Bhatti, S. A., Shan, Q., Glover, Ian, Atkinson, R. and Rutherford, R. (2009) Extraction of Middleton Noise Model Parameters from Measurement of the Noise Environment in an Electricity Substation. In: IET-URSI UK Festival of Radio Science, December 2009, Birmingham. (Unpublished)
Abstract

An impulsive noise model of the environment inside an electricity transmission substation
will be presented. The results of a measurement campaign will be summarised and the
existing Middleton Class A and Class B models [1] will be reviewed. A statistical modelling
process for parameter extraction from the field measurement database will be described.
Parameter extraction is based on the Wave Packet Transform [2] and expectation
maximization (EM) algorithm [3] and the resulting statistics of first order impulse features
(amplitude, duration and inter-arrival time) will be presented.
The most obvious source of impulsive noise close to power systems equipment is partial
discharge (PD) caused by degraded insulation. There is evidence in the measurement
database, however, of an unexpected quasi-periodic impulsive process. It is the aggregate
process that is of interest here, irrespective of physical origin. If the duration of PD, or other,
impulses is sufficiently small [4] their spectrum will extend into the microwave region where
they may degrade the performance of WLANs, WPANs and similar technologies operating in
the 2.4 GHz ISM band. There is evidence in the measured data that this is indeed the case.
The degradation appears to be greater than expected for additive white Gaussian noise in the
high SNR regime. In the low SNR regime the converse seems to be the case. The primary
purpose of the noise model is the assessment, by simulation, of these technologies for control,
monitoring and surveillance applications in electricity substations. Some example results from
these simulations showing the degradation of bit-error-rate due to impulsive noise relative to
that due to additive white Gaussian noise will be presented.

Information
Library
Statistics
Add to AnyAdd to TwitterAdd to FacebookAdd to LinkedinAdd to PinterestAdd to Email