Radio location of partial discharge sources: a support vector regression approach

Abstract

Partial discharge (PD) can provide a useful forewarning of asset failure in electricity substations. A significant proportion of assets are susceptible to PD due to incipient weakness in their dielectrics. This paper examines a low cost approach for uninterrupted monitoring of PD using a network of inexpensive radio sensors to sample the spatial patterns of PD received signal strength. Machine learning techniques are proposed for localisation of PD sources. Specifically, two models based on Support Vector Machines (SVMs) are developed: Support Vector Regression (SVR) and Least-Squares Support Vector Regression (LSSVR). These models construct an explicit regression surface in a high dimensional feature space for function estimation. Their performance is compared to that of artificial neural network (ANN) models. The results show that both SVR and LSSVR methods are superior to ANNs in accuracy. LSSVR approach is particularly recommended as practical alternative for PD source localisation due to it low complexity.

Partial Discharge Detection and Localization: Using Software-Defined Radio

“…Several localization algorithms can be adopted. Examples of algorithms are: time of arrival (TOA), angle of arrival (AOA), time difference of arrival (TDOA), received signal strength (RSS), etc., [20][21][22][23][24]. Recently the RSS algorithm approach has been preferred for indoors and outdoors localization due to cost-effectiveness because it does not require the employment of antenna arrays or synchronization, thus hardware cost is minimized, [23][24].…”

Improving RF-Based Partial Discharge Localization via Machine Learning Ensemble Method

“…In this approach, the network autonomously creates a spatial map of the RF characteristics of the radio signals and uses sophisticated machine learning techniques to estimate the location of PD sources. In previous work we have used machine learning algorithms to build a bespoke propagation model for the radio environment from the perspective of each node [13]. Based on the relative received signal strength of a PD pulse at each node, a multilateration approach can be deployed to infer PD location.…”

Received Signal Strength Intensity Based Localization of Partial Discharge in High Voltage Systems

“…If PD events occur in the long term they can result in failure of high voltage systems. Developing weaknesses in the materials used as dielectric is one of the major causes of PD in high voltage systems [1]. There are three main types of PD, and these are categorised as surface, void and corona discharge [2].…”