7.2 Variable-Frequency Drives in Refrigeration Plants

Where VFDs Show Up in an Ammonia Plant

A variable-frequency drive (VFD) converts incoming three-phase AC to DC and synthesizes a new AC output whose frequency (and matching voltage) sets motor speed. On CIRO-level plants you will see VFDs most often on evaporative-condenser or cooling-tower fans, condenser-water or overfeed pumps, and some screw packages that use speed for capacity in addition to, or instead of, a slide valve. The exam is not asking you to commission a drive. It is asking whether you understand what speed does to flow and power, what the drive must not be allowed to do, and what it cannot replace.

Affinity Laws as Operating Intuition (Fans and Centrifugal Pumps)

For centrifugal fans and pumps, with the system curve in the same neighborhood:

• Flow rate varies approximately with speed: $Q \propto N$
• Head (or fan pressure) varies approximately with speed squared: $H \propto N^{2}$
• Power varies approximately with speed cubed: $P \propto N^{3}$

Those three lines are the reason a plant can float condenser pressure by slowing fans instead of hammering a fixed-speed fan on and off. Work a speed cut an operator will actually see:

• Speed to 80%: flow ≈ 80%, head ≈ 64%, power ≈ 0.80³ = 51% of the original fan or pump power.
• Speed to 70%: power ≈ 0.70³ = 34%.
• Speed to 50%: flow ≈ 50%, head ≈ 25%, power ≈ 12.5%.

The cube law is why "slow the condenser fans 20%" is a bigger kW move than it sounds — nearly half the fan power is gone at 80% speed, not 20%. It is also why you do not claim the same cube-law savings on a screw compressor. A rotary screw is positive displacement. Refrigerant mass flow is roughly proportional to rotor speed (and to slide-valve position, if the slide is not locked at 100%). Shaft power does not collapse with $N^{3}$. Treat affinity $N^{3}$ as fan and pump intuition. On a screw, think capacity nearly linear with speed, with oil, discharge temperature, and motor amps as the limits.

Minimum Speed: Oil, Cooling, and Seals

Every VFD application on this exam has a minimum speed (minimum Hertz) that is not "as slow as the keypad will go."

• Screw packages: oil injection, oil-separator velocity, rotor cooling, and sometimes a shaft-driven oil pump all assume a floor speed. Below that floor you can lose oil ΔP, starve bearings, or overheat even if the slide valve is unloading. The oil-pressure interlock still applies. If ΔP collapses at low speed, the compressor must drop — the drive is not a license to run without oil.
• TEFC fan motors: the motor's own cooling fan is on the shaft. Sustained extremely low speed can overheat the motor even at modest torque. Condenser fans also need enough airflow to be stable on the fan curve (stall and vibration).
• Pumps: mechanical seals and some overfeed circuits need a minimum flow. Dead-heading a pump on a VFD at low frequency still makes heat. NPSH margins also change with speed; slowing a pump does not automatically cure cavitation if the liquid is already too close to saturation.

Use the OEM minimum Hertz. Do not "save energy" by forcing 5 Hz on a machine whose oil system was designed around 30 Hz.

Skip Frequencies and Mechanical Resonance

A long condenser-fan shaft, a cooling-tower stack, a pump base, or a discharge pipe can have a natural frequency inside the operating speed range. If the drive dwells there, vibration grows, bearings spit, and welds crack. Skip frequencies (rejected speed bands) tell the drive to pass through those RPM ranges and not linger. On a CIRO item, skip frequencies are mechanical-resonance avoidance, not utility billing tricks and not harmonic frequencies the drive is supposed to amplify.

If a fan "only shakes at 38 Hz," the correct operations move is to widen or add a skip band (and inspect the mechanical problem), not to keep running there because the PID wants that speed for head-pressure control.

Harmonics and EMI

A VFD is a non-linear load. It draws current in pulses and injects harmonics (classically 5th and 7th on six-pulse front ends) back toward the plant bus. Effects you should be able to name:

• Extra heating in transformers, capacitors, and some motors on the same bus
• Distorted voltage that fools analog meters and some older power-factor capacitors
• Electromagnetic interference (EMI) on nearby 4–20 mA level, pressure, and temperature loops if motor leads are unshielded or analog cables share a tray with VFD output cable

Mitigations you will see in a machinery room: line reactors or DC chokes, VFD-rated motor cable, short motor leads or dV/dt / sine filters on long runs, shields landed at the drive, and analog pairs kept away from the inverter output. IEEE 519 is the usual harmonic-limit conversation with the utility and the electrical engineer — operators need the symptom list, not a memorized THD table that IIAR does not publish for CIRO.

Interlocks Still Win

Never assume a VFD bypasses interlocks. The drive's speed reference, PID, and "auto" HOA are capacity and energy tools. Safety is still the series path taught in the ladder section: high discharge pressure, high discharge temperature, high level, oil ΔP after start delay, motor overload, and e-stop. Those contacts should drop enable, Safe Torque Off (STO) if the drive has it, or the run permissive — not merely ramp the speed reference to 0.0 Hz. A coast-to-zero on a live enable is not a safety stop if the drive can be told to run again by a PID bump.

If the package has an across-the-line bypass contactor for "VFD failed, run at 60 Hz," the same cutouts must still drop that contactor. Bypass is not a holiday from oil pressure or high-side cutouts. A VFD also does not make a motor immune to single-phasing on the line side; input-phase loss should trip the drive, and you still treat a downstream output fault as a stop, not as a reason to jumper faults.

HOA at a VFD: Hand is a local speed or start that must still see the safety chain; Off is a stop; Auto follows the plant controller. Hand is not permission to run a high-level vessel or a machine with the oil-pressure switch jumped.