Implementation of Solar Energy Grid Facilities Towards Smart City Development: A Preliminary Study for Kuala Lumpur City Using the NNARX Method

The key factor in smart city development is implementing clean and renewable energy to keep the environment safe from pollutants. The energy mentioned is precisely on electricity and also transportation services. According to the statistics retrieved from The World Bank Data, the carbon emission released by both electricity generation and transportation for each country in the world is increasing every year. This has shifted the energy generation using the renewable energy source which is solar energy. With the development of solar energy as the main energy source for electricity and transportation, there are also some concerns about its vulnerability and its availability. The continuity of electricity generation and electric vehicles are totally reliant on the solar energy source. A preliminary study must be carried out to establish secure solar energy grid facilities for smart cities. Dealing with the non-linear time-based dataset requires the application of the Machine Learning method. Neural Network is one of the Machine Learning methods which uses the black-box model for predictive analysis. In this paper, the solar energy potential forecast is developed by using the Neural Network Autoregressive Model with Exogenous Input (NNARX) method. NNARX is the evolutionary Multi-layer Neural Network that has been widely applied in non-linear time-based predictive modeling. The input for the solar energy potential forecast using the NNARX method is the historical geographical and meteorological dataset for Kuala Lumpur, the capital city of Malaysia. Based on the modeling simulation results using MATLAB R2019a, the NNARX manage to forecast the solar energy potential output with the regression results of 94.45%. This indicates that the NNARX could be used as a forecast model for the preliminary study.