Fractal Techniques In Arcing Fault Analysis And Online Power Cable Diagnostic Method

Fractal geometry is one of the most important discoveries of mathematicalscience. Nowadays fractal techniques are widely used in natural and social science.This paper discusses some applications of fractal in electric power system, which isincluding three aspects.Firstly, study of arcing fault with fractal tools. The sampling data of arcing faultto earth of transmission line is analyzed with R/S analysis. The characteristics ofpower spectrum and fractal dimension of fault signals are discussed. To a certaindegree, the spectrum of the sampling data, which is sampled at monitored point whena fault occurs at a point of a transmission line, embodies an internal description of thefault wave to the fault position. The fractal dimension of the fault wave is aquantitative representation of the sampling signal to the fault location. The fractaldimension of sampling data shows the spectrum of this sampling signal wave, withwhich an acing fault location can be estimated. The functional relationship betweenfault location and fractal dimension of a fault signal, of which the basic component isfiltered, is a quadratic curve. The R/S method and spectrum analysis of arcing faultsignal is a good candidate for protection algorithm and fault location estimation oftransmission line.Secondly, study of hot-line diagnostic method for XLPE power cables. In thispart, an accurate water tree model has been created .We successfully develop a newhot-line diagnostic method called "high frequency superposition method". Wesuperpose a voltage with high frequency but low amplitude on the shielding layer ofpower cables to make the loss current more obvious. Then we use fractal methods toanalyze loss current signals collected. Degradation of XLPE cable by water treesgives rise to harmonics in the loss current. Harmonic components arise as a result ofnonlinear voltage-current characteristic of water trees. The fractal dimension of losscurrent waveform increases because of the growth of harmonics. We calculate threefractal dimensions including capacity, information and correlation. We discuss theinterrelationship between fractal dimension and harmonics. The fractal dimensionprovides an extremely good indication of the state of degradation, which can be usedas a judgment for hot-line dielectric degradation.Thirdly, study of power cable dielectrics breakdown and online diagnosticmethod. The forky directions of water tree in XLPE dielectrics have been established.It is a fractal structure, which is attribute to DLA (Diffusion-Limited Aggregation)model. We develop an online diagnostic method called "Static Transient detection". Itmakes use of the inherent and frequent transients of power systems and measures theloss current waveform. As a result, we can estimate the dielectric state based on thefractal dimension of loss current.