CCG Facilities Integration

Innovating Mission Critical Systems

Research & Whitepapers

Use of Modular Design and its Relation to Energy Performance and Construction Costs

01.15.2013

Authored by:
Bruce Edwards, RA
President


Background
“Modular” design is a topic of intense interest in the data center industry.  Products have been developed and placed into the market by numerous manufacturers and equipment integrators.  The range of products referred to as “modular” is evolving rapidly, and a common definition is lacking.  An industry-agreed technical standard is farther from reality.

“Modular” concepts and products have been present in the data center industry for decades, but the current discussion is being driven by the emergence of various new offerings and their attendant marketing efforts.  Two trends are evident: rapid product development and commercialization, and a certain amount of exaggeration with respect to the performance and qualitative merits of the products.

In order to consider the topic, some general background discussion is useful to establishing shared terminology.  Modular elements may be defined as “identical or equivalent components which are interchangeable, and which are combined to produce a larger functioning entity”.  Additional modular elements can be added to the initial installation at subsequent points in time, and with minimal disruption to its functioning, thus providing an expansion capability, or “scalability”. 

Standardization and interchangeability of the modular components leads the way to mass production, off-site fabrication, and stockpiling or pre-positioning of elements prior to on-site need.  These attributes may improve build quality, reduce production cost, and shorten jobsite delivery times. 

Based on this description, conventional field erected facilities already incorporate a great deal of modular content into the construction process, and realize the benefits of doing so. 

Modular Component Products
From a chronological perspective, the off-site prefabrication of mechanical and electrical “modular” components pre-dates that of IT deployment components, and prefabricated modular buildings have been available for a broad range of uses and occupancies for a very long time. 

Prefabricated elements such as skid-mounted pump systems or complicated piping assemblies have been provided to the construction industry for decades.  Larger and more complex assemblies such as pre-fabricated chiller plants that may be split for shipping but mated up once on-site also have a relatively long history on the market.

More recently, prefabricated electrical assemblies have emerged to a greater extent.  Diesel engines, UPS systems, combination diesel engine-rotary UPS systems housed in containers are examples.  Very large multi-unit switchgear arrangements have been prefabricated off-site and skid-mounted for jobsite delivery.

Off-site pre-fabrication of components that are larger or more complex than the previous modules has emerged as the defining characteristic of today’s very public “modular design” or “modular construction” discussion.  It is useful to distinguish initially between IT equipment modular components and mechanical and electrical infrastructure modular components, although there is an increasing level of integration between the two types.

With respect to IT equipment deployed in large modules, a number of manufacturers have begun to offer standard ISO shipping containers modified for the purpose of housing racks of IT equipment.  Most such products incorporate terminal cooling and power equipment to serve the IT equipment within.  In addition to allowing deployment of large numbers of pre-configured, pre-racked IT equipment, these containerized solutions also typically permit a high degree of supply and return air separation by their inherent physical arrangement, enabling potential cooling energy efficiencies. 

Key to the expansion of containerized IT equipment deployment was the emergence of a small number of extremely large users able to exert significant influence on IT equipment manufacturers with respect to the physical characteristics of that equipment.  This influence was due to the very large volume of equipment purchases they controlled.  Related to that is the relatively high degree of centralized planning and control these entities are able to exercise over the selection of IT equipment to be utilized by the enterprise business units, encouraging a great deal of standardization in that regard.

In response to the physical constraints of the standard ISO container dimensions, some manufacturers have begun to gang together multiple containers with their sides removed, or to develop prefabricated IT equipment enclosures not based on the standard ISO containers and with greater dimensional flexibility.  Restrictions on over-the-road loads still limit the size of such products, however.

Energy Use and PUE

Manufacturers of pre-fabricated IT equipment enclosures may tout reduced energy consumption as a key advantage of using their systems, often claiming some sort of inherent advantage over conventionally constructed facilities.  For the purpose of discussion, we will consider here annualized energy consumption in excess of that utilized by the IT equipment itself.  This corresponds to the useful metric of “power utilization effectiveness”, or PUE.  A lower PUE represents better energy utilization performance.

PUE performance depends on many factors, chief among them being the efficiencies of the major mechanical and electrical systems equipment (transformers, UPS systems, chillers, cooling towers, evaporative coolers, fans, humidity control, etc.), the use of cooling economizing strategies (water-side, air-side), operating temperatures and voltages, and the physical configuration of the cooling and power systems, including cooling supply and return air separation.  These factors may be influenced by the project team during design of the facility.  External factors are also present, primarily the local climate and air-quality conditions.

All of these factors exist in both pre-manufactured or containerized solutions and conventionally constructed facilities, and similar opportunities for lowering energy consumption exist in both scenarios, as well as in hybrid solutions.  The increasing recognition of the issue of energy consumption and a greater understanding of the techniques for improving performance are yielding more energy effective facilities of all types.

The following table summarizes an energy consumption analysis of various cooling system approaches in conventionally constructed data centers.  Examination of the Annualized PUE performance for this range of solutions (PUE between 1.16 and 1.23) demonstrates the results that can be achieved.



Cost

As with the subject of energy use, suppliers of pre-fabricated IT enclosures and mechanical and electrical systems components portray these products as lower cost, both in terms of capital and operating cost. 



There are far too many products and combinations thereof to comprehensively review the cost implications of using pre-fabricated “modular” systems here, but a high level analysis will allow for a basic understanding on the subject. 

 

The capital costs of the typical data center facility are allocated approximately 25% to general construction of the building, 25% to mechanical systems, and 50% to electrical systems.  The mechanical and electrical infrastructure options are generally the same for both pre-manufactured “modular” and conventional field-erected solutions at a technical level. These costs account for approximately 75% of total facility construction cost.  Of the remaining 25% of costs attributed to general construction of the building shell, fit-out, and site development, perhaps half might be allocated to the IT equipment areas.  Thus, little opportunity for cost variance occurs between the two approaches, particularly if careful attention is paid to how the systems are deployed in stages over time. 

The small, scalable increments of pre-fabricated IT space may present a cost avoidance opportunity if building shell construction can be deferred, but not necessary a long term reduction when the facility is fully developed if the unit costs for the spatial components are higher.

The facility operating costs of the typical data center are dominated by energy costs and infrastructure equipment maintenance costs.  As noted earlier, for new construction, the mechanical and electrical infrastructure options and their associated equipment and energy needs are similar for both pre-fabricated and field constructed approaches, and present little opportunity for differentiation. 

With respect to total cost of ownership (TCO), factors of capital and operating costs are considered, as well as any tax implications (as relates to depreciation rates for differing types of construction classifications) and replacement schedule (useful life of pre-fabricated, moveable elements as compared to conventionally built, fixed elements).

Summary

The current situation is this:  The data center designer and user now have a very wide range of “modular” options available with which to accomplish project goals.  These options range on a continuum from a myriad of small scale elements employed in conventional field erected fashion to nearly complete off-site prefabrication of an entire data center by a single source with the discrete modular elements to be assembled on a prepared site. 

Where the optimal design solution falls on this continuum is a matter to be determined based on the specific project requirements, including building footprint, power and cooling load magnitudes, IT mission and hardware growth rates, future mission plans, the nature of supporting personnel functions, capital and operating budget constraints, climate and physical site conditions, and risk tolerance, among others.

To download, please click CCG Whitepaper_ Use of Modular Design and its Relation to Energy Performance and Construction Costs-2013-07-12.pdf.