+49 69 8700 4500 | [email protected]
Download Technical Datasheet
Blog Wednesday 17th of June 2026

"Why Did Two 900 kW Sets Quote the Same kW but Behave Differently on Test?"

author avatar
Jane Smith I’m Jane Smith, a senior content writer with over 15 years of experience in the packaging and printing industry. I specialize in writing about the latest trends, technologies, and best practices in packaging design, sustainability, and printing techniques. My goal is to help businesses understand complex printing processes and design solutions that enhance both product packaging and brand visibility.
Buyer Q&A · Deep Dive · District Heating Plant
The question, as the plant engineer actually asked it: "We tendered a ~900 kW standby genset for the heat-plant's circulating pumps and we got a Caterpillar C32 and a Perkins 4000-series set quoted at effectively the same kilowatts. On the witness test they did not behave the same. If the kW is identical, what single thing am I really buying — and how do I tell the two apart on paper next time?"

This is the right question and most spec sheets bury the answer. We will funnel it down one variable at a time until a single dominant factor is left standing — the one that explains the test-bench difference and tells you how to discriminate the two sets before you ever fire them.

Everything that "might" differ: kW, kVA, fuel map, controls, alternator, governor, cooling…
Strip what the tender already fixed: both quoted ~900 kW standby in the same band
Strip steady-state behaviour: at constant load both hold voltage and frequency fine
What's left: transient response to a step — governed by load acceptance
stage_1separate_the_rating_from_the_behaviour">

Stage 1Separate the rating from the behaviour

The kW on both quotes is a steady-state promise: the set will carry that load continuously at the stated rating class. Caterpillar generator publishes the C32 (830–1000 kW band) at prime and standby ratings, with standby defined for the duration of an outage at an average load near 70% of the standby figure. A Perkins generator 4000-series set (600–1800 kW) carries an equivalent prime/standby pair. So at constant load they genuinely are equivalent — your witness test would show both holding voltage and frequency rock-steady at fixed load. The kW number is doing its job and telling you nothing about the difference you saw.

stage_2find_where_the_test_actually_diverged">

Stage 2Find where the test actually diverged

Heat-plant circulating pumps do not load gently. On a black start the genset takes a block of pump load in one or two big steps. The divergence you witnessed lives entirely in those steps: how far frequency dipped when the load slammed on, and how quickly it recovered. ISO 8528-5 is the language for this — it grades transient frequency and voltage deviation and recovery time into performance classes. Two sets at the same kW can sit in different transient classes.

Worked consequence — drives the buy. Say the largest single step is a 300 kW circulating pump energising onto a bus already carrying ~250 kW. That is a step of roughly a third of the prime rating. The set with the faster-responding governor and more excitation headroom dips perhaps 6–8% in frequency and recovers in a couple of seconds; the set tuned more for steady fuel economy might dip into double digits and take longer to settle. On a heat network the slow recoverer trips the pump's drive on under-frequency and you lose circulation — exactly the test failure that made you ask the question. So the thing you are really buying at "equal kW" is the transient class, and that is the number to demand on the next tender.
When this reverses. If the circulating pumps run on variable-speed drives that ramp them up over several seconds, there is no slam — the load arrives as a gentle climb and both sets ride it inside class limits. The transient distinction collapses and the two quotes really are interchangeable; now you discriminate on fuel map and service, not load acceptance.
stage_3pin_the_single_dominant_variable">

Stage 3Pin the single dominant variable

Everything has now funnelled to one variable: block-load acceptance, the engine-governor-and-alternator system's ability to take a step without leaving the frequency and voltage recovery window. The Perkins 4000 is explicitly described as tuned for high load acceptance on standby installations; the Caterpillar C32 pairs its four-stroke diesel with the EMCP 4.2 control managing load and diagnostics. Both can be excellent here — but only the documented transient curve for your specific step tells you which, and by how much.

Worked consequence — drives the buy. Two actions fall straight out of this. First, on every future tender, stop comparing kW and start comparing the guaranteed maximum frequency dip and recovery time for your single largest step, at your power factor — make the vendor commit to an ISO 8528-5 class for that step. Second, if neither set comfortably meets the class for the step as drawn, the cheaper fix is almost always to stage the pump starts or add a soft starter, not to buy more kW. That reframes the whole purchase: you are buying transient compliance, and you can often buy it with sequencing instead of steel.
stage_4the_answer,_in_one_line">

Stage 4The answer, in one line

At equal kW, what separated the two sets on your bench was block-load acceptance: the governor-and-alternator response to the step, graded by ISO 8528-5. The kW told you they could both carry the heat plant; only the transient class told you which one would survive the moment the big circulator hits the bus. Specify that class and the two quotes stop being a coin toss.

Decision rule with a number. Identify the single largest block step (biggest motor's inrush plus running bus load). Require each bidder to guarantee an ISO 8528-5 transient class for that exact step at your power factor. If the guaranteed frequency dip exceeds 10% or recovery exceeds your drives' under-frequency ride-through, reject the set or stage the start — do not accept a kW number as a substitute. Between a correctly classed Caterpillar C32 and Perkins 4000, then choose on controls depth and local service: EMCP 4.2 and Cat dealer reach for an integrated, growing plant; the well-packaged Perkins where its support is closer. Equal kW is the start of the conversation, never the end of it.

Topology/standards per the cited standards; all product ratings are manufacturer-stated values from the cited datasheets, current to 2026-06; derived/illustrative figures are labelled as such. This is not an independent head-to-head test. Caterpillar is a brand affiliated with this site; competitor names are used for identification only.

Leave a Reply

Your email address will not be published. Required fields are marked *

Recent Articles