Abstract - D. McQueen, MSc, Energy Management
Mathematical Modeling of Residential Electricity Demand and Engineering Applications
Dougal McQueen - 2002
The objective of this thesis is to develop new methodologies for Low Voltage network design. These methodologies, utilizing the Monte Carlo method, are intended to replace the existing conservative design methods that are not suitable in a competitive environment, where design optimization is necessary.
A metering study of residential electricity demand and received power quality is performed with the aim of developing mathematical models to simulate residential electricity demand. The demand on a distribution transformer is also monitored for purposes of model verification. Mathematical models are progressively developed to characterize differing aspects of residential electricity demand that contribute to diversity. Modelling methodologies include Markov chains, Fourier series analysis, and non-homogeneous Poisson equivalent rectangular pulses.
Justification of the models is provided by simulation of the distribution transformer’s demand and matching this against the measured demand and historical MDI measurements. Justification is further provided by simulating the voltage drop in a LV network and comparing this against the voltage drop predicted using the traditional engineering methodology and measured values.
Further applications include: simulation of the LV network profile while changing the hot water cylinder element size, simulation of the LV network profile under a controlled period demand experiment, voltage unbalance due to diversity, and the effect of the sampling interval on demand measurement.