Additional information about standby power for mission critical facilities

When utility power is interrupted, standby power system failure is not an option for mission critical facilities. Presenters from the March 19, 2015, webcast respond to questions left unanswered during the live event.

By Dwayne G. Miller, JBA Consulting Engineers, and Kenneth Kutsmeda, Jacobs March 25, 2015

Requirements of mission critical standby power systems transcend those of standard commercial projects, which are typically designed to merely comply with applicable building codes. Many times, high expectations of systems for mission critical facilities are influenced by the desire for increased levels of reliability and efficiency.

Dwayne G. Miller, PE, RCDD, AEE CPQ, CEO, JBA Consulting Engineers, Hong Kong, and Kenneth Kutsmeda, PE, LEED AP, Engineering Design Principal, Jacobs, Philadelphia, respond to questions that did not get answered in the live webcast event.

Question: What does ATS mean?

Kenneth Kutsmeda: Automatic transfer switch.

Question: Can a bank of capacitors be used to correct power factor (PF) problems in generators?

Dwayne Miller: This would be an unusual way to address a power factor issue particularly with a generator or small generation system. If the generator has a PF problem, the generator manufacturer needs to address it. For a large generation system with load induced power factor challenges, this might be a solution, however, I’d be very reluctant to attempt it, particularly with a variable load. If you have a steady state fixed load, you could use fixed capacitor banks to correct; however, I would recommend a system model be developed to understand how the capacitors will affect system impedance and system behavior during a fault. I typically avoid switched/automated capacitors due to transients associated with them. In the context of mission critical systems and standby power systems, adding capacitor banks could be a risky proposition, particularly with harmonic rich load (servers, UPS units, lots of VFDs, etc). In addition, power factor for mission critical facilities typically will be high (north of 95%) as a good portion of the loads are close to unity PF so capacitors for power factor correction are probably unnecessary.

Question: How does mission critical prime meet the unlimited run time stipulation set by Uptime Institute?

Kenneth Kutsmeda: For the record, Uptime Institute is a guideline and not a code requirement. Uptime Institute does state that "engine generators for Tier III and IV sites shall not have a limitation on consecutive hours of operations when loaded to N demand." ISO standby rated engine generators have a limitation of the total number of hours and therefore are not Uptime Institute certified. ISO prime rated engine generators do not have that limitation on runtime and as long as they are de-rated to N demand can be Uptime Institute certified. The mission critical or data center rating was developed primarily for marketing purposes. Different manufacturers have different limitations on their mission critical rating. In most case the manufacturer will have to review the loads and other aspects of the application to determine if the generator can run without limitation on hours. In the certification process, Uptime Institute will generally accept the generator if the manufacturer documents that there is no limitation on runtime. However, it is recommended you verify this with Uptime Institute before purchasing the generator.

Question: I have a facility that when we transfer to SEP (open transition), the input breaker for the online uninterruptible power supply (UPS) trips. The ATS vendor blames the UPS system and the UPS vendor blames the ATS system. Any suggestions?

Dwayne Miller: I will assume SEP stands for site emergency power. If you are transferring open transition, in rush from the UPS might be an issue. I’d also want to understand the transfer settings on the ATS. In my experience, the quickest way to get to the bottom of these sorts of issues is to measure and capture the event resulting in the trip. If you capture the 3-phase waveform that results in the trip, you will be in a much better position to determine what the problem is. I’m not sure what size your UPS is, but the input breaker would be another thing to review closely to see if you have adjustable settings. I would start at the UPS and work my way toward the source/generator to better understand how the generator, transfer switch, and UPS are acting as a system.

Question: If a legally required system’s generator is placed in an outdoor walled area, can other types of equipment (such as chillers) be put in the same area?

Dwayne Miller: NFPA 110 addresses exterior locations in 7.2.2. The short answer is yes. National Electrical Code (NEC)-required clearance needs to be addressed as a minimum. In addition, the authority having jurisdiction (AHJ) for the specific location should be consulted as there are often local codes and ordinances that will create a more stringent requirement. The big issues for generators in enclosed areas are adequate clearance and clear path for air intake, radiator discharge, and exhaust. Walled areas can create problems with generator overheating and tripping off line due to over temp. Air-cooled chillers or other equipment that raise the temperature, equipment/obstructions that restricts airflow, the prevailing wind conditions, etc. all need to be reviewed to ensure the generator can run and operate properly.

Question: What is your recommendations for MV (15 kV) paralleled generators grounding for better reliability: Low-resistance grounding, high-resistance grounding, or hew hybrid (high- and low-) resistance grounding?

