Two Economists at the University of Alberta in Canada published a study in the journal Energy studying the energy efficiency of Canadian Commercial Buildings. The paper, entitled The potential for energy efficiency gains in the Canadian commercial building sector: A stochastic frontier study, was published in 2005.
(J. Buck, D. Young, Energy 32 (2007) 1769–1780)
I trust that nobody is against energy efficiency, but the matter is somewhat complex. While it is relatively easy to determine a ratio for instance of units of energy (e.g. joules, kilojoules, gigajoules, etc.) per unit of volume (cubic meters or cubic feet) or maybe less ideally (but more commonly) floor space square meters or square feet, it is far more difficult to determine why a particular building is less efficient than another. Since the resources for construction and design (and/or retrofitting) are not infinite, it is important to have economic efficiency, to determine whether it is better to put on a different type of roof, a different type of furnace, better insulated windows etc.
In this effort, the authors attempt to quantify parameters in a mathematical fashion that will help to elucidate the relative contributions of various strategies.
The authors write in the body of the paper:
...not all buildings will use energy this efficiently, or as efficiently as their bestperforming counterpart. Some differences in energy use will be due to random factors such as unusually good or bad climatic conditions, occasional malfunctions of equipment,etc. Other differences may be due to differences inthe incentives faced by building managers and other owner specific factors... For instance, some owners may circulate energy-saving tips to building occupants and some owners may include utilities in the rent...
To capture these two different types of impacts on energy efficiency, the stochastic frontier approach decomposes the random portion of energy use into two components. The first of these components, which is included in the stochastic frontier, is a general random shock that can be positive or negative. The second component consists of a non-negative random shock, which is a function of characteristics that determine the extent to which energy use in a particular building exceeds the efficient amounts indicated by the stochastic frontier...
...Space conditioning (heating and cooling) is the most significant factor in terms of energy consumption, accounting for more than half of all energy use. Heating, ventilation and air conditioning (HVAC) design is therefore a candidate to play a major role in terms of energy efficiency. While hot water plays a minor role in energy consumption (4–6%), artificial lighting may play an important role (about 15%)...
The data set is described as follows:
Our data on energy use in Canadian commercial buildings are extracted from the CIBEUS survey conducted by Statistics Canada in 2001 on behalf of the Office of Energy Efficiency at Natural Resources Canada. The CIBEUS data set contains detailed information pertaining to 4101 commercial and institutional buildings in major metropolitan areas of Canada. The information gathered includes data on several owner and occupancy characteristics, physical building characteristics, types of energy efficiency technology in place, recent retrofit decisions, and energy use. For our study, these data have been supplemented with city-specific degree-day data, available online from Environment Canada [25]...
Some of the variables included are quite detailed, issues such as window glazing type, district steam availability, the nature of the wall studs (wood or metal), the nature of lighting ballasts, the ratio of window space to wall space, dimmers...
The presence or absence of more than 50 parameters (including types of use and types of owners were evaluated.)
The results of the study are interesting and in some cases, counter-intuitive.
The stochastic frontier portion of the model used in this study indicates which building and exogenous environmental characteristics have significant impacts on the intensity of energy utilization. The results indicate that larger buildings use energy less intensely...
...It appears that only a few of the specific physical characteristics available in our data set have significant individual impacts on the shape of the stochastic energy frontier. Buildings that use district hot water or packaged units for heating purposes use energy more intensely than those using conventional furnaces. Compared to fuel/ heating oil, the use of electricity or liquid petroleum gas or propane leads to a lower intensity of energy use, while the use of natural gas increases energy use intensity...
...Unfortunately, building specific energy cost data were not collected in the survey. While the type of heating equipment and the choice of main heating fuel matter, specific HVAC conservation features do not. Similarly, lighting conservation features have no significantly distinguishable impact on the location of the stochastic energy frontier...
...While the type of heating equipment and the choice of main heating fuel matter, specific HVAC conservation features do not. Similarly, lighting conservation features have no significantly distinguishable impact on the location of the stochastic energy frontier. In terms of building envelope thermal integrity variables, the window-to-wall ratio has an impact on the intensity of energy use, but the type of window chosen does not. Wall type is a factor in concrete and curtain walls, which leads to increased intensity of energy use. In addition, buildings with walls adjoined to other buildings tend to use energy less intensely. None of the roof type variables was significant, either individually or jointly...
...Are Canadian commercial sector buildings energy efficient? The minimum possible inefficiency attainable according to Eq. (3) is 1.0, while the maximum is not bounded. Inefficiencies estimated from our stochastic frontier model range from 1.13 to 51.94 with 785 buildings (or more than two-thirds of the sample) having inefficiency ratings below the average of 2.70...
I thought this paper interesting.