The world demand for energy is steadily increasing but the contributions of energy sources are continuously changing. There are several trends in energy development among the trends is change in energy types: from high- to low-carbon, or from fossil to non-fossil energy sources due to the world wide concern of the carbon foot-print and climate change. Other trends include the technological advances in energy production such as fracking, advances in solar power generation, and wind energy. Moreover, improving the efficiency of production processes as exemplified by the design of larger turbines that increase the energy production per turbine as well as improving the efficiency of the plant are major challenges in wind energy.
Data Science for Wind Energy addresses the production process of wind energy. The author’s background and training in mechanical engineering and data analysis and modeling positioned him to develop accurate models for wind energy production, prediction, and efficiency.
There are many books and resources on “data science” written by authors from different disciplines including computer science, statistics, business, and engineering. Similarly, there is an equivalent list of books on wind energy. However, none exits that links the two areas. This book fills the gaps with strong physical and statistical models. The book begins by providing an overview of wind energy, its importance and issues in predicting its output energy due to many sources of uncertainties. The following chapters provide forecasting models which can be applied for energy output and demand. The new contributions are in the spatio-temporal models where determination of wind directions and magnitude at different locations in wind energy “farms” is critical to the overall performance of the energy system. This is well explained and presented in Chapter 3 which focuses on covariance functions with application to wake effect in the energy systems, it contains interesting and useful statistical approaches such as kriging and Gaussian spatio-temporal autoregressive models which are then applied in local wind field taking into consideration the geometry of the wind turbine blades and other factors. The following chapter focuses on regime-switching approach where nonstationary processes (such as wind speed) is changed to stationary process (regime change) to be able to obtain computationally efficient models. This where “data science” is effectively applied.
This theme continues in the following chapters where power curve modeling and analysis are treated in-depth. The physical model of the wind turbine is incorporated to obtain realistic power curves. One of the most challenging aspects of the physical structures of the wind energy system is the maintenance and replacements of parts specially when operating under off-shore environments or in large “wind” farms set up. Monitoring and prediction of failure times of the critical components is data-rich environment and the use of data science is effective in identifying failure modes and the failure times. These are carefully and clearly addressed in this book.
This is the first book that focuses on the data science methodologies and their applications in a growing field, wind energy. It is well-organized and well-written. It will enhance the knowledge base of data science and its applications in the wind energy field.
Rutgers University