Kenneth Kutsmeda: Recently we have started to recommend the hybrid system; however, it is more expensive and requires more due diligence when setting up the system. In generally we would recommend low-resistance grounding systems for large redundant medium-voltage systems where limiting damage to the equipment during a ground fault is important but it is acceptable to take the system offline for a ground fault.

Question: What are the requirements for the fuel systems for serving NEC 700 and fire pumps loads?

Kenneth Kutsmeda: NEC Article 700 requires the alternate source of power have a minimum of 2 hours (full demand operation) of on-site fuel storage. If power is needed for the fuel transfer pumps to deliver fuel to the generator day tank, NEC 700 requires that the fuel pump be connected to the emergency power system. NEC 700 also states that the generator can’t be solely dependent on a public utility gas system for their fuel supply unless acceptable by the AHJ where there is a low probability of simultaneous failures of power and off site fuel delivery.

Question: I didn’t hear much discussion on the NFPA 70E arc flash requirements and how it affects designs and considerations. Please expand.

Dwayne Miller: NFPA 70E clearly has maintenance implications. In a mission critical facility, concurrent maintainability; in effect, de-energize for maintenance purposes while still serving 100% of the load can alleviate some of the issues. The owner and end user must be advised of NFPA 70E implications to ensure they understand maintenance impacts for decisions they make for the electrical distribution. Personal protective equipment (PPE), boundary issues, labeling of equipment, and similar issues have to be addressed in the context of operations and maintenance of the facility.

Kenneth Kutsmeda: Having redundant components that allow you to shut a component to work on it without disruption will alleviate some concerns. Other things to consider is separating the low-voltage generator control form the switchgear breakers sections. This can be especially beneficial when using medium voltage (MV) because it moves the operator away from the front of the MV gear while breakers are opening and closing.

Question: What are some common modes of failure (CMF) as opposed to single points of failure?

Dwayne Miller: Single point of failure is one device or component that takes the entire system down. Another way look at it is a piece of equipment or component in which all of the current associated with the building load must pass through. A building with a single transfer switch for the entire emergency system has a single point of failure for the emergency system due to the single transfer switch. In effect, if the transfer switch fails, the emergency and normal power is not available to the load side of the transfer switch. A building with a single utility transformer has a single point of failure for the normal power system in the form of the transformer. If the transformer fails, there is no other means to get normal/utility power to the facility.

Common mode failure point is defined as the simultaneous failure of several components due to a single external cause. One way to look at it is a part of the system that does not have current associated with the load going through it, but its failure causes several elements of a system to fail. Examples would be failure of a paralleling switchgear (PSG) programmable logic control (PLC) rendering the PSG system inoperable. Another is a dc power system (station batteries) failure resulting in none of the power circuit breakers in the PSG being able to operate. A third could be a temperature sensor that turns on air conditioning to keep a UPS from tripping out on over temperature. The sensor fails and, as a result, the UPS trips out on over temp. The air conditioning failing would also be common mode failure.

Examples of systems to be concerned with for both single point of failure and CMF: fuel systems relying on bulk storage and pumping to multiple day tanks, remote radiator arrangements, PLC controls, single generator bus, single bus for generator source, utility source and loads, etc.

Question: Are there concerns with locating the paralleling equipment in the same room as the generators?

Dwayne Miller: NFPA 110 Chapter 7 needs to be reviewed for specific application. In addition, in my experience AHJs will have a lot of input and there are often local ordinances around this specific issue. Code aside, I always push to separate the generators and the paralleling switchgear. I prefer interior conditioned dedicated space for the paralleling switchgear separate from the space for the generators. The switchgear space is typically cleaner than the generator area. Having it inside also facilitates maintenance or troubleshooting activities that will often be required during inclement weather. In addition, troubleshooting controls, programming PLCs, and similar activities with engine noise present is challenging. From my perspective this is more of an operational/practical concern. Additionally, with all of the electronics driving complex paralleling systems, a conditioned space also has positive implications for both availability and reliability of the system.

Question: This is more of a comment: NEC 517 life safety and critical are both governed by 700 and are emergency systems by definition.

Dwayne Miller: Thanks for the comment. My comparison was more general in nature to relate 700, 701, and 702 loads to the three branches in a hospital. I agree that critical branch loads are much more directly critical to human life than 701 loads, particularly in intensive care units, surgery suites, patient monitoring, and, in the context of article 517, are treated very similar to NEC 700 loads (NEC 517.31, 10-second requirement, and NEC 517.30.C.1 wiring requirements such as independent conduit, raceways, etc.). My general statement was rooted in NEC 2014, 517.26 which specifically states "the life safety systems of the essential electrical systems shall meet the requirements of Article 700, except as amended by Article 517." Again, I agree the critical branch is more akin to article 700 loads versus 701 loads.