HOME-ELECTRICITY-USAGE-PREDICTION-AND-ANOMALY-DETECTION

As a developing country, Srilanka has faced many difficulties in the power management field. The main reason is still Srilanka is using the old traditional power grid system. As a result of these old mechanisms, nobody is aware of the future energy consumption rates, whether the current power generation points are enough to fulfill those needs. When compared to the USA, they are converting their power grid system to smart grid systems because of as a country, they experience that the traditional one-way communication grid is not enough to prevent blackouts. Even though Sri Lanka has experienced a few blackouts in recent history, somehow, Sri Lanka managed the conditions by monitoring the grid using a load dispatch center to balance the supply-demand. Though this method helps to survive for tiny lands, this doesn’t help full for largest countries. But Sri Lanka also will not survive lengthy blackouts in the future. Because Still major thermal power plants repeatedly get inactive due to mechanical issues in the turbine and equipment part. On the other hand, most significant power plants work using Coal and diesel fuel to power up plants. But these natural resources are decreasing day by day from the earth very quickly. On the other hand, these power plants will definitely be banned due to environmental issues like global warming in the near future. The most optimal solution to overcome this problem is to predict our future energy demand and give AI power to machines to control themselves according to the Current grid statistics and add various backup capabilities (Solar panels, wind turbines) to the system. As the first step, I am going to predict the household energy consumption of the house. For that, I will use the Individual household dataset as support data in this project. This data set location is France, and it contains nine attributes .they are date-time, global active power, global reactive power, voltage, global intensity, submeter 1, submeter 2, and sub-meter 3 values. But this data set didn’t contain the local weather data stats for each consumption. Therefore, I will use my previous computer engineering project smart meter system database value. It is stored in a cloud database. Both data sets attributes are the same, and my data set will be more suitable because the voltages and other factors are different from one country to another.