Industry Solutions / Industrial and Manufacturing

Process UPS, Infrastructure UPS and Switchroom Cooling for Australian Plants

Vendors blur two completely different problems into one word. When a plant asks for "a UPS", what usually arrives is a rack-mounted unit sized on the IT cabinet in the office, installed in the comms room, and quietly expected to also keep the PLCs alive. It will not, and the day it matters you find out in the most expensive way available: mid batch, mid shift, with product in the pipe.

The distinction that matters on an industrial site is between process power and infrastructure power. Process power keeps the plant controllable. Infrastructure power keeps the plant visible and connected. They fail differently, they cost differently when they fail, and they should be specified, sized and maintained as two separate systems. The second thing vendors skip entirely is the switchroom next door, where a wall of variable speed drives is pushing continuous heat into a room that was often designed with a wall fan and an optimistic assumption.

Indigi Power and Cooling is an Indigenous and Veteran owned critical power and cooling specialist, registered with Supply Nation and ICN Gateway. We work across both halves of the problem because on an industrial site they are physically adjacent. The UPS that protects your control system usually lives in the same building as the drives that are heating it up, and the reliability of one depends on the ambient conditions created by the other.

Two separate UPS problems in every plant

Start by classifying the load, not by picking a product. Once the loads are classified honestly, the specification largely writes itself. The table below is the classification we use on site surveys.

Process UPS

A process UPS is small in kVA and enormous in consequence. A typical PLC panel draws a few hundred watts, sometimes only tens of watts at the 24 V DC rail. What it needs is not capacity, it is quality and certainty: a stable sine wave, tolerance of the poor supply conditions found downstream of a plant transformer, and a runtime long enough to either ride through the disturbance or drive an orderly, controlled shutdown of the process. On a continuous process, that runtime is not about waiting out the outage. It is about giving the control system enough time to bring the plant to a safe, restartable state rather than a random one.

Process UPS units also live in bad places. Control panels sit on the plant floor, not in a comms room. Ambient temperature inside a panel routinely sits well above the 20 to 25 degrees Celsius at which valve regulated lead acid battery life is quoted, and battery life falls sharply as temperature rises. That is why panel-mounted UPS batteries fail years earlier than the datasheet suggests, and why we treat battery location and panel ventilation as part of the UPS specification rather than an afterthought. Where the environment is genuinely hot we specify equipment rated for it, covered on our high temperature and industrial outdoor UPS page.

Infrastructure UPS

The infrastructure UPS is a more conventional problem and a well understood one. It carries servers, switches, firewalls, CCTV and access control, it lives in a comms room or a small server room, and it is sized on a steady load with a known peak. The runtime target is usually whatever it takes to reach generator changeover, or to complete a graceful shutdown of the virtual estate. It is the same engineering as any commercial site, and it should be specified that way.

The point is not that one is harder than the other. The point is that they are different, and a single UPS cannot satisfy both. Size one box for both loads and you get an oversized unit running at a low load factor, where double conversion efficiency falls away, sitting in the wrong building, with a runtime that is a compromise between two requirements that were never the same requirement.

Why one shared UPS is an expensive mistake

Four consequences follow from sharing, and we see all four regularly.

• The wrong location. A shared UPS is almost always installed where the IT is, which means the cable run to the control panels is long, and the thing keeping your process alive is now dependent on a distribution route nobody surveyed for that purpose.
• The wrong runtime. IT wants enough time to shut down. Process wants enough time to reach a safe state, which on a continuous process can be a very different number in either direction.
• A shared failure mode. One UPS means one battery string, one inverter and one maintenance bypass. A fault or a service outage takes both systems at once, so you lose the process and the ability to see what happened to it in the same instant.
• Maintenance conflict. The IT UPS can be taken to bypass during a change window at 2am. The process UPS on a continuous plant may have no such window for months. Combining them means the more restrictive constraint governs, and in practice that means the maintenance quietly does not happen.

