Article Preview
TopIntroduction
Most of the worldwide energy is derived from fossil fuels. In 2018, the world's electricity consumption amounted to approximately 23.4 trillion kilowatt-hours (Sönnichsen, 2020). Various countries employ different methods of electricity generation such as steam, nuclear, biomass, geothermal, coal, etc. Along with the United States, China is one of the highest per capita consumers of electricity in the world. In 2019, the China consumed 5,564 billion kWh of electricity and US consumed 3,902 billion kWh of electricity (Sönnichsen, 2020). With the exponential increase in electricity consumption, countries have implemented new ways to decrease emissions stemming from power generation through the use of fossil fuels. Carbon dioxide emissions from fossil fuel sources grew by 2.7% in 2018 which was much faster compared to the growth of 1.6% in 2017 (Levin, 2018). In 2019 the growth somewhat slowed to 0.6% for the first 6-10 months (Friedlingstein et. al., 2019). This decreased by -7% in 2020 due to the pandemic (Friedlingstein et. al., 2020). The dip though is seen a temporary, and steps are needed to reduce the carbon emissions. Renewable energy sources like solar energy will make a significant contribution to this effort.
Figure 1.
US Photo Voltaic Installation (Wood Mackenzie, 2021)
In recent years, solar energy has been increasing in popularity. Figure 1 shows the photo voltaic installations in the use both in terms of history and future projections (Wood Mackenzie, 2021). The residential share of this market can be seen as steadily increasing. Power generation from solar increased 22% in 2019 and this constituted 3% of the total electricity generation in 2019 (IEA, 2020). The demand for solar systems has increased due to added interest from homeowners and businesses. Also, the likely pairing with battery storage to extend electricity availability throughout the off-hours has increased demand for the solar systems. By 2025, more than 25% of all behind-the-meter solar systems will be paired with battery storage, compared to under 5% in 2019 (SEIA, n.d.). Strong federal policies and incentives, such as Solar Investment Tax Credit, have increased the likelihood that consumers will become interested in solar power. Increased solar energy usage offers a myriad of benefits including environmental safety, decreased pollution, and financial benefits, such as decreased utility bills (Stevović, 2017). Solar power has become more affordable, accessible, and prevalent in many parts of the world. Solar panels are usually integrated with smart grids. The main goal of smart grids is to substantially increase the adaptability of renewable energy. Due to off-time challenges and dependency on nature such rainy or cloudy days, it becomes challenging to run renewable energy integrated smart grids efficiently in the absence of any predictive analysis. The problem with substantial renewable integration is that the electricity generated from renewables is not easily predictable and will vary based on weather conditions and site-specific conditions (Jolliffe, 2017). This is where predictive analysis can add significant value to the operation of a smart grid and allow for a steady and balanced supply of electrical power. The predictive analysis carried out in this manuscript can be employed at consumer locations which are not covered by projects such as the Google Sunroof Project (google.com/get/sunroof). Even for those locations that are covered by such projects the granularity of prediction at consumer sites can provide some additional value. The analysis presented in this manuscript may be used to provide a more accurate prediction of solar radiation using feature set from a consumer grade weather station in conjunction with day length forecast from an Internet weather website of choice. The feature set used does not require measurements using complex equipment as usually done in several studies discussed later in this manuscript.