An Explanatory Study Approach,Using Machine Learning to Forecast Solar Energy Outcome

AI(artificial intelligence)techniques play a crucially important role in predicting the expected energy outcome and its performance,analysis,modeling and control of renewable energy.Solar energy usage has grown exponentially over the years.In the face of global energy consumption and increased depletion of most fossil fuel,the world is faced with the challenges of meeting the ever-increasing energy demands,also utility companies who provide solar energy have a challenge of unstable input of solar energy to the grid due to its intermittent nature,unlike other sources,hence the difference between expected generation and actual generation,demand and supply can lead to an unbalanced grid.Therefore,incorporating accurately machine learning technology to predict the expected outcome of solar energy from the intermittent solar radiation will be crucial to keep a balance grid operation between supply and demand,production planning and energy management especially during installations of a photovoltaic power plant.However,one of the major problems of forecasting is the algorithms used to control,model,and predict performances of the energy systems which are complicated and involve large computer power,differential equations,and time series.Also having unreliable data(poor quality)for solar radiation over a geographical location as well as insufficient long series can be a bottleneck to actualization.To overcome these problems,we employ the Anaconda Navigator(Jupyter Notebook)for machine learning which can combine large amounts of data with fast,iterative processing and intelligent algorithms allowing the software to learn automatically from patterns or features to predict the performance and outcome of Solar Energy which in turn enables the balance between supply and demand on loads,efficient operation of the utility company as well as enhances power production planning and management.

Effects of Rainfall and Underlying Surface on Flood Recession——The Upper Huaihe River Basin Case

The effects of rainfall and underlying surface conditions on flood recession processes are a critical issue for flood risk reduction and water use in a region.In this article,we examined and clarified the issue in the upper Huaihe River Basin where flood disasters frequently occur.Data on 58 rainstorms and flooding events at eight watersheds during 2006–2015 were collected.An exponential equation(with a key flood recession coefficient)was used to fit the flood recession processes,and their correlations with six potential causal factors—decrease rate of rainfall intensity,distance from the storm center to the outlet of the basin,basin area,basin shape coefficient,basin average slope,and basin relief amplitude—were analyzed by the Spearman correlation test and the Kendall tau test.Our results show that 95%of the total flood recession events could be well fitted with the coefficient of determination(R2)values higher than 0.75.When the decrease rate of rainfall intensity(Vi)is smaller than 0.2 mm/h2,rainfall conditions more significantly control the flood recession process;when Vi is greater than 0.2 mm/h2,underlying surface conditions dominate.The result of backward elimination shows that when Vi takes the values of0.2–0.5 mm/h2 and is greater than 0.5 mm/h2,the flood recession process is primarily influenced by the basin’s average slope and basin area,respectively.The other three factors,however,indicate weak effects in the study area.