Occupancy-sensing controls can substantially reduce building energy use and peak demand, especially when controlling lighting equipment. To determine if occupancy-sensing lighting controls are economically desirable, information on the potential savings for both energy charges and demand charges as well as changes in heating requirements must be considered. To evaluate the changes in peak demand and in energy use, peak occupancy schedules should be part of the building energy models. To help predict the demand savings due to occupancy-sensing building controls, peak building occupancy rates can be modeled, given the number of rooms and average occupancy rate.

The potential energy savings when occupancy sensors are implemented in buildings have been investigated. With occupancy sensor controls for the lighting equipment in one or two person offices, the estimates of energy savings vary from 25% to 50%, [CEC, 1993]. According to surveys conducted by the US Energy Information Agency, nearly half of all commercial buildings have the lights on over the entire period of operation, while most of the remaining have the lights on over 50% [EIA, 1992]. The potential savings due to the application of occupancy-sensing lighting controls are significant, especially since the same controls can also control the HVAC system, as described by Morrow [1997].

Cost effectiveness of occupancy sensors should be based on a combination of electrical energy savings, electrical demand reduction, and changes in heating/cooling energy use. In representative utility rate structures with direct charges for electrical demand, energy charges comprise from 75% down to 30% of the overall charges for electricity, with the balance, from 25% up to 70%, determined by the monthly charges based on demand. The increasing complexity of utility rate schedules can complicate the determination of how much of total charges are due to peak demand, but the trend over the recent years indicates that this portion is increasing [PSC, 1997]. Therefore, for design decisions on control suitability, evaluation of the economics of occupancy controls requires that the effect on both energy use and peak demand be established. The standard technique which simplifies electric rates to a single composite charge per kilowatt-hour [IESNA, 1993] may lead to significant errors in predicting the potential savings. In an EPRI report entitled "Lighting Controls: Patterns for Design" [EPRI, 1996], the example of the economics of occupancy sensors for controlling lighting uses a single composite rate for estimating savings. A more detailed discussion is provided in an appendix in the EPRI report. That appendix does caution:

Including imaginary demand savings could overestimate the potential savings significantly, while ignoring potential reduction in demand may lead to a false conclusion that the controls are economically undesirable.

The electrical rate which is most commonly applicable to commercial customers of the local utility, Public Service Company of Colorado (PSC), is Secondary General (SG). According to PSC surveys, for the average SG ratepayer and for the period from 1994 through 1996, between 60 to 70% of the total electrical charges are attributable to demand charges [PSC, 1997]. The trend over this period shows that the portion due to demand charges is increasing.

Reduction of the lighting load increases heating requirements and costs, while decreasing cooling requirements and costs, as described by Rundquist, Johnson and Aumann [1993]. These changes should be considered in the economic evaluation of occupancy-sensing controls. Cooling typically corresponds to annual electrical demand peak, while heating corresponds to natural gas use. Therefore, the effects of occupancy-sensing controls on these systems are distinct and should be considered separately. This is especially true when the significance of the peak demand increases, either through high monthly demand charges or carrying over one month's peak to other months' charges (i.e. ratchets).

The complexities of estimating peak electrical demand for a facility include the interaction of multiple systems and the variations inherent in daily and seasonal cycles, and the weather, as well as the idiosyncrasies of any business or institution. Existing building simulation software can model such facilities and provide information on these interactions, but typically use limited occupancy information. Average occupancy rates, either daily or hourly, do not represent the variability which in some facilities contributes to peak demand. Average occupancy rates generally correspond to energy use savings, while peak schedules correspond to peak demand reductions. Depending on the level of controls in the model and in the relative significance of lighting in the overall electrical demand, there may be substantial difference between using average values or peak values for establishing peak demand effects. The critical issue in evaluating the effects of controls on the peak demand is the temporal coordination: what is the potential reduction at the time when the peak would have occurred if the controls were not in place? The effect of the reductions may include changing when the peak would occur, so this question should be sequentially applied to each period of peak demand over the duration of a model, until the reduction due to the controls is established

Occupancy sensing controls provide an instantaneous change in the electrical demand for those systems that are controlled, as well as a thermal effect of longer duration. In closely controlled spaces, with private or semi-private offices, occupancy controls shape the demand profile for the controlled lighting, as well as modify the profile for the HVAC system. For that portion of a facility which is offices, the controlled lighting load can be substantial enough that reduction in peak demand can carry through to a contemporaneous reduction in demand charges.

Previous work by Stannard, Keith and Johnson [1997] for the Electrical Power Research Institute, included information concerning the occupancy patterns for office facilities. The available reliable information was limited, and addressed average office occupancy, in hourly values for a typical profile over workdays or weekends, as presented in the ASHRAE/IES Standard 90.1-1989 [Table 13-3].

Peak occupancy data has not been available for use in modeling the performance of occupancy-sensing controls. Therefore, evaluation of the economic benefits available in applying such controls was difficult or incomplete.

