A Prediction of Power Demand using Weather Forecasting and Machine Learning: A Case of a Clinic in Japan

There is a need to accelerate research and development for an energy saving and a demand response using a renewable energy system because the supply and demand of a power is tight worldwide since 2021. Since hospitals and clinics are heavy power demand, there are promote energy conservation by using distributed energy that actively utilizes renewable energy. Recently, clinics installing a photovoltaic (PV) in combination with a diesel generator (DG) for an energy saving and/or a peek cut of a demand have been more. So, it's necessary for medical staff to predict a power demand, assuming an energy saving and/or a short-term operation. Therefore, this paper proposes the prediction method of a power demand for medical facilities using the weather forecasting data. A neural network (NN) of the prediction method is used the weather forecasting data in the area of medical facilities announced by Japan Meteorological Agency (JMA) are gathered for a long-term as an input. As a result, it is shown that power demand can be predicted with high accuracy.