Exogeneity investigation and modeling energy demand via parallel dynamic linear models for maximum simultaneous power demand

As a solution to power system planning and control problems, it is essential to know the whole demand for the electrical energy needed in a country. Besides governments, also for the private sector interested in energy supply industry, it is important to know the set of social, economic or technical variables affecting the maximum demand for electric energy. This paper starts a discussion of the problem based on an oil-producing country, Iran. It is an attempt to recognize exogenous variables in the demand system. The paper investigates exogeneity of a variety of quantified variables through 13500 parallel models. Different combinations of several exogenous variables that are guessed to be effective on the system generate these models. First, models are filtered applying coefficient sign significance criterion and error validity. Finally, a fuzzy decision-making process selects winner models among 480 remaining models. The selected model introduces the most vital elements that can be used later to establish a more complicated nonlinear model.