ABSTRACT
This chapter presents a point process characterization of peak electricity demand. It discusses an analysis of power systems reliability in the presence of random extreme peak loads. The chapter focuses on long-term peak electricity demand forecasting using additive quantile regression models. Modeling the upper tail distribution of peak electricity demand is important as there are times when high load values exceed the maximum generated electricity, which could cause problems to the system, including grid instability. A severe stress on the grid system occurs when the peak electricity demands' amplitude exceeds a predetermined sufficiently high threshold that is determined using extreme value theory methods. Some of the important measures in reliability modeling include reliability indexes such as loss of load expectation, which is derived using daily peak load, including among others the extreme peak load frequency, which is the average number of peak loads above a sufficiently high threshold.
