Submitted: March 2014
Abstract
Nowadays the ability to predict the electrical consumption becomes more and more important, in a time where the whole electric power system is currently undergoing a fundamental change—resulting in many decentralized supply units. Along with this, several components are subject to an increasing load. These modifications require to think about various so far efficient ways to guarantee trouble-free service.
This thesis is therefore devoted to the investigation of a factor of great importance used precisely to estimate the maximal stress that parts of the system have to be able to withstand—the so-called coincidence factor. The objective is to shed some light on the nature of this very value, in particular in relation with stochastic consumption processes.
We will start with some basic considerations to understand the information value of the coincidence factor. This fundamental part is followed by some illustrative notes on the importance of the group size. Afterwards we will discuss the meaning of the coincidence factor as a random variable and which consequences this notion implies. Especially the current application of this value is questioned in the setting of stochastic consumption. In the last part we will introduce the so-called maximum entropy method. We will explain how this technique could help us in finding appropriate probability density functions for the stochastic consumption model by bringing all the insights together that we have gained in the previous chapters.