To address additional sources of risk due to deregulation of the retail electricity market, a major US utility approached Prescio Consulting to improve their load profiling efforts.

The utility wanted to develop a new load profiling model and application software in order to aggregate the present and future customers into finer load profile classes. Load profiling classifies a customer into a given group according to certain key characteristics. The finer the aggregation within a group, the higher the accuracy of the load profiles. The primary purpose of a finer aggregation is to attempt to reduce the biases resulting from assigning a class load shape to a subgroup of customers whose shape may be different from the class shape.

Prescio Consulting designed a load profiling model to statistically attempt to segment different classes of customers into finer sub-classes with homogeneous and functional characteristics This was done to minimize misclassification errors. Additionally Prescio Consulting fully incorporated weather patterns and economic indicators, as well as time characteristics into the profile curve models to make them dynamic and realistically driven by these key factors. The load model was also designed to forecast monthly average usage per profile class. These monthly aggregates could then be distributed over hours in each forecasted month, factoring in expected weather conditions, time of day characteristics, and economic activity trends.

A customer-switching model was developed to calculate switching impacts on the number of customers in each profile class. The switching model was used to generate the dynamics of the population over time within each profile class. Customized application software was then designed and developed to be seamlessly incorporated within the existing hardware and software architecture of the utility. The model was coded in C with a VB interface and Oracle 8i database.