Detecting Electric Power Theft in Households Using Hidden Markov Models

Electric power theft is a severe problem affecting the energy sector of a nation and continues to propagate and the energy sector tends to lose a chunk of its revenue resulting in higher tariffs for consumers innocent and guilty alike. Since electrical energy is widely used across the world and its demand is increasing rapidly, generation of electric power has to be increased or the already available one, economised. An effective system for detecting electric power theft is therefore needed so that when implemented, will either provide the nation with adequate funds to increase generation when households pay for energy that is consumed, or effectively utilise what is already available when households economise their energy consumption. This paper provides a means whereby power consumption of consumers can be monitored and recorded over a period of time. Based on that, the Hidden Markov Model (HMM) was used to predict the consumption with the deviations between real and predicted values giving indications of theft or otherwise. Appropriate sanctions can then be meted out to perpetrators of electric power theft. This way, higher tariffs are not imposed on innocent consumers and consumers will pay for the right amount of energy consumed.

Keywords:

Electricity tariffs , Households , Hidden Markov Model (HMM) , Power consumers , Power theft detection