Design of Artificial Neural Networks to Recognize Electrical Load

The human need for electrical energy is currently very large. Especially for the needs inthe field of industry, research and education and household. In these fields, electrical energy can be said to be the main requirement. Any use of electrical energy, required fees paid to the government company. The fee is paid based on the amount of electricity consumed. There is a study built to detect the amount of electricity consumption. Jumrianto, built a 1 phase AC electric energy measuring device using a microcontroller whose results are displayed to the LCD. The calculated energy involves the value of the power factor, so that it can be more accurate in calculating the use of inductive electrical loads [1]. R.Bhavani dan Prashanth B.U.V built a tool for calculating the consumption of electrical energy wirelessly that can be monitored remotely through computers and cellphones [2] [3].In this study, a system is built that not only measures the consumption of electrical energy, but also can recognize the type of load that is active. The system consists of a series of sensors that are processed using an Arduino microcontroller for and using the Artificial Neural Network (ANN) Algorithm with a Backpropagation method built into the computer to identify load usage. ANN is a reliable method for pattern recognition [4] and has been used in previous studies with a success rate of 86% [5]. Expenses are recognized by analyzing the shape of the current pattern of each load taken by several samples. The result of the recognition of the load is displayed to the computer