PUE and Cabinet AC
PUE = total facility energy ÷ IT equipment energy. Cabinet AC, lighting and power distribution are all 'overhead'. Dropping PUE from 1.6 to 1.3 roughly halves cooling overhead — significant for edge sites.
Four Savings Levers
| Lever | Saving |
|---|---|
| EC fan vs AC fan | 20%–35% |
| Variable-speed compressor | 15%–25% |
| Setpoint 27℃ vs 22℃ | 10%–15% |
| Indirect free cooling | 30%+ (shoulder) |
Deployment
- Replace fixed-speed units with EC VSD combos.
- Raise setpoint within allowed band to cut runtime.
- Enable free-cooling in cool seasons for big annual savings.
查看 HCK 机柜空调产品 或 联系昊诚工控 获取选型方案。
Sizing, selection and what the datasheet does not say
The selection that survives a hot summer starts from total heat dissipation rather than from the nameplate rating of the equipment. A drive's nameplate is its output; the heat it puts into the cabinet is a few percent of that, and assuming otherwise is the most common way to oversize a system by a factor of ten.
The working method has four steps. First, list every component that generates heat and take its dissipation from the datasheet rather than from its rating. Second, add the envelope gain of the enclosure itself, which for a sealed cabinet in a hot climate is often the largest single term. Third, apply a safety factor once, to cover the uncertainty in the first two steps. Fourth, round to the nearest sensible unit, and check that the result leaves the unit running part of the time rather than flat out.
| Step | What you are solving | Common error |
|---|---|---|
| Component heat | What the equipment puts into the air inside | Using nameplate power instead of dissipation |
| Envelope gain | Heat from outside through the cabinet walls | Forgetting it on small sealed boxes |
| Safety factor | Uncertainty in the estimate | Applying it twice, or not at all |
| Installed capacity | Rounding to a real catalogue unit | Picking a unit that will run at 100 % duty |
Then ask the three questions that turn a calculation into a specification: what ambient was the quoted capacity measured at, what airflow does the unit need, and how much clearance does the condenser require. Specifications that answer all three tend to be honest about their capacity; specifications that answer none of them are describing a product rather than a performance figure.
Airflow, placement and sealing
A unit that is correctly sized can still fail to hold setpoint because of how the air moves inside the cabinet. The evaporator takes in whatever air is beside it. If the warm air leaving the equipment rises into that intake, the refrigerant has no cold surface to evaporate against and the controller sees a cabinet that will not cool, whatever the rating on the box.
The three symptoms to look for are familiar: the setpoint is reached at the bottom of the cabinet but never near the top, the compressor runs continuously while the panel reads close to ambient, and the unit's own intake temperature is close to the ambient.
- Mount high, where the warm air rises
- Fit a baffle so air passes across the full height of the equipment rather than around it
- Direct the discharge across the load, not down into the floor of the cabinet
- Seal the perimeter between the unit and the cabinet wall; an IP-rated cabinet with an unsealed unit is not an IP-rated cabinet
- Keep the intake grille clean and unobstructed; flow resistance at the intake quietly costs capacity
Measure the temperature spread from the bottom to the top of the cabinet after a full working day. A spread above 8 °C points at the airflow rather than the capacity, and it is a cheaper fault to fix than a larger unit.
Maintenance intervals that fit the site
There is no single correct interval. There is only an interval that is written down, and a filter and condenser that can be reached without dismantling the panel. What a unit in a clean air-conditioned plant room needs and what a unit under a machining centre needs are two different schedules.
| Environment | Filter | Condenser | Drain |
|---|---|---|---|
| Clean indoor plant room | Quarterly | Twice a year | Monthly |
| Workshop with dust or swarf | Monthly | Quarterly | Monthly |
| Outdoor coastal | Quarterly | Every two months | Monthly |
| Desert, sand-laden air | Monthly | Every two months | Monthly |
| Wash-down or high humidity | Monthly | Quarterly | Weekly |
Two habits do most of the work. The first is recording the ambient at each visit, which converts next year's argument about whether the unit has degraded into a settled question. The second is never pressure-washing a dry, dusty coil from the front, which drives the dust into the fins instead of shifting it. Clean dry first, rinse gently, and let it dry before the unit runs again.
Fault diagnosis: symptom, likely cause, what to check
The failure that costs the most is the one that always looks like the compressor. In practice the compressor is rarely the first thing to fail. What fails first is airflow, then the drain, then the sensor, and the compressor is usually simply working against conditions that have quietly become impossible.
| Symptom | Most likely cause | Check first |
|---|---|---|
| Setpoint never reached, compressor runs flat out | Recirculation into its own intake | Intake temperature against ambient |
| Bottom of the cabinet cold, top no warmer than ambient | Air bypassing the equipment | Baffle position and discharge direction |
| Capacity fell without an obvious cause | Fouled condenser | Coil surface temperature and dust build-up |
| Trips in humid weather | Blocked condensate drain | Drain path, and whether it is graded to drain |
| Controller reads a wrong or frozen temperature | Sensor fault | Sensor resistance, then the wiring |
| Ices up inside | Low airflow, or a setpoint below the dew point | Filter condition and setpoint |
Log the duty cycle before any of this. A unit that logged 100 % through the previous summer had the answer to the question already written down, and nobody had read it.
What to ask before the order is placed
Five questions rarely change the price and frequently change whether the cabinet still works in year six. Ask them at quotation stage, when they are still questions, rather than after commissioning, when they become arguments.
- At what ambient temperature, and at what supply voltage, is the quoted capacity valid?
- What is the measured airflow in cubic metres per hour, and what clearance does the condenser need to be installed with?
- What is the control accuracy, and what does the unit do if its own sensor fails? A unit that assumes a safe reading turns a five-minute wiring fault into a burned panel.
- What is the compliance documentation, and does the test report actually exist for the unit being quoted — not for a sibling model?
- What is the lead time on a replacement unit, given where the cabinet physically is?
Those five answers, written down and signed off at quotation, are usually worth more to the project than a few percentage points of extra capacity.