Sustainable building is essential in today’s “carbon neutral” world. To find out more about responsible design, Mission Critical sat down with Mark Heinrich, senior director of engineering at CPG. Drawing on 25 years of experience in the engineering, design, and construction, he discusses some critical considerations.
1. How has traditional data center design and construction evolved to accommodate the pre- and post- pandemic shift to the cloud and ongoing digital transformation efforts?
Traditional data center design and construction started to evolve prior to the pandemic as demand for businesses, government entities and individuals to move from local data centers to the cloud grew exponentially. This shift accelerated the speed of design and construction in the data center market over the past 10 years, with a huge increase in the last 5-10 years and an even larger increase once the pandemic hit.
Around 10-15 years ago companies started to realize that they could stop building, managing, maintaining, cooling and paying for electricity within data center spaces that reside in their own facilities, and, instead, shift these services to a colocation or hyperscale facility that someone else could build, manage and maintain for them. This shift was found to work economically and allowed the companies to concentrate their resources on other aspects of the business. Others saw the success in this shift and started doing the same thing, creating an extreme demand for more data center space.
Once the pandemic hit, employees started working from home, consumers increasingly ordered goods online and at-home entertainment became more of a necessity, increasing the demand for even more cloud computing. The post-pandemic world has created a data center building boom. The immense demand has been difficult for the data center market to keep up with, and has required that data centers be built and delivered as quickly as possible. In order for the design and construction processes to keep up with the volume, new processes are required to keep pace and to improve the time-to-market of data center project builds while keeping costs low.
2. What strategies are helping data center developers improve the time-to-market and costs of their new data center projects?
Two successful ways to decrease the time-to-market and costs for a data center are to use the design-build construction model, versus a traditional design-bid-build model, and to incorporate a modular and prefabrication design approach.
The design-build model puts the design and construction of the project under a single umbrella. This allows the design and construction teams to work closely together, along with the client, from the very beginning to the very end of the project in order to streamline the design while watching constructability and costs. It also eliminates the time between the design being complete, the bid process, and the contractors being brought on board. This allows the design-build team to design and procure long lead equipment (LLE) early in the process so the equipment can get to the site as quickly as possible to decrease the overall schedule. This is increasingly important with the supply chain issues that are impacting manufacturers and commodities. Designing and building the project faster results in less project cost.
To simplify the design and decrease the time required for the build, companies can also use a modular and prefabrication design and installation approach to projects. With this approach, modular mechanical and electrical equipment are prefabricated in a factory environment instead of on-site. Labor rates within a factory environment are often lower than field labor rates and challenges in the field, such as weather, can be minimized. This equipment can either be built before a project starts or be built while project site activities, such as groundwork and concrete slab pours, are being completed, thus saving time on the overall build and working ahead to help with supply chain issues. The modular equipment is built, inspected, and tested before it is shipped to the job site where it is essentially plug-and-play equipment. This modular and prefabrication design approach also results in a repeatable and scalable data center project.
3. How does using this design approach result in a more repeatable and scalable data center?
A data center build will vary due to the project location, with every location having different considerations based on weather design conditions, site configurations, soil, code requirements along with many other factors. The goal is to design data center projects where the build varies as little as possible no matter the location and design conditions. Modular equipment, such as modular electrical rooms (MERs) or modular mechanical air conditioning units, can be designed to minimize the design modifications that are required due to the project location and design conditions. While the building may have to be configure differently due to a site configuration, that same modular equipment can be used at various project sites.
The MERs house an entire electrical lineup including the switchgear, UPS and battery equipment along with the associated cooling and controls equipment and can be located almost anywhere with minimal variations based on geographic location. Modular mechanical conditioning units are designed to be flexible in the type of cooling that is utilized. The modular equipment is also designed to be scalable where they are set at certain capacities that are applicable to most data center projects.
A repeatable and scalable data center design can decrease design time, construction time and overall project cost. Using this approach is one of the many critical factors that goes into the design of a data center project and the selection of the electrical and mechanical systems.
