A 90 kW VSD Proposal on a Leaky, Load/Unload System
A plant has two 75 kW oil-free screw compressors rated 11.0 m³/min at 7.5 bar; one runs as lead in load/unload and the other is standby. Measured at the lead compressor: 62 kW loaded, 18 kW unloaded, with a cycle of 52 s loaded and 9 s unloaded during production. Average production air demand is 9.5 m³/min over 6,000 operating hours per year. An ultrasonic leak survey with production running tags 41 leaks and estimates 2.9 m³/min total. End uses need 6.2 bar at the farthest point; storage is a single 1 m³ receiver. The vendor proposes a 90 kW VSD compressor for $95,000 and claims 35% savings. Electricity is $0.12/kWh with a $14/kW-month demand charge.
1. Evaluate the vendor's 35% claim using the measured load/unload data. What does the 18 kW unloaded power tell you?
Read the worked answer
Establish the current operating point from the cycle.
- Duty = 52/(52+9) = 0.853, which matches 9.5/11.0 = 0.864 - the measurements are consistent.
- Average power = 18 + 0.853 × (62 − 18) = 55.5 kW
- Specific power = 55.5 kW / (9.5 × 60 = 570 m³/h) = 0.097 kWh/m³ - reasonable for a 75 kW screw, so the machine is not internally faulty.
The 18 kW unloaded power is the real story. It is 29% of full load, being paid for every hour of every cycle, and it also sets the marginal value of anything that reduces flow: every m³/min removed saves roughly (44 kW ÷ 11 m³/min) × duty reduction, not the full-load 0.097 kWh/m³.
On the 35% claim: the leaks are 2.9 of the 9.5 m³/min. A VSD still compresses leaks - it cannot save flow that is being wasted. A large share of the claimed saving is leak removal that the vendor has priced into a $95,000 machine. Also, a 90 kW VSD (~14 m³/min) sized for un-reduced demand would run at ~47% of capacity after leak repair, a poor turndown point. Demand the vendor's savings calculation: what flow, what hours, what specific power curve, and what baseline.
2. Quantify the leak and pressure actions and sequence them against the VSD purchase.
Read the worked answer
Step 1 - Leak repair (2.9 → ~0.9 m³/min)
- Demand falls 9.5 → 6.6 m³/min; new duty = 6.6/11.0 = 0.60
- Average power = 18 + 0.60 × 44 = 44.4 kW → saving ≈11.1 kW
- Energy: 11.1 × 6,000 = 66,600 kWh ≈ $8,000/yr, plus ~$1,900/yr demand
- Cost ~$15,000 → payback ~1.5 years
Step 2 - Pressure reduction (7.5 → ~6.9 bar)
- 0.6 bar is ~4% of power at roughly 1% per 0.14 bar, but only if nothing downstream is starved
- Saving ~1.5–2 kW ≈ 9,000–12,000 kWh ≈ $1,100–1,400/yr for a few thousand dollars of regulators and switch adjustment
- Mandatory check: minimum pressure at the worst-case end use during peak demand, and filter/regulator pressure drop
Step 3 - Storage and sequencing
The 1 m³ receiver is small for 11 m³/min. Add storage (~2–3 m³) to stop short-cycling, then sequence the two fixed-speed machines so the trim is not a lightly-loaded 75 kW unit.
Step 4 - Reassess the VSD
Only now. Re-measure demand (day/night/weekend profile). If the post-leak profile still swings widely, a correctly sized VSD trim (probably 45–55 kW, not 90 kW) may clear the hurdle at the 12% rate; if the profile is flat at 6.6 m³/min, load/unload plus good sequencing already captures most of it.
Sequence: leaks → pressure → storage → re-measure → maybe VSD. Fixing demand first shrinks both the capital and the risk.
3. What measurement and verification plan would you write to prove the savings?
Read the worked answer
Use IPMVP Option B (retrofit isolation with continuous measurement) on the compressed-air system.
Baseline (2–3 weeks, all at once)
- kW logger on the lead compressor, 1-minute interval, with a separate digital input for the load/unload status - this is what turns cycles into an average power and lets you verify the 18 kW unloaded figure.
- Flow: either an installed thermal-mass meter or a pump-up test at a production stop (isolate the system, time receiver pressure rise against known volume) to get actual leak load before and after repair.
- Simultaneous production record, so the baseline can be normalised.
Independent variable and adjustments
- Production hours/units, not calendar hours. Savings must be reported at the post-retrofit production level.
- Ambient temperature if air-cooled compressors are involved.
Post-installation
- Repeat the same measurements for an equal period, at the same season if possible.
- Report power per m³ (kWh/m³) and total kWh separately - fixing leaks reduces kWh but barely changes kWh/m³, and a reviewer will ask.
- Confirm the demand-charge effect from the interval bill data (three highest 15-minute intervals), not from nameplate.
Guard rails: do not accept vendor-logged data alone; log independently at the compressor terminals. State the assumption that production stays within the baseline range; if it moves, adjust, don't compare raw kWh.