The initial problem is the collection of occupancy data with sufficient scope and duration to provide peak occupancy information. This requires an extended period of data acquisition in a facility which is suitable for providing such data. The next step is the application of this data in models to evaluate the significance of using different occupancy schedules in predicting the economic benefits of applying occupancy-sensing controls to lighting systems in offices. If the effect of applying these peak occupancy schedules in models is significant, then the next problem is to develop a simplified tool for estimating the monthly peak occupancy schedule for an office facility, based on the number of office rooms and the average occupancy of those rooms during the month. For the period between 8 AM and 5 PM, Monday through Friday, peak occupancy values could be estimated for each hour or for the entire period to provide daily or hourly information as desired. The final problem is to validate the results from the simplified tool against data collected for this work and other suitable and available occupancy data.

In summary, the primary objective of the work presented in this thesis is to determine the significance and usefulness of peak occupancy data in evaluating the economic benefits of applying occupancy-sensing lighting controls in offices. The secondary objective is to develop and validate a simplified tool for estimating the peak occupancy data and to make it accessible to designers and engineers to evaluate the benefits of occupancy-sensing lighting controls in offices.

The literature survey conducted for this research work has four parts. The first part is a review of literature concerning the energy and economic opportunities in using occupancy-sensing lighting controls. The second part is a review of existing work that dealt with the significance of applying average or peak occupancy rates. The third part is an investigation of published data on occupancy profiles: hourly or monthly, average or peak. Finally, information on the local utility rates applied to the facility for which the occupancy data was researched.

There is great potential for application of occupancy-sensing controls for lighting equipment. The Illuminating Engineering Society of North America (IESNA) states that 20-25% of all the energy used in buildings is used for lighting, corresponding to 5% of the total energy use in the United States [1993]. In addition the heat from lighting accounts for 15-20% of the cooling load for buildings. The California Energy Commission reports that over 40% of the electric energy use in California is for lighting [1993]. The U.S. Energy Information Agency reports in Commercial Buildings Energy Consumption and Expenditures 1992 that 63% of the building have 100% of the lights on during the building operating hours, and 19% more of the buildings have between 51% and 99% of the lights on. While only 17% of the buildings in that survey are identified as offices, many of the other categories listed will have some office spaces as part of the building. The potential for peak demand savings is substantial, since for the buildings with 100% of the lights on, the median peak Watts per square foot is 5.37, while the 75th percentile is 10.00 peak Watts per square foot. This difference indicates that approximately 25% of the buildings in this survey have peak demand that is nearly twice the peak demand of an average building.

Many sources report the potential for energy savings due to occupancy sensing lighting controls, but information on peak demand savings is limited. The California Energy Commission (CEC) shows manufacturer's estimates of 25-50% potential energy savings for small offices for 1 or 2 people [1993]. Even greater energy savings, 35-65%, are reported for private offices in the Electric Power Research Institute's Lighting Controls: Patterns for Design [1996].

The second part of the literature review summarizes the existing work that investigated the significance of applying average or peak occupancy rates. Stannard, Keith and Johnson [1997] provide a brief discussion of the impact of occupancy profiles on the reduction in peak demand and energy use due to lighting controls. The results of a survey by the Lighting Research Center discusses the importance of considering occupancy use patterns in predicting potential savings [Mannicia, 1993]. This survey was made by inspection of rooms once per hour, and the results include occupancy profiles from schools taken over four day period. While this information indicates energy savings, the following economic evaluation uses a single kWh charge, and the conclusion is that occupancy-sensing controls are not justified. The CEC Advanced Lighting Guidelines, Occupant Sensors section provides a figure with relative lighting energy profiles for manual and occupancy-sensing lighting controls which shows a significant (25-35%) difference in the percent of maximum lighting energy used during the period between noon and 6 PM [1993, p. 11]. This may be considered as a peak demand profiles for the lighting equipment alone, and shows a potential savings in the peak demand which is significant but less than the estimated energy savings.

The investigation of published data on occupancy profiles includes some of the work discussed above. The most widely recognized source for occupancy data is the ASHRAE / IES 90.1-1989. In this document, Table 13-3 includes hourly profiles for office occupancy and office lighting. For weekdays between 8 AM and 5 PM, the average occupancy rate is 0.76, with a peak of 0.95. For the office lighting profile, the average is 0.89, with a peak of 0.90. These two profiles are shown in Figure 1.1. The occupancy profiles in ASHRAE/IES 90.1-1989 are intended to apply to every weekday over the entire year.

The final portion of the literature survey is a review of information concerning electrical charges from the local utility. Public Service Company of Colorado (PSC) uses secondary general (SG) rates for typical commercial customers. Information from PSC survey's of its SG rate customers for the period 1994 through 1996 shows that the typical customer's monthly bill is around 67% due to demand charges and 33% due to energy charges [PSC, 1997]. The trend over these years indicates that the demand portion increases, due to increases in demand and in rates, as shown in Figure 1.2.

Figures and Tables for Chapter 1

Chapter 2

Table of Contents