Kenneth H. Tiedemann
Engineering algorithms, Regression modelling, Statistically adjusted engineering estimates, End use electricity consumption
Electric utilities need end use consumption information for planning, forecasting, operational, and evaluation purposes. Statistically adjusted engineering (SAE) analysis has been proposed as a means of overcoming the weaknesses of conventional engineering analysis while retaining the strengths of rigorous and empirically grounded engineering algorithms, but it has rarely been applied to industrial customers. This study reports on the results of a detailed SAE analysis based on data for industrial facilities in British Columbia. Data for the study came from detailed site audits and engineering analysis at 198 industrial sites, which provided detailed end use consumption estimates for nine end uses for each of the sites. This information was used in the SAE model which was modelled using ordinary least squares. The study has three main results. First, the estimated regression model had excellent statistical properties: the signs for the regression coefficients were consistent with prior expectations, the magnitudes of the regression coefficients were reasonable, and the regression had excellent explanatory power. Second, large electric utilities often have information for a sample of customers on end use electricity consumption, and the present research suggest that the relatively little used SAE approach can be used to cost effectively estimate end use electricity consumption. Third, total and end use electricity consumption estimates vary substantially by industrial sector, which suggest that disaggregated information may be more useful than the aggregated data which is typically collected through load research activities.
Important Links:
Go Back
