Wavelet analysis of trees

In this paper the discharge currents accompanied with tree developments are discussed based on the wavelet transformations. Tree phenomena are a very important object for studying solid insulating materials, and one of the physical quantities related with tree developments is the discharge currents. In this paper we apply the wavelet transformations to the discharge currents accompanied with the trees developing from the needle tip of a needle-plane electrode system in PE. The lack of the discharge current (pulses) were estimated by a neural network method. The results obtained are: The development of trees is active at the applied time with great wavelet coefficient /spl alpha/ (I), and thus the development can be estimated from the variation of the corresponding wavelet coefficient /spl alpha/ (I). The wavelet transformations are a very useful method for obtaining the time dependent information of local tree developments.<
>