What a process trip actually costs

We deliberately do not quote a dollar per minute figure, because the honest answer is that it varies by an order of magnitude between plants and anyone quoting a universal number is guessing. What is consistent is the mechanism, and the mechanism is what you should be costing internally.

On a batch process, an uncommanded stop mid batch usually means the batch is gone. Not delayed, gone, because the recipe cannot be resumed from an unknown state. You lose the raw material, the energy already put into it, and you inherit a disposal problem. On a continuous process, the cost is restart time. Extruders, furnaces, dryers, kilns and reactors do not resume, they have to be brought back up through a controlled ramp, and that ramp is measured in hours with off-specification product coming out of the end of it for much of the way. On a food or beverage line, an unplanned stop with product in the equipment frequently triggers a full clean down and requalification before you are allowed to restart, and that clean is on the critical path whether or not the outage itself lasted thirty seconds.

There is also a category that does not appear on any downtime spreadsheet. A hard, uncontrolled stop is mechanically violent. Loaded conveyors, full pipework and spinning masses stopping without a controlled ramp put stress into gearboxes, couplings and seals, and that shows up as maintenance weeks later without anyone connecting it back to the power event. A process UPS that buys enough time for a controlled shutdown avoids that entirely, which is a benefit that is real and almost never counted.

Switchroom and MCC room cooling, the heat load nobody sizes

This is the section our competitors skip, and it is the one that causes the most avoidable equipment failure on Australian industrial sites. A switchroom or motor control centre room full of variable speed drives has a large, continuous, year round thermal load, and it is routinely underestimated because the heat is invisible and the equipment does not complain until it fails.

Where the heat comes from

A variable speed drive is efficient but not free. Typical VSD efficiency sits around 97 to 98 per cent, so roughly 2 to 3 per cent of the power passing through the drive is dissipated as heat, mostly in the rectifier, the DC bus and the IGBT power stage. That heat goes into the room. It is continuous whenever the drive is loaded, it is not seasonal, and unlike an office it does not go away at night on a plant running around the clock. Add switchgear and busbar losses, control transformers, panel lighting, the odd charger and any solar gain through the roof and walls, and a switchroom that was fitted with a wall-mounted extraction fan is now well beyond what that fan can move.

The table below is a sanity check, not a substitute for a proper heat load calculation, but it is enough to tell you quickly whether your existing arrangement is plausible.

What high ambient does to drive and contactor life

Drives are usually rated for full output up to around 40 degrees Celsius ambient, and above that the manufacturer requires derating, meaning the drive can no longer deliver its nameplate current. An unconditioned Australian switchroom in February does not stay below 40 degrees Celsius. It is common to walk into one and find it 10 to 15 degrees above outside ambient because the room is full of heat sources and has no meaningful ventilation path.

The failures that follow are predictable. Electrolytic capacitors in the drive DC bus are the classic temperature-limited component, with life falling steeply as operating temperature rises. Cooling fans in the drives run harder and fail sooner. Contactor and relay contacts, terminations and busbar joints all age faster. None of this trips an alarm. It simply moves a wave of drive and starter failures forward by several years and turns them into unplanned outages, which is exactly the cost you were trying to avoid when you bought the UPS.

Ventilation, comfort cooling or precision cooling

Mechanical ventilation moves outside air through the room. It is cheap, it needs no refrigerant, and on a mild site with a modest load it can be adequate. It has two problems in an industrial setting. The room can never be cooler than outside ambient, which in most of Australia is not good enough in summer, and you are now pulling unfiltered plant air, dust, salt or corrosive vapour across live switchgear. Where mechanical ventilation is used in a building context, hygiene and maintenance obligations under AS/NZS 3666 apply to the air handling system.

A comfort split system is the usual compromise and the usual disappointment. It is designed for a load that varies through the day and mostly disappears at night, and it is optimised for a mix of sensible and latent cooling suited to people. A switchroom load is almost entirely sensible, constant and continuous. Comfort units short cycle, over-dehumidify, and they have no redundancy, no meaningful filtration and often no alarm path, so the first anyone knows about a failure is when the drives start tripping. Precision cooling is built for exactly this duty: high sensible heat ratio, continuous operation, tight control and proper monitoring. That is the case we set out in detail on our precision CRAC cooling for switchrooms and MCC rooms page.

