A Novel Approach to Residential Load Modelling for Grid Planning Purposes
Challenge and objective
Dramatic increase in grid connection demand.
Need for optimized grid utilization - Understanding the characteristics of both current and future loads is crucial for determining available capacity.
Smart meter data is collected, but underutilized for grid planning.
Work performed
A load modelling methodology was developed which utilizes smart meter data and additional customer information without demanding information unknown to the DSO.
The available data was randomly divided into test/training datasets for performance.
Significant results
The methodology offers more detailed output than traditional methods, including hourly load curves, coincidence factors and peak load estimation with uncertainty for different customer groups.
Research provides DSOs with more comprehensive estimates about the behaviour of new customers.
Comparison to traditional model shows the advantage of differentiating between customer groups with different traits.
Impact for distribution system innovation
DSOs can use the methodology to estimate the electricity demand behaviour of new customers seeking connection to the distribution grid to make better decisions of whether it is operationally safe to connect them.
