Planning for modular power system growth
How to meet current standby power and electrical distribution needs while also accommodating future expansion.
“In this age of electrical development, estimating future electrical load growth for buildings is becoming an important duty of building designers—a duty that is both difficult and controversial.” This was the summary statement made in a 1957 report by the Building Research Institute on “Electrical Load Growth in Buildings.” The report investigated the electrical demands for office buildings in metropolitan New York in an attempt to assist administrators and designers facing the problems of estimating future electrical load growth. The report concluded that the electrical demand loads increased as much a three times in the previous 10 years with the primary causes being air conditioning, higher light intensities, business machines, and appliances.
More than 50 years later, consulting engineers and administrators are still struggling with the difficult task of predicting the future and estimating electrical load growth. How much future expansion do I allow for? How much do I pre-invest on future expansion? How do I design and implement an electrical system that meets the current load requirements and allows for future expansion? These are all important questions that need to be considered when planning and implementing an electrical system.
The first step in planning an electrical system that can accommodate future growth is to determine the amount and timing of the expansion. Although the technology has changed since the Building Research Institute report was published, the factors that made up the extent of the growth remain similar. These factors include building expansion, function of building, and equipment technology within the building.
Load increase due to building expansion is probably the easiest of the three factors to predict and plan for when designing a power distribution system. Most buildings or facilities are now designed around a master plan detailing any and all potential for future growth. In this case, the load growth can be calculated or estimated based on the size of the expansion. If no master plan was developed, the likelihood and magnitude of the expansion can be based on the location of the building.
In general, buildings on a small property in a suburban environment have less opportunity for expansion than those on a large property in a rural or campus-type environment. Therefore, for a small property in a suburban environment, a 0% to 10% building expansion factor may be used versus a 10% to 25% building expansion factor for a large property in a rural environment. Building expansion can also be internal, especially where certain spaces within the facility were left as shell space to be fit-out in the future.
The next two factors, building function and equipment technology, are interrelated. Certain types of buildings are more prone to load growth than others due to the tasks being performed and the ongoing advancement of the equipment technology being used. Lab and hospital equipment continue to become more electronic with the need to store and access substantial amounts of information. Data center equipment continues to become more compact with higher power use in a smaller space. The following are projected load growth factors for different types of building functions based on a 10- to 15-year period.
- Office building: 5% to 15%
- Laboratory building: 15% to 35%
- Hospital: 15% to 35%
- Data center: 50% to 200%
Changing the function of the building or portion of the building is less predictable and usually not accounted for in the design load analysis unless specifically identified during programming. For example, standard office space could be allocated for conversion to data center/laboratory should the data center/laboratory need to expand.
Advancements in technology have generally been associated with an increase in load density. More components are packed in the same space, drawing more power. Over the past few years however, energy-saving initiatives and regulations have forced technology to do the opposite. More efficient equipment, energy-saving techniques (variable speed drives, occupancy and daylight sensors, harmonic mitigation, cloud computing, etc.), and HVAC taking advantage of more free cooling opportunities have lowered the overall load of facilities. This reduction in overall energy use is excellent for facility owners and the society as a whole, but it further complicates the estimating and planning for future load growth.
Although there is no intricate analysis or formula to accurately predict the future, the growth factors stated above and personal knowledge of the facility can be used to estimate the maximum and minimum potential for future load growth. The estimated load values can be integrated into an energy profile for the facility showing the initial load and the projected load (minimum and maximum) as it increases over time. The design of the electrical system should ultimately fall somewhere in between the two load profiles.
The next step in planning an electrical system is to determine how much pre-investment capital should be spent on future growth. Three basic theories are associated with this determination.
Denial theory: This theory involves the denial from the designing engineer and owner that the building will ever increase in load. Under this option there is no pre-investment in future growth. The engineer designs and implements an electrical system that serves the current load only and does not allow for future load growth. This option may be the most cost-effective solution on day one, but it makes accommodating any future load increase very expensive—and maybe even impossible. This theory can often be the result of people believing the existing design is already oversized.
Figure 1: The big bang theory suggests that you pre-invest in all the equipment necessary to serve final projected load. Courtesy: KlingStubbins
Big bang theory: The next theory involves going for the big bang at the start of the project. Under this option, the pre-investment is for all of the equipment necessary to serve the final projected future load. The engineer designs and implements an electrical system that can serve the final projected future load capacity (see Figure 1) while using only a portion of it to serve the current load. This option has a large up-front capital cost, wastes available capacity, and can actually increase operating cost due to certain equipment operating inefficiently at lower loads. This option also involves a level of risk because if the load does not reach its projected growth, that pre-investment of capital in the equipment is lost (Figure 2).
Figure 2: Pre-investment of capital is lost when load does not reach its projected growth. Courtesy: KlingStubbins
Pay-as-you-go theory: The final theory involves a more stepped modular approach that allows the electrical system to grow with the load. Under this option the pre-investment includes equipment to address short-term growth and options within the distribution system to allow for future expansion (see Figure 3). The engineer designs and implements an electrical system to meet the current load demand with the ability to add equipment in steps as the load increases. This option can reduce operating cost by allowing the equipment to operate at its maximum efficiency and eliminate risk because there is the minimum pre-investment of capital in equipment. If the load does not increase to the projected growth, there is no need to build out the facility or expend additional capital for new equipment (see Figure 4).
Figure 3: In the pay-as-you-go theory, a stepped, modular approach is used to grow the electrical system with the load. Courtesy: KlingStubbins
The pay-as-you-go theory is the preferred method of design because it minimizes up-front capital costs. This modular approach to the electrical systems is very common in data centers where load growth and densities have increased exponentially over the past decade.
