VFD retrofit: affinity-law deep-dive

VFD retrofits on centrifugal pumps, fans, and blowers are the single deepest ECM category in Opnor's library — the highest-frequency match across every industry, the most reliable savings, and the cleanest math. The savings come from one well-known equation. Understanding it is the difference between "VFDs save 20% always" (wrong) and knowing which assets actually qualify.

Why VFDs save energy at all

Most centrifugal loads in industrial plants — process pumps, cooling tower fans, induced-draft fans, blowers — are sized for peak conditions and run most of the time at partial load. Without a VFD, the standard way to handle partial load is mechanical throttling: a valve, damper, or vane that restricts flow downstream of the pump or fan.

Throttling wastes energy. The motor still runs at full speed; the restriction dissipates the surplus pressure as heat. The pump itself is operating at the wrong point on its curve, also dissipating energy as recirculation and turbulence.

A VFD removes the restriction. Instead of running at full speed and throttling output, the motor runs at the speed needed to deliver the required flow. The energy you don't spend overcoming an artificial restriction is the savings.

The affinity law (cube-of-speed)

For centrifugal loads, three relationships govern how flow, pressure, and power scale with motor speed:

The cube relationship is the magic. Halve the speed → use 1/8th the power. Run at 80% speed to deliver 80% flow → use 51% of the original power. Reduce flow by a small amount → cut power by a much larger amount.

Which assets qualify

The Opnor ECM library matches a VFD candidate when:

• Asset type is centrifugal (pump, fan, blower) — confirmed via type field or auditor walk-through
• Existing drive type is not already a VFD (i.e. has_vfd = False) — avoids re-recommending what's already there
• Nameplate kW ≥ 5 — below this, capex per VFD doesn't pay back even at generous load profiles
• Load factor ≤ 0.85 — at full load, there's nothing for the VFD to recover (you're already at 100% speed). Below 0.85 LF means the asset spends meaningful time at partial load.
• Operating hours ≥ ~3,000/yr — short-runtime assets don't accumulate enough savings to clear payback

The savings calculation

For a candidate asset, the annual energy savings from a VFD retrofit is:

savings_kwh = nameplate_kw × hours × qty × (1 − LF³) × realization

Breaking down each term:

Worked example. A 22.5 kW centrifugal pump running 4,200 hours/year at LF = 0.68:

savings = 22.5 × 4,200 × 1 × (1 − 0.68³) × 0.80 = 22.5 × 4,200 × (1 − 0.314) × 0.80 = 22.5 × 4,200 × 0.686 × 0.80 ≈ 51,879 kWh/yr

At a typical Quebec industrial rate of $0.07/kWh, that's $3,632/year in electrical savings on a single pump.

Realization factors and the 0.80 multiplier

The cube law is theoretical. In practice, VFD installations don't hit 100% of theoretical savings because:

• VFDs themselves have a small efficiency loss (~2–4% at typical operating points)
• The motor running below its design speed sees efficiency degradation, especially below 50% speed
• System pressure isn't always perfectly proportional to N² — real piping has static head that doesn't scale
• Operators sometimes set a minimum-speed limit higher than necessary, leaving savings on the table
• The actual load profile isn't always exactly what the audit assumed — bursts of high-load operation that weren't captured in the LF estimate

The 0.80 realization factor is the empirical adjustment we apply to bring the theoretical number in line with measured post-implementation results. It's calibrated against years of post-VFD M&V data on industrial installations and lands savings within ±10% of measured on most projects.

When VFDs don't pay back

Even on a qualifying centrifugal load, a VFD retrofit can fail to clear the payback hurdle. Common reasons:

• The asset already runs at near-constant high load (LF > 0.85). The cube law gives you almost nothing back.
• Operating hours are too low. A 30 kW pump running 800 hours/year saves very little even at low LF.
• Existing system has significant static head that VFDs can't reduce. Net savings can be 30–50% of the cube-law expectation.
• Power quality issues require harmonic filters, doubling the capex. Pushes payback past 4 years on smaller assets.
• Process needs constant flow regardless of demand (e.g. sealing fluid pumps). No partial-load operation = no VFD savings.

The Opnor ECM library checks for these conditions during matching. If we're recommending a VFD on your asset, the asset has already passed the qualification checks. The estimate isn't universal — it's the result of running the matching engine against your specific asset list.