4. What are the critical factors that go into the selection of the electrical and mechanical systems for a data center project?
Time-to-market and cost are the two most critical factors when selecting the electrical and mechanical systems and their associated equipment. Specifying standardized equipment is key in the design approach as it is much more available in the supply chain compared to a customized counterpart. Designing the equipment so multiple manufacturers can meet all the specification requirements allows for better cost and lead times through both competition and availability/capability.
Reliability and redundancy are also critical factors that go into the selection of electrical and mechanical systems for data center projects. Typical design practice is that if a piece of equipment fails then the data center must continue to operate, so the design must implement redundant equipment and systems that meet the clients’ needs. Electrical and mechanical equipment should also be “concurrently maintainable” meaning if a piece of equipment is taken offline for service, the remaining equipment should have ample capacity to maintain the critical load of the data center.
The outside weather design conditions within a geographic location, including but not limited to dry bulb temperature and humidity levels, along with the required design conditions inside of the data center dictate the type of mechanical systems that can be utilized. Once the design conditions are established, the efficiency and cost of each mechanical cooling system option can be examined to determine the best mechanical system for the data center.
Most data center critical equipment these days can operate within higher temperatures and higher humidity levels than they could in the past. Data centers no longer have to be 65-75⁰F like they did 20 years ago but can be up to 95-100⁰F in many cases. This allows the option to use higher efficiency mechanical cooling systems such as evaporative cooling if a location has lower humidity and ambient temperature design conditions. This higher efficiency results in a lower data center PUE, which is also a critical factor that goes into the selection of electrical and mechanical systems.
5. What is data center PUE and why is it important?
PUE stands for Power Usage Effectiveness which is basically the efficiency of a data center. It is the amount of power that is put into a data center divided by the amount of critical power that is delivered to the data center critical equipment (the critical equipment typically being the servers located within the equipment racks). The lower the PUE, the more efficient the data center, the lower the operating cost and the lower the impact is on the environment. Some items that go into a PUE calculation are power required for mechanical cooling, lighting and electrical equipment inefficiencies. The more efficient these items are, the lower the PUE of the data center.
6. Are there certain smart infrastructure investments that would help ensure a more reliable, efficient data center?
There are two main data center systems for smart infrastructure: Electrical Power Monitoring System (EPMS) and Building Automation System (BAS). The EPMS system, as the name implies, monitors the electrical power throughout the data center. The more electrical systems that are monitored, the more the overall health of the electric system can be monitored. An EPMS can allow a data center operator to identify a problem and determine where the problem exists to streamline repair. It can also help to determine if the data center is properly utilizing all the electrical power being provided to the project, and thus determine if there is stranded capacity that can be used elsewhere.
The BAS system is the automatic central control of the data center mechanical systems, along with other systems like lighting and security. This control allows for a reduction in energy consumption, a reduction in operating and maintenance costs, and improved life cycle of equipment. Both the EPMS and BAS systems can be monitored remotely and can store historical performance data for future analysis.
As part of the design-build process it is beneficial to have an EPMS and BAS design team and integration team on board to produce fully comprehensive EPMS and BAS design documents versus the traditional “performance specifications” that delegate the design to the BAS installation contractor. The design team and integration team work together in the design-build model along with the BAS installation contractor to facilitate the design and construction of the data center project and increase time-to-market.
7. Based on your responses, it’s clear that the modular, design-build approach is the natural step in the evolution of mission-critical facility construction projects. What would you say is the most critical factor in the design and build of a data center today?
While most of the responses here are mainly focused on improving time-to-market, cost, efficiency and operation of the data center, the most critical factor in the design, build and operation of a data center is safety. The design of a data center must consider how the equipment will be installed, operated and maintained in a safe manner. The design-build model naturally allows the design team, construction team and client (operations team) to work together throughout the entire design and construction process, with the expertise of the installation and operation, so that safety is maintained. Safety issues can be identified early and addressed during the design process more effectively using the design-build model.