Deconvolution method applied to the study of surface
current density in a low voltage breaker arc using magnetic measures

MACHKOUR-DESHAYES Nadia

Manufacturing Metrology Division (822).

National Institute of Standards and Technology, Gaithersburg, Maryland 20899

An electric arc is an unstable and transitory phenomenon, but circuit breaker manufacturers need to know the exact nature of the arc. Preferably the arc current density and arc dimension are known; because the arc nature - diffuse or concentrated strongly influences the erosion of the contacts, they must know the nature. At the LEM Laboratory (Laboratoir d’Electrotechnique de Montluçon), a number of experimental methods [1-2] are being developed for reconstituting the shape and for knowing the dynamics of a low voltage breaker electric arc. They all assume the arc is a succession of rectilinear segments.

Starting with measurements of the magnetic induction created by an arcing device, our objective study was to develop ways to reconstitute the current density. The electric arc is modeled as a rectangular current sheet, having a variable density according to the arc propagation direction. The other elements of the arcing breaker are assumed to be threadlike current segments. The measurements used Hall effect sensors, which were placed close to but outside the arc.

We found the magnetic induction created by the arc represents a convolution product between current density, a function depending on the experimental device geometry and the model used for the arc. Further small measurement errors lead to large errors in the solution. Lastly, we had to use optimization to regularize the problem. To choose the best method, several optimization methods were tested.

Our method also allows calculation of the arc size, and determination of some arc parameters. One is the electric conductivity, which is used to estimate the arc temperature.

Presenting Author's information

Name : MACHKOUR-DESHAYES Nadia

Division : Manufacturing Metrology Division (828)

Room and Building address: Sound (233), Room A151

Mail Stop : 8223

Telephone : (301) 975-6627

email : nadia@nist.gov

Sigma Xi member? : Non