up, which is to study the change of arcing sound, light, current ofmedium-low voltage switchgear under different conditions of discharge test. Experimentationis provided that it is feasible to make arcing sound as a state parameter of arcing fault.A nonlinear analysis of method based on chaos and fractal theory is used to extract thefeature of arcing sound. The paper built a system based on duffing-homels chaotic oscillatormodel to detect the Characteristic frequency of arcing sound. The paper put forwardsound's correlation dimension could be the criterion of the existence of arc fault, then builtup an on-line system to detect arcing sound by chaos and fractal theory which improvestraditional passive fault electric arc detection to early active detection. As a two-criteriaapproach, arc lighting and current of the switchgear are used to detect the fault electric arc.When the features of fault electric arc are detected, the system would send signal to break offthe relay which controlled the circuit breaker, and protect the Transformer and switchgear fromfurther damage.