All electrical work associated with the room is carried out to AS/NZS 3000. Any refrigerant handling on the mechanical side is carried out by ARCtick licensed technicians as required under the Ozone Protection and Synthetic Greenhouse Gas Management Regulations. Where the switchroom also houses a UPS with a vented battery installation, room ventilation and gas detection need to be considered together, which we cover under gas detection for battery rooms.

Harmonics, UPS sizing and cable derating

Every variable speed drive and every rectifier in the plant, including the rectifier inside the UPS itself, draws current in pulses rather than smoothly. That non-sinusoidal current is harmonic current, and on an industrial site with a lot of drives it is not a rounding error. It changes how you size the UPS, how you size the cables and transformers feeding it, and whether the generator behind it behaves.

Why it matters for UPS sizing

Two separate effects have to be handled. The first is what the UPS injects upstream. A UPS with a six-pulse rectifier front end is itself a harmonic source, and on a site already carrying significant drive-generated distortion, adding another one can push voltage distortion at the switchboard past the point where sensitive equipment misbehaves. Modern transistorised or active front end UPS designs bring input current distortion down substantially and largely remove this concern, which is why input current distortion should be an explicit line in the specification rather than something you discover at commissioning.

The second is what the UPS is asked to supply. Non-linear downstream loads mean the UPS output sees a crest factor well above that of a resistive load. A UPS sized on kW alone, ignoring the load power factor and crest factor, will be undersized in practice. This is the same issue that makes an industrial UPS a genuinely different product rather than a marketing label: it needs the output filtering, the overload tolerance and the fault clearing capability to sit in front of motor starters, drives, welding equipment and lamp loads without going to bypass every time something starts.

Transformer and cable derating

Harmonic current does real thermal damage upstream. Transformers suffer additional eddy current losses that rise steeply with harmonic order, which is why transformers on drive-heavy sites are commonly derated or specified with a K-factor rating. Neutral conductors in three-phase four-wire systems carry triplen harmonics that do not cancel between phases, so neutral current can approach or exceed phase current, and a neutral sized on the old assumption of near-zero current runs hot. Cables carrying distorted current suffer additional heating from skin and proximity effects. None of this is visible on a standard clamp meter reading true RMS current alone, and all of it shortens the life of assets nobody is inspecting.

In Australia and New Zealand, harmonic current emission limits for equipment connected to public low voltage systems are set out in the AS/NZS 61000 series for electromagnetic compatibility. AS/NZS 61000.3.12 covers limits for harmonic currents produced by equipment with an input current above 16 A and up to 75 A per phase, and AS/NZS 61000.3.2 covers equipment at or below 16 A per phase. Distribution network service providers may also impose their own connection conditions on larger industrial sites, so the practical answer on a specific site is measurement first, then design. Our power integrity inspection service exists precisely for this: measure what the site is actually doing before anyone specifies anything.

Enclosures, ingress protection and SCADA integration

IP rating and material selection

Standard UPS and control enclosures are built for offices. On a plant floor they are consumables. The ingress protection rating tells you what the enclosure will tolerate, and the right choice depends on what the environment actually does rather than what the site drawing says it does.

• Dusty environments such as cement, quarrying, grain handling, timber and dry powder processing need at least IP54 for dust protection, and conductive dust environments need more. Dust is the single most common cause of premature electronics failure on Australian industrial sites, because it blankets heat sinks and turns a thermal design into a thermal problem.
• Wet and washdown areas in food, beverage and rendering need IP65 or IP66, because washdown is a directed jet of water and often a hot caustic or acidic one. An IP54 enclosure in a washdown bay will be full of water within a month.
• Food-grade and hygiene areas need 316 stainless steel enclosures with sloped tops, continuous welds and hygienic seals so that they can be cleaned rather than merely wiped. Painted mild steel fails audit and then fails structurally.
• Corrosive and coastal environments including water and wastewater treatment, chemical handling and anywhere with chlorine, hydrogen sulphide or salt air need corrosion-protected coatings or stainless, and the cooling equipment on the enclosure needs the same treatment as the enclosure itself. A stainless panel with a standard coated enclosure cooler simply fails at the cooler.

