Context-Aware Query for High-Voltage Transmission Line Fault Detection using Wireless Sensor Network Paulo R´ egis C. de Ara´ ujo Department of Telematics Instituto Federal do Cear´ a Fortaleza, Cear´ a, Brasil Email: pregis@ifce.edu.br Raimir Holanda Department of Computer Science (PPGIA) University of Fortaleza Fortaleza, Cear´ a, Brasil Email:raimir@unifor.br Antonio Wendell de Oliveira Rodrigues and Andr´ e Luiz Carneiro de Ara´ ujo and Jo˜ ao Paolo C. M. Oliveira Department of Telematics Instituto Federal do Cear´ a Fortaleza, Cear´ a, Brasil Telephone: 55 085 33073607 Jos´ e de Aguiar Moraes Filho and Angelo Brayner Department of Computer Science (PPGIA) University of Fortaleza Fortaleza, Cear´ a, Brasil Telephone: 55 085 3477.3268 Abstract—Due to vandalism or weather conditions, disruption or damage in high voltage transmission line insulators may occur, causing a risk to the electricity supply for many cities. Studies show that the insulators ′ states can be known by the value of the leakage current which passes through them and by other variables. In this paper, we present a context-aware query strategy in a Wireless Sensor Network (WSN), installed on the electricity supply system, in order to detect faults in insulators on the high voltage transmission line. The approach presented is based on the strategy of continuous and real-time query of some variables. We have created two new modules, the Data Prediction Module (DatPredMod) and the Data Aggregation Module (DatAggrMod) to support this query strategy. I. INTRODUCTION Due to vandalism or weather conditions, disruption or damage in high voltage transmission line insulators may occur, causing a risk to the electricity supply for many cities. The damage to the insulator generates a high leakage current which is sufficient to open switchgear in the electricity sub- station. Such a situation interrupts the power supply. One of the ways to investigate the health (state) of insulators is to know the leakage current which passes through them. Besides the physical and structural damage to the insulator, the leakage current can vary with temperature, humidity and the environmental pollution. Other important characteristic of the leakage current is that when it passes through the insulator, it generates a radio frequency signal. As the leakage current varies with environmental conditions, it requires an analysis of the environment context to know whether the increase in leakage current is due to the insulator state or if there are environmental conditions causing the increase. As we have used the WSN to monitor the insulators, we have to observe the most critical point in WSNs, which is en- ergy consumption [5]. The largest part of energy consumption in a sensor node occurs during data transmission or reception. For this reason, the main goal of most algorithms designed for WSN applications is communication cost reduction in terms of energy consumption [2], [3], [4]. For example, the use of in-network aggregation operators [2], [3], [4] is an efficient strategy to reduce the volume of data transmitted in a given WSN, and consequently, the energy consumption. Even the WSN installed on an electricity transmission line, the possibility of harnessing this energy is by magnetic induction which presents a very poor energy conversion. Thus, our work addresses two issues. The concept of data aggregation and data compression, whose coordination is realized by DatAggrMod (embedded in the cluster head and base station), is used due to the large amount of information that should be passed to the base station. The other issue is related to the concept of data prediction, whose coordination is realized by DatPredMod (embedded in the base station). This concept is also important to minimize the cost of packet traffic in the WSN and to improve the quality of information provided. This work is structured in five sections. In section II, the basic concepts of this work are depicted. Section III demon- strates all the components of the approach and its functionality. In section IV, the evaluation stage is presented, with graphics that demonstrate the efficiency of our approach. And, finally, section V presents the conclusion and future works. September 11, 2014