An Analysis of Grocery Store Energy Use

Abstract

Approximately 3% of the United States' commercial building energy consumption is attributable to food sales facilities. Although this is one of the smallest consumption percentages, it is still significant, amounting to about 151 trillion Btu, or $2.17 billion per year. Food sales facilities ranging from 10,000 to 100,000 sq. ft. use 3 to 5.5 W/ sq. ft. (32 to 60 W/ sq. m.) of electricity two to three times what

typical office buildings of the same size use (EIA 1986). Identifying potentials for energy savings in food sales facilities is therefore a worth-while pursuit.

Why do people study energy consumption? According to Haberl et al. (1990), there are five different groups of people who can benefit from building energy monitoring and at least seven basic applications of energy monitoring projects. The five groups of beneficiaries are: the energy analyst; the energy consumer; governmental agencies; engineers, manufacturers, and contractors; and, utility and fuel suppliers. The seven basic applications are: energy consumption and load forecasting, evaluation of end-use energy data, the monitoring of energy savings from retrofits, determining system .efficiencies, environmental quality issues, analyzing the human factor, and diagnosing operational and maintenance problems.

This thesis is a study of the energy use in supermarkets, which fall into the category of the energy consumer. This study is of interest to the energy analyst and the manufacturers of grocery store equipment, and to utilities which can use the results of energy consumption modeling procedures developed herein as inputs to load-predicting models. Many papers and reports have been written about the energy use in grocery stores. In general, they addressed three major issues: energy use surveys and market analyses, refrigeration and HVAC system improvements, and energy use modeling methods. This thesis extends the foregoing work by first performing a general energy use survey of over 90 grocery stores, and presenting statistics regarding their energy use characteristics. Then, several of the previous methods of energy consumption modeling are adapted and applied to the whole-

building and sub-metered component load data from two case study grocery stores. Two methods of modeling, multiple linear regression and principal component analysis are evaluated. Finally, a new method is developed and tested that allows for the accurate estimation of sub-metered loads without incurring the expense of collecting many months of hourly, sub-metered data.