The trade-off to be aware of: a higher IP rating means a sealed enclosure, and a sealed enclosure cannot dump heat by exchanging air with the room. That is why enclosure air conditioners and air-to-air or air-to-water heat exchangers exist, and why increasing the IP rating without addressing the thermal path just relocates the failure from water ingress to overtemperature.

Dry contacts and Modbus into SCADA

A UPS or cooling unit that is not visible in the control room is not being monitored, no matter what the manual says. Industrial sites do not want another vendor portal and another login. They want the asset in the same SCADA screens as everything else, with alarms that route through the existing alarm management.

There are two practical routes and most sites should use both. Volt-free dry contacts give you the handful of signals that must never depend on a protocol stack: on battery, low battery, on bypass, general alarm, and for cooling, unit fault and high temperature. They are hard-wired into a spare digital input on the PLC and they work when everything else does not. Modbus RTU over RS-485, or Modbus TCP over the plant network, then gives the full data set: input and output voltage and current per phase, load percentage, battery voltage, temperature, autonomy remaining, and for CRAC units supply and return air temperature, humidity, compressor and fan status and filter differential pressure. SNMP is available on most equipment as well and suits the IT side, but on the process side Modbus is almost always the shorter path because the PLC already speaks it.

Specify the interface at purchase order stage. Retrofitting a communications card to an installed UPS means an outage on the thing you installed specifically to avoid outages.

What Indigi does on an industrial site

Indicative pricing. Cooling design and installation runs from $8,000 to $15,000 for small rooms, $25,000 to $45,000 for medium installations and $45,000 to $60,000 plus for large projects, with switchroom and MCC room work usually landing in the small to medium bands unless the room needs redundancy. UPS maintenance is $720 ex GST for a one-off visit or $1,940 per year ex GST for a standard annual plan. Single-phase UPS installation starts from $850 up to 3kVA, $1,250 for 3 to 10kVA and $1,800 for 10 to 20kVA. Three-phase installation starts from $3,200 for small systems and $6,500 for medium systems. Sites in hazardous areas, food-grade zones or with high IP requirements are quoted individually because the enclosure and installation method drive the cost more than the equipment does.

Frequently asked questions

What is an industrial UPS?

An industrial UPS is an online double conversion uninterruptible power supply engineered for the electrical and physical conditions of a plant rather than an office. It is built to sit in front of non-linear and inductive loads such as motors, variable speed drives, welding equipment and lamp loads, with the output filtering, overload tolerance and fault clearing capability those loads demand. It also carries a wider operating temperature range and is usually supplied in an enclosure rated for dust, moisture or corrosion. On an industrial site the term covers two quite different applications: a small process UPS protecting control equipment, and a larger infrastructure UPS protecting IT and networking.

What is the difference between an industrial UPS and a commercial UPS?

A commercial UPS is designed for a predictable, mostly linear IT load in a temperature-controlled room, and it is optimised for cost, footprint and efficiency at that duty. An industrial UPS is designed for a hostile environment and a difficult load. The practical differences are a wider input voltage and frequency window, higher tolerance of overload and inrush from motor starting, better output performance into non-linear loads with a high crest factor, wider operating temperature range, ruggedised components, higher IP rated enclosures, and industrial communications such as Modbus and volt-free dry contacts rather than only SNMP. Putting a commercial UPS in front of a drive-heavy load is the most common cause of a UPS that spends its life on bypass.

Why does a PLC need a UPS?

