Energy Audit in Households using Machine Learning

Abstract

Maintaining the energy usage with minimal power loss throughout the supply chain is of the major issues faced in many small-scale sectors or even in households of today’s world. Even though Power transmission can play a cardinal role in the supply chain, monitoring the transmission lines for energy leakage or any faulty connections is critically important. There have been several measures taken to come up with a better solution but, the problem of finding a consistent method for monitoring the power leakage is still at peril. There are actually many ways of saving the energy by mitigating the usage and preventing the loss of energy due to over usage and wastages, for this a thorough monitoring and study of the usage should be done. If the electricity usage pattern of the concerned is identified, then it will be facile to come up with a solution for the problem at hand. The electricity wastage constituted by all the countries aggregated is found out to be around 8.25%, which is considerately large given that many places around the world does not even have access to electricity. So, there is a need to find a better solution for this problem. After conducting a thorough study on the electricity usage pattern of several households we are proposing a method which is an ensemble of machine learning algorithms, Internet of Things, sensors, Embedded systems. Using an IoT device we’ve designed we monitoring and collecting electricity usage in households in a time based manner. These collected data is stored in the database and is processed and fed into machine learning algorithm to predict the upcoming month’s electricity usage. This predicted data is then fed into another algorithm to provide recommendations to the user to reduce the electricity consumption according to their usage interests. Thus reducing the cost significantly.