Your article in the April issue, “Top Nine Reasons Generators Fail to Start,” had quite a bit of excellent information on gen-set maintenance. Unfortunately, the title and subheads were a bit misleading.
Standby generators for mission-critical applications have maintenance and failure issues that can be broadly split into two categories based upon load impact. The second, and generally less serious (in terms of number of load losses) category is long-term run failure potential. This category is similar to “prime” or construction rental applications, which generator service technicians at rental yards (such as the author's) face on a regular basis. This article dealt largely with these issues.
The first and far more serious category of standby gen-set availability issues is the short-term run failure group, which includes fail-to-start. This group is most critical only because most utility or site-generated power failures are very short term, meaning seconds, minutes, or a few hours. Causes can be shorts from rodents or tree branches, circuit breaker trips (utility or user site), lightning hits, etc. IEEE ITEC, EPRI DPQ, and other sources give a good summation of time and frequency of most utility disturbances and confirm that the overwhelming majority of these (97% to 99%) are substantially less than one hour in duration, typically less than 1 minute. This is not enough time to uncover the types of cooling system problems discussed in the article. Certainly cooling issues, and even some of the fuel issues discussed, are important, but have little to do with “failure to start” of standby systems, which is what the article reportedly was focused on.
In my experience, there are very few reasons standby gen-sets will typically fail to start. All are so simple to avoid, making a reasonable person wonder why they are listed. Nonetheless, here they are. The single largest cause is, as the author accurately stated, battery failure. Well-designed and maintained mission-critical systems normally have redundant batteries, air-start, or at least constant monitoring, and early change-out of batteries, but not rental sets. We recently managed a fast-track upgrade for a Tier IV site that required two separate 500kW gen-sets to power bridge (15 second) UPSs for temporary power. These gen-sets were the best the local dealer had, but happened to have start batteries that were 2½ years old. Lead acid batteries in a typical generator storage yard are not air-conditioned and therefore likely approaching end of life somewhere in two to three years. Because this was summer — and temperatures could exceed 100°F — we changed both to new and commissioned the systems thoroughly. We felt this was a small price for significant insurance.
Beyond battery and charger issues (charge turned off and no one listening to the alarm!), the other major causes of fail-to-start in critical standby systems are controls/ATS not in auto mode, and errant breaker trips, assuming normal commissioning/testing was done. All of these can be addressed with basic maintenance, site security, and MOP/scripting for critical work actions common in higher criticality sites.
Good standby gen-set maintenance will include all the things the author mentioned for reliable long-term running, but avoiding “fail to start” is easier than you may think.
— Dennis DeCoster, Mission Critical
West, UPS & Power Protection
Technologies, Redondo Beach, Calif.
Author's reply: I can see where the title seemed a little misleading. Just to clarify, we do not just run a rental yard. We service more than 1,300 generators in the San Francisco Bay Area and Silicon Valley. This article started off as my own little research project to identify the real reasons our customers' generators don't start.
Looking at the data for our rental fleet was entirely different. This data came from permanently installed generator sites only. Most sites were data center facilities of some kind. I researched our real “calls for service,” which had the description “my generator won't start.” In reviewing two years' worth of calls, I found most were not really “fail-to-start” conditions. Most were perceived as a condition that would cause the generator not to start. The key point I am trying to make is maintenance is cheap insurance, and training the onsite staff is key to operating an emergency power system.
However, I disagree with your comment about most power outages are not long enough to uncover some of the fuel or cooling system issues. As soon as a power outage of any duration causes the generator to come online, it will be online for a specified period of time. The automatic transfer switch will usually have a 30-minute delay timer that is set up to retransfer to the utility feed. This timer doesn't start until the normal utility feed is back online. What we typically find is a leak in the cooling system — that was not previously picked up during regular inspections — reveals itself once the gen-set is called on to start. The engine will start but then immediately shut down once the low coolant level alarm triggers.
— Darren Dembski,
Peterson Power Systems
In the April article, the author sort of tap danced around the “Controls not in Auto” subject (item number 5). One thing that most diesel (i.e., prime mover) gen-sets have is a rack trip (manual) feature. If this is tripped (sometimes to test in emergency shutdown situations), then the engine is not going to start. This might not show up in electrical safety due to it being a fail safe due to diesel runaway. I've been caught a couple of times with this in the marine engineering field.
— Steven A Mauch, electronic industrial control mechanic/RFE, USDA Forest Service, Forest Products Laboratory, Madison, Wis.
Author's reply: Steven, you're correct. Although we don't see many mechanical trips on standby generators, they are out there. All of the old 2-cycle Detroit Diesels have the air door on top of the blower, which needs to be manually reset in the event of an emergency shut down. I've actually seen an old Cat D399 accidentally trip one of two air doors while online. The engine kept running, but was drastically low on power.
— Darren Dembski,
Peterson Power Systems