Using Load Profiles to Determine EPSS Peak Demand Load

David Stymiest, PE CHFM FASHE, DStymiest@ssr-inc.com

The author has used load profiling to determine EPSS peak demand load for more than 20 years.  The discussion below is intended to highlight some of the lessons learned.  For a full discussion, refer to the author’s 2009 ASHE Management Monograph, “Managing Hospital Emergency Power Systems – Testing, Operation, Maintenance and Power Failure Planning” available at http://www.ashe.org/.

NFPA Disclaimer: Although the author is Chair of the NFPA Technical Committee on Emergency Power Supplies, which is responsible for NFPA 110 and 111, the views and opinions expressed in this message are purely those of the author and shall not be considered the official position of NFPA or any of its Technical Committees and shall not be considered to be, nor be relied upon as, a Formal Interpretation. Readers are encouraged to refer to the entire text of all referenced documents.  NFPA members can obtain NFPA staff interpretations at www.nfpa.org.

Hospitals should document their actual EPSS peak demand load.  It is not enough to assume that the highest emergency generator kilowatt (kW) demand during an early-morning monthly test represents the true peak EPSS demand load.  This is a poor assumption due to the variability of mechanical, building, and clinical process loads during a typical hospital workday.  If an EPSS test time is chosen due to low clinical activity, then that avoided clinical load will not be reflected in the EPSS test loading.  Additionally, some equipment, such as smoke control systems and fire pumps, will not operate except during atypical situations.

Determining the actual emergency demand load cannot usually be done through one simple measurement because the EPSS automatic transfer switches will usually be connected to the normal power system along with other normal power loads.  However, portable recording instrumentation can be used to meter the loads on each transfer switch for 30 days in accordance with NEC^® 220.87, a much better approach than short term sampling with a hand-held ammeter. 

The differences between the 15-minute demand load and the instantaneous peak loads should be noted so that the hospital engineer can assess how much additional allowance should be made for short-time variations. The assessment should differentiate between the “instantaneous” peak loads that last only a few seconds, typically representing motor starting inrushes, and the longer but still short-duration peak loads that last for one or two minutes.  These one or two minute peak loads should be noted for later determination of an overall allowance for short-time load variations.

If a power management system is not available, record the load side of each automatic transfer switch for 30 days in accordance with NEC^® 220.87.  Add the separate transfer switch load profiles together, along with allowances for other unrecorded loads, to determine the total EPSS load.  If a power management system is available, then a 30 day report of simultaneous time-of-use transfer switch load readings will provide a reliable load analysis.  In BOTH cases, adjustments will have to be made for variations between the loads that were running when the measurements occurred and other loads that could run during a normal power outage as stated below. 

If 1 year of maximum demand information is available for each transfer switch, these demands can be added together in accordance with NEC^® 220.87, but there is a downside to this approach.  The maximum demands for each transfer switch are not likely to coincide with each other, and the calculated sum of the non-coincident demands (also called the “sum of the peaks”) will be higher than the real EP System peak loading (also called the “peak of the sums”) that one would obtain by adding together the actual load profiles.

The referenced load profile template illustrates an approach for determining the daily load profiles of separate branches in a sample hospital building.  It also illustrates allowances for certain other types of loads.  Both a chart of the individual load profiles and a chart of the composite overall EPSS load profile are included.  The total EPSS load profile results when the individual load profiles and allowances are added to obtain a composite generator load profile for that same building.    

This strategy usually gives repeatable peak values because most hospital loads and processes are repeatable.  The author’s experience reviewing thousands of load profiles indicates that daily ATS load profiles taken in the same hospital building over time tend to show similar characteristics and values.  The author’s experience is that load growth and space or occupancy changes are the major factors that generally cause the load profiles to change significantly.

The load profiles of one building or branch should not be used to predict the load profiles of other buildings or branches, since variables such as building size, specific occupancies, occupancy patterns, and energy conservation features all affect the load profiles.   Note that certain types of loads are not likely to be running during normal operating conditions (i.e., the fire pump, the smoke control system, the fire alarm system in “alarm” condition.)  These atypical but necessary items need to be modeled with allowances.

Certain EPSS ATS loads do not normally vary during the day.  The code limitations on what kinds of loads may be connected to the life safety branch usually make its demand stable, except for the impact of a fire alarm condition on the fire alarm system demand.  One healthcare facility was recording the load profile on a life safety transfer switch in a high rise building when such a fire alarm condition occurred and the load demand profile that resulted doubled as a result of the fire alarm.

Similarly, critical branch loads in many patient care areas tend not to change very much throughout the day unless portable radiology equipment is brought into the area and plugged into the critical branch outlets.  Operating room critical branch loads, of course, vary significantly depending upon the status of the operating room.  All patient care unit critical branch loads will vary, of course, based upon the unit’s census and patient acuity.

Equipment System loads, however, will often vary during the day.  This is particularly true if energy management strategies are used to turn off loads like operating room ventilation fans when the operating rooms are not in use.  Elevator and radiology loads are the most variable, even during regular working hours.  In addition, elevator and radiology loads can provide the highest inrush impact on the generator.  There can be a major difference between generator loading with and without elevators, as well as with and without radiology loads.  Load profiles indicate that monitoring of elevator loads under normal power conditions does not allow one to predict the elevator loads under emergency power conditions with much accuracy.  A month of radiology load metering is usually required before one can accurately model the radiology load impact on the emergency generator.

Please be sure to read the above post before obtaining the referenced load profile template.   

Again, it must be noted that NEC^® 220.87 requires 30 days of continuous load monitoring per transfer switch if the past year’s peak demand load is not available.

NFPA Disclaimer: Although the author is Chair of the NFPA Technical Committee on Emergency Power Supplies, which is responsible for NFPA 110 and 111, the views and opinions expressed in this message are purely those of the author and shall not be considered the official position of NFPA or any of its Technical Committees and shall not be considered to be, nor be relied upon as, a Formal Interpretation. Readers are encouraged to refer to the entire text of all referenced documents.  NFPA members can obtain NFPA staff interpretations at www.nfpa.org.