Figure 4: Capital is not spent if load does not reach its projected growth. Courtesy: KlingStubbins
Design and implementation
The following are five key components in designing a successful modular electrical system:
- Establish load steps that match the electrical components.
- Don’t break the electrical system into too many steps.
- Avoid stranding available capacity.
- Include all levels of the distribution system in the modular design.
- Ease and isolation of the expansion.
Selecting load steps that best match the electrical components within the electrical system is a key component in designing a successful modular electrical system. Generators, UPS systems, and power distribution units (PDUs) all have standard ratings. The goal is to find the “sweet spot” where all those components can grow at the same pace so that you don’t once again pre-invest. This can also be true with the mechanical systems. Recommend coordinating the modular growth of the electrical system with the mechanical systems so that they also grow at the same pace.
Some systems, like a UPS, are already available in a modular design. Small, modular UPS systems require a pre-investment in the cabinet, incoming feeders, and distribution but defer the actual UPS block cost until they are actually needed to meet the loads. Larger UPS systems work in a similar fashion with the pre-investment including the incoming and distribution systems and the controls to parallel multiple modules. The UPS module and battery costs are deferred to a later date.
Another key component in designing a successful modular electrical system is not to break the electrical system into too many steps. If the future total load is estimated at 3000 kW, design options include installation of three 1000 kW generators or 12 250 kW generators. The raw cost of a 250 kW generator is about 30% to 35% of a 1000 kW generator. Adding the cost of paralleling controls, breakers, feeders, and installation can easily increase that number to 40%. Simple cost analysis (excluding investment of the funds not spent and inflation) will show the final cost of 12 generators is greater than that of three 1000 kW generators. There are also more indirect costs associated with 12 generators, such as maintenance, that have an adverse effect on breaking the system into multiple small steps.
The next key component is to avoid designing a modular electrical system that strands available capacity. This is frequently observed in standby power systems that are sized to meet an estimated calculated load. Often that load does not materialize due to functions not being performed, equipment not being installed, or even the cooling system not operating during the winter (free cooling), creating available capacity that could serve additional loads. If the distribution system is not set up to adjust for that fluctuation in load, then the capacity gets stranded within the system.
To compensate for that fluctuation in load and avoid stranding capacity in the standby power system, a load add and load shed scheme can be used. This method categorizes the standby load into priorities. The load on the generator is constantly monitored, and if capacity becomes available within the standby system, loads are automatically added. If the load increases above the system capacity and/or a generator is lost, then lower priority loads (less critical) are automatically removed from the system. This technique allows the owner to use the full capacity of the standby system and not pre-invest in capacity that is stranded when certain systems are not operating. It is very common in mission critical facilities with large standby power systems that use programmable logic controllers (PLC) to open/close electronically operated breakers. On smaller systems this can be accomplished by utilizing a dedicated controller or the BAS to open breakers, transfer automatic transfer switches (ATS), or turn on/off motors based on the load.
The incoming service and standby power systems are the largest and probably the most common portions of the system for a modular design; however, the design should include all levels of the distribution system. Switchboards, panels, feeders, and even branch circuits can be modular, designed to allow for stepped growth as required to meet the increased load demand. As with larger systems, in these cases the modular design can be additional equipment or space/capacity left in the existing equipment. Switchboards and switchgear can be designed with space for additional breakers or for additional sections. Panels can be provided with 25% space in load capacity and/or breaker poles.
The following are additional ideas that can be implemented into the electrical system to allow for future growth:
- Use modular generators with integrated paralleling controls. This allows for the addition of generators to a common output without the expensive traditional paralleling system.
- Tie the fire alarm system into the standby system load controls. This allows for the use of stranded capacity dedicated for the fire pump (only used during a fire) to serve new/low priority loads. When the fire pump is activated, shunt-trip the low-priority loads to allow the fire pump to run.
- Install transformers that have the ability to add fans to increase their rating.
- Install a disconnect switch within long runs of plug-in or track-type busway, allowing for an increase in load density. One 400 A busway can be divided into two sections of 400 A busways.
- Spare breakers or spaces for additional breakers to allow for additional loads.
- Oversize conduits to allow for larger cables in the future.
- Provide spare conduits to allow for additional cables.
- Provide pathways and chases to allow routing of addition of conduits and cables.
- Limit the initial loading of branch circuits to allow for additional loads (lights, receptacles, etc.) to be installed in the future without having to run new homeruns back to the panel.
Two key components that are often overlooked in the design are ease and isolation of the expansion. The modular approach is not optimized if you have to shut down the facility or tear apart the electrical system to make it work. In large systems like data centers, this might not seems that complicated because you can provide completely separate distribution systems in each step. At some point, however, all those systems must be tied together, so you should make sure the design includes points of disconnect that will allow for the installation and commissioning of those separate systems without interruption to the existing facility. In smaller systems this becomes more complicated, yet the principle remains the same. Provide points of disconnect and means within the electrical system that allow you to expand without requiring major renovations.
The bottom line is there is no crystal ball or magic formula to predict the future load growth of a building. Prediction, by definition, is to foretell with precision of calculation, knowledge, or shrewd inference from facts or experience. Engineers and owner/administrators must rely on the growth factors, their knowledge of the facility, and their experience to determine a range of potential future growth (minimum and maximum) and then use that range to develop a modular plan that will allow the building to reach the projected future load in economical steps.
Kutsmeda is an engineering design principal at KlingStubbins in Philadelphia. For more than 18 years, he has been responsible for engineering, designing, and commissioning power distribution systems. His project experience includes 7x24 mission critical facilities, emergency operations, and call centers; highly specialized research and development buildings; and large-scale technology projects.