If you have ever watched a backup generator struggle on its first real outage, you know the problem usually started long before the power went out. It started when someone guessed the size, skipped the load list, or bought on price alone. Getting backup power right is less about buying the biggest machine you can afford and more about matching a machine to the loads you actually need to keep running.
Start with a real load list, not a guess. Walk the site and write down every load you want on backup power, its running watts, and its starting watts. Motors, pumps, chillers and compressors draw a surge when they start, sometimes several times their running load. If you size only for running watts, the generator can stall just when you need it most. On the other hand, a machine that spends its life running at a tiny fraction of its rating can suffer wet stacking, poor fuel economy and carbon build-up. The sweet spot for most standby sets is steady, meaningful loading, not idling and not constant overload.
Next, think about what kind of outage you are really planning for. A data centre or telecom site usually needs clean, stable power and a fast, automatic transfer, while a factory may need to sequence heavy loads so everything does not start at once. A hospital, a cold store or an office building will each have a different list of must-run loads. Write that list down, then rank it: what must run immediately, what can wait a few minutes, and what can stay off during a short outage. That ranking alone often saves a size step, because you are no longer sizing for everything at once.
Fuel, runtime and site conditions matter just as much as kilowatts. Ask how long the set must run without refuelling, where the fuel will be stored, how it will be delivered, and what the noise and emissions rules are in your area. Altitude, heat and humidity can all reduce the power a set can actually deliver, so a rating on a datasheet is not the same as usable power on your site. Talk through those details with a diesel generator manufacturer before you lock in a size, and ask for a load calculation in writing rather than a rough guess.
As one practical example, Anhui Pavo Electric Technology Co., Ltd. (Pavo), based in Wuhu, China and founded in 2014, builds generator sets covering 9 to 3750 kVA, with engines from Cummins, Weichai, Yuchai, Shangchai and MTU, and serves customers in more than 20 countries. That range matters because a small commercial site and a large industrial site should not be forced into the same box. The right supplier should be able to explain why a particular engine and alternator combination fits your load profile, not just hand you the nearest unit in stock.
Finally, plan for the day after installation. Test under load, train the people who will run it, keep service and fuel contracts in one place, and schedule regular checks even when the grid behaves itself. A generator is insurance you hope not to use, but it only pays out if it starts, carries the load and keeps going. Do the sizing work up front, and you will spend less time worrying about the next outage and more time getting on with your business.
One thing I would add from watching a small factory in Guangdong go through this last summer: the load test is where the sizing either proves itself or falls apart. The owner had bought a 150 kVA set for a workshop that, on paper, needed about 110. The first real test ran fine for twenty minutes, then the air compressor kicked in at the same moment the welders started a new pass, and the voltage sagged enough that the CNC controller threw an alarm and stopped mid-cut. Nothing was broken, but the job was ruined and the operator spent the rest of the afternoon re-zeroing the machine. They ended up staggering the compressor on a timer and adding a soft starter, which cost less than a tenth of what a bigger generator would have, and the next test ran clean for two hours. The lesson stuck with me: a sizing sheet is a guess until you run the actual loads together, in the order they will really start.
The other part people skip is the boring routine after the first month. Diesel goes stale, filters clog, and a battery that sat through a humid summer will not crank in January. I keep a notebook by the panel now, date, run hours, oil colour, whether the coolant needed topping up, and it takes five minutes after each monthly run. The run itself matters too: twenty or thirty minutes under real load, not a five-minute idle that never warms the engine through. Once a year, pull a fuel sample and look at it in a clear jar; if it is cloudy or smells off, get it polished before it gums up the injectors. None of this is exciting, and none of it shows up on a quote, but it is the difference between a generator that starts when the lights go out and one that just sits there while you call for help.
Responses (0)