A Temporal Approach to Characterizing Electrical Peak Demand: Assessment of GHG Emissions at the Supply Side and Identification of Dominant Household Factors at the Demand Side

Abstract

When electricity demand is at its highest it is most costly to generate and transmit and is usually considered to produce the greatest greenhouse gas (GHG) emissions, so reducing peaks can have a double benefit. In most nations, these peaks in demand occur daily in the mornings and evenings. This PhD research developed a new way of assessing GHG emissions produced by the generation of electricity at different times of the day, including at peak time. It also investigated the dominant factors driving peaks in household demand and developed a new analytical approach to identify these factors so as to assist in developing targeted demand management.

Two contrasting countries - New Zealand and Bangladesh - were chosen to apply the research. These two countries have very different climatic conditions, economic conditions, socio-demographic characteristics, electricity generation sector, and emissions from electricity generation, so the research findings could be tested and compared.

To assess GHG emissions, an analytical approach was developed - ‘time-varying carbon intensity analysis (TVCIA)’ - to explore the relationships between GHG emissions and peaks in demand. Applied to 2015 data from New Zealand, a country with around 80% renewable generation dominated by hydro, it was found that New Zealand’s carbon intensity was largely uncorrelated with demand. This finding was counter to some perceptions in the electricity sector in New Zealand where it is assumed that peak demand always means higher GHG emissions.

In contrast, when the method was applied to Bangladesh, which has an electricity system dominated by fossil fuel generation, it showed that daily peaks in demand had the highest GHG emissions. Therefore, reduction in demand at peak times could be a potential option to reduce GHG emissions in Bangladesh. In New Zealand, seasonal demand management could be beneficial as GHG emissions can increase significantly in a dry year when hydro lakes are low.

For the latter part of the research, a methodology called ‘time-segmented regression analysis (TSRA)’ was developed to identify the dominant factors driving peak electricity demand in households. Applied to a New Zealand dataset, the analysis revealed that methods of water and space heating were the dominant factors in determining peak demand in New Zealand’s households. In contrast, the number of occupants and the number of electrical appliances were dominant factors in determining peak demand in Bangladeshi households.

Together the new approaches that have been developed can assist nations in determining the GHG emissions from electricity generation at different times (over days, weeks, months or years) and also determining what factors in households are driving peaks in demand. This is important to help design more effective, targeted energy efficiency and demand management strategies. Together these methods can help in devising programmes for reducing GHG emissions from electricity use.