Because a PLC losing power does not fail cleanly, it fails at whatever point in the sequence it happened to be at. The process stops in an unknown state, outputs drop out in an uncontrolled order, and the plant has to be brought back manually rather than resumed. A PLC also does not need a supply interruption to be upset. Voltage sags of a few cycles caused by a large motor starting elsewhere on site, or by a fault on the incoming network, are enough to reset a controller or corrupt an I/O scan. A process UPS rides through those disturbances, and where the outage is genuinely long it buys the time to execute a controlled, sequenced shutdown so the restart is a procedure rather than an investigation.

How long should a UPS run a PLC panel for?

Set the runtime from the process requirement, not from a default. If the intent is to ride through short disturbances and hand over to a generator, then a few minutes past the generator changeover time is enough. If the intent is a controlled shutdown, the runtime has to cover the longest controlled shutdown sequence on the plant with margin, which on batch equipment is often 10 to 20 minutes and on some continuous processes considerably more. If the intent is to keep the control system alive and observable through a longer outage while the process itself is down, that is a different and larger number again. Also allow for battery ageing, because a battery at the end of its service life will not deliver the runtime it delivered when new, and for the panel temperature, because heat is what gets it there faster.

How much heat does a VSD generate?

A variable speed drive typically operates at around 97 to 98 per cent efficiency, so roughly 2 to 3 per cent of the power passing through it is dissipated as heat into the surrounding room. As a working figure, a switchroom with 250 kW of installed drives running at about 70 per cent loading is putting somewhere near 4 to 5 kW of continuous heat into the room from the drives alone, before switchgear losses, transformers, lighting and solar gain through the building fabric are added. That heat is continuous whenever the plant is running, which is why a switchroom cooling load behaves nothing like an office cooling load and why it should be calculated rather than estimated from room volume.

Why do VFDs cause harmonics?

A variable frequency drive rectifies incoming AC to a DC bus before inverting it back to a variable frequency output. That rectifier does not draw current smoothly across the whole waveform. It conducts only when the instantaneous supply voltage exceeds the DC bus voltage, so it draws current in short, high pulses twice per cycle per phase. A pulsed current is mathematically a fundamental plus a series of harmonics, and with a six-pulse rectifier the dominant ones are the fifth, seventh, eleventh and thirteenth. Those harmonic currents flow back through the site distribution, causing extra heating in transformers, cables and neutrals, and distorting the voltage waveform that every other piece of equipment on the board has to live with. Mitigation options include line reactors and DC link chokes, passive or active harmonic filters, and multi-pulse or active front end drive designs.

Does an MCC room need air conditioning?

In most Australian climates, yes, once the room contains a meaningful quantity of drives. Mechanical ventilation alone can never bring the room below outside ambient, and it pulls plant dust, salt or corrosive vapour across live switchgear. Drives are generally rated for full output only up to about 40 degrees Celsius ambient and must be derated above that, and an unconditioned switchroom in an Australian summer will exceed that comfortably. The right answer is a room load calculation followed by dedicated cooling sized for a continuous, almost entirely sensible heat load, with monitoring and alarms brought back into SCADA. A comfort split system is a common substitute and a poor one, because it is designed for a variable daytime load with a significant latent component and it has no redundancy and no alarm path.

Where we work

Related services and industries

• Precision CRAC cooling for switchrooms and MCC rooms
• High temperature and industrial outdoor UPS systems
• Three phase UPS installation and commissioning
• Single phase UPS installation and commissioning
• UPS maintenance plans and battery maintenance and replacement
• Power integrity inspection for harmonic and power quality measurement
• Socomec UPS Australia, including industrial ranges built for non-linear loads
• Gas detection for battery rooms and BESS
• UPS battery replacement cost calculator
• Mining and resources UPS and cooling for pit, plant and remote sites
• Warehousing and logistics for distribution centres and automated materials handling

Get your process and infrastructure power separated properly Send us a list of your control panels and their loads, the installed kW of drives in each switchroom, what is currently cooling those rooms and your shutdown windows. We will come back with a load classification, a switchroom heat load position and an installed price. Indigi Power and Cooling is Indigenous and Veteran owned, and registered with Supply Nation and ICN Gateway. Contact Indigi Power and Cooling Switchroom Cooling