SWEP B185

Too small shows first as a high CO2 leaving temperature. The gas leaves the exchanger well above the entering water temperature, the high-side pressure climbs to compensate, and compressor power rises with it. On a heat recovery duty the water never reaches setpoint at full flow.

Too large costs money up front and can cost performance at part load. Spread across too many channels, the water side runs slowly, which lets sediment settle and fouling start. Send the minimum load along with the design load.

Mount it vertical and pipe it counter-current. CO2 cooled above its critical point changes temperature all the way through the exchanger, so a co-current hookup gives away a large share of the capacity.

The side matters more on this frame than on most. The channels are asymmetric, 0.1106/0.1169 dm3, and the two connection pairs are different sizes for that reason. Piping the CO2 to the water ports puts the refrigerant in the wrong channel set and loses the performance the frame was chosen for.

Support the lines independently and isolate compressor vibration. A 140-plate unit weighs about 72 kg (160 lb) dry, and pulsation fatigues braze joints over time.

Water, glycol mixtures and oils on the secondary side all suit the 316/316L plate and copper braze. That covers the usual gas cooler, heat recovery and heat pump arrangements.

Two fluids do not suit it. Ammonia attacks the copper braze, so a CO2/ammonia cascade exchanger needs a copper-free construction. Deionized or ultrapure water leaches copper for the same reason. Chloride-bearing water is a plate problem instead - 316 pits in it - so heat recovery to pool water or brackish water calls for a different plate material.

Quite a lot, because the catalogued builds of this frame vary in only one field. Every B185 we have on file is an H plate, single pass, SC material and U pressure class, with the same connection layout: 2x42.4&35.1 on one side and 2x48.3&42U on the other. The plate count, from 40 to 140, is what separates them.

So the number after the x is the thing to recover. Failing that, count the plate edges along the side of the pack. Also copy the stamped design pressure if it is legible. U is a category spanning 64 to 140 bar, not a rating, and the replacement has to meet the figure the system was designed around.

In place, by circulating a cleaning solution through the water or glycol side against the normal flow direction, then flushing untill the water runs clear and neutral. A brazed pack cannot be opened. The CO2 side is a closed, clean circuit and is left alone.

Because the braze is copper, the cleaner has to be one that is safe for both copper and stainless. A strong acid that suits bare stainless will attack the braze.

The signs to act on are a rising water-side pressure drop and a CO2 leaving temperature that drifts up at the same load. The published particle limit is 0.8 mm, so a strainer finer than that on the water inlet prevents most of the problem.

When the duty is not high-pressure CO2. The U class is what this frame is for, and on an ordinary water-to-water or R410A duty a standard-pressure frame of similar size does the same work without the cost of the heavier construction.

It is also wrong where the secondary flow excedes the 38.8 m3/h (170.83 GPM) port ceiling, and on any ammonia cascade, where the copper braze rules it out.

One installation point on the duties it does suit: UL approval for CO2 duty wants a relief valve on each side set at 0.9 times design pressure. Plan those into the piping before the unit arrives. Supermarket CO2 systems shows where this frame sits in a rack.