Analysis of customer profiles on an electrical distribution network

Abstract

any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. Additional information:Use policyThe full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-prot purposes provided that:• a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.Please consult the full DRO policy for further details. Abstract-It has become increasingly important for electrical distribution companies to understand the drivers of demand. The maximum demand at any given substation can vary materially on an annual basis which means it is difficult to create a load related investment plan that is robust and stable. Currently, forecasts are based only on historical demand with little understanding about contributions to load profiles. In particular, the unique diversity of customers on any particular substation can affect load profile shape and future forecasts. Domestic and commercial customers can have very different behaviours generally and within these groups there is room for variation due to economic conditions and building types. This paper analyses customer types associated to substations on a distribution network by way of principal component analysis and identification of substations which deviate from the national demand trend. By examining the variance spread of this deviation, data points can be labelled in the principal component space. Groups of substations can then be categorised as having typical or atypical load profiles. This will support the need for further investigation into particular customer types and highlight the key factors of customer categorisation.

Analysis of clustering techniques on load profiles for electrical distribution

“…Third are the principal component values from [12] for each cluster centroid. As explained in section III, this allows for an idea of the customer make-up to be understood for each cluster.…”

“…For K-means and hierarchical clustering, these are the average values of PC1 and PC2 for each cluster. Associating the customer classification principal components from [12] is not as straightforward for fuzzy clustering because of the allowance for partial membership. However, the impact of the PCA can still be discussed with simple matrix multiplication.…”

“…By using the PCA investigation in [12] it is possible to attribute a customer make-up to each of the primary substations. In particular, there is a need to distinguish between domestic customers and commercial/industrial customers.…”

“…Within domestic customers, there is also a distinction between rural and urban domestic customers. In [12], it was determined that the first principal component (PC1) shows a distinction between domestic and commercial/industrial customers where a positive values indicates a greater influence from commercial/industrial customers and a negative value indicates greater influence from domestic customers. The second principal component (PC2) shows a distinction between rural and urban customers where a positive value indicates a greater influence from rural customers and a negative value indicates a greater influence from urban customers.…”

“…Similarly to other methods used, the load profile centres in Fig. 6 can be explained using general knowledge of load profiles but also using the preliminary work done in [12]. By performing a summation across the membership matrix U , Table III each cluster rounded to the nearest whole number.…”