Automated Monitoring of Electricity Consumption Using LSTM-RNN and IoT

Today we are living in a world where technology is dominating every sector. The need of automation is increasing day by day and the way of working of the whole world is moving towards the automation of different tasks, which can be done with expert knowledge without any need for human efforts. Due to this, the electricity demand is increasing, but this includes a lot of wastage of electricity that can be saved. The problem which we have identified here is wastage of electricity and to solve this problem we simply need a system which can be used for monitoring the usage of electricity. At first place, this problem looks very simple, and it seems it can be solved easily by some manual work done by a human but, this problem is very complex in reality as the consumer is not able to identify the exact point where electricity is being wasted or else it will be identified once electricity is already wasted which is of no use. These traditional systems are not efficient enough as they cannot identify a potential electricity wastage in advance, for example, if we charge mobile and we forget to turn it off then the charger will consume electricity for several hours and the wastage of electricity will be identified when we turn off charging. To solve these problems many models have been proposed by so many researchers that are BP Neural Network model, EPSO-BP neural network model and there are many more models that were used to solve this problem. The working and drawbacks of previously proposed models will be discussed further in the related work section of this paper. To solve this problem in this paper we have proposed a model that includes 3 sections. In the first section, we have created an IoT based device to measure and store the electricity usage of each appliance. In the second section, we have used the LSTM version of RNN which is very accurate and efficient to create a model that can work in real-time with very high accuracy. In the last section, this paper includes a web app as the frontend of this whole work done in previous sections.