Electric Load Forecasting I

Fundamentals and Best Practices

The standard course outline is listed below. The course is being customized based on the background and needs of attendees. See THIS BLOG POST for my teaching method.

Course Outline

Introduction to Electric Load Forecasting
  • An overview of the electric power industry
  • Business needs of load forecasting
  • Driving factors of electricity consumption
  • Classification of load forecasts
Salient Features of Electric Load Series
  • Data pool
  • Trend
  • Seasonality
  • More about load profiles
Mulitiple Linear Regression
  • Naïve models
  • Trend
  • Class variables: weekday, month and hour
  • Polynomials of temperature
  • Interaction regression
  • Rolling regression
A Naive Benchmark for Short-term Load Forecasting
  • Motivation
  • Criterion
  • A Naïve MLR model for STLF
  • Applications
  • Two more salient features
Customizing the Benchmarking Model 
  • Recency effect
  • Weekend effect
  • Holiday effect
  • Case studies
  • Two more salient features
Very Short-Term Load Forecasting  
  • Hour ahead load forecast
  • Weighted least squares regression
  • Dynamic regression
  • Two-stage method
  • Extensions
Medium/Long-Term Load Forecasting  
  • Macroeconomic indicator
  • Weather normalization
  • Forecasting with weather variation
  • Forecasting with cross scenarios
Variables, Methods, Techniques, and Further Readings 
  • Load, weather, calendar, macroeconomic indicators, etc.
  • Similar day and hierarcy
  • Regression
  • ARIMA
  • Exponential smoothing 
  • Support vector machine
  • Artificial neural networks
  • Fuzzy systems and fuzzy regression
  • Relevant and readable books
  • Load forecasting papers
  • Fuzzy systems and fuzzy regression
Frequently Made Mistakes
  • Counterexamples
  • Expectation
  • Data
  • Models
  • Decisions
Software Applications