Scoping a Level 2 Audit at a Three-Shift Bakery
A 180,000 ft² bakery runs three shifts, seven days a week, with two natural-gas tunnel ovens, a 400 hp steam boiler, ammonia refrigeration, and 1,200 cfm of compressed air. Annual utility spend is $2.4 million: 22 GWh electricity and 850,000 therms gas. The owner commissions an ASHRAE Level 2 audit with a six-week schedule and two auditors. A prior audit five years ago listed 40 measures; only four were implemented, mainly because production would not accept downtime. Utility data exist only for the main electric and gas meters; ovens, refrigeration, and compressed air have no sub-meters. The production manager allows equipment shutdown only on Sundays from 02:00 to 06:00, and site safety induction consumes half a day. As-built drawings are incomplete. The owner wants 'quick wins' for this budget year plus a defensible list of capital projects for next year.
1. Define the audit boundary, baseline period, and level of effort for this Level 2 audit. What would you exclude or defer, and why?
Read the worked answer
Boundary: the whole bakery site, from incoming electric and gas meters through process ovens, refrigeration, compressed air, HVAC, lighting, and domestic hot water. The baseline should be 12 consecutive months, normalized to production tonnes, because output varies and the prior audit is too old to use as a baseline.
Level of effort: a Level 2 energy balance, not investment-grade. Use main-meter interval data plus targeted logging; do not attempt a full process simulation or meter every plug load. Allocate effort to the dominant process loads and to measures that production will accept. Defer detailed envelope analysis, plug-load surveys, and renewable screening to a later phase unless they are cheap to include. Exclude capital projects with no feasible shutdown path from the quick-win list, but keep them in the capital list with the downtime requirement stated.
The prior audit is a source of candidate measures, not evidence. Re-screen its 40 measures against current production and utility data. The deliverable should separate no-cost or low-cost operational measures from capital measures.
2. Given the Sunday shutdown window and missing sub-meters, outline a data-collection plan. How will you treat missing data and quantify uncertainty?
Read the worked answer
Use 15-minute utility data for 24 months and production records for the same period. Install temporary power loggers on the two oven lines, the ammonia rack, the compressed-air plant, and a sample of AHUs; log for at least two weeks, including a Sunday shutdown. Use the Sunday window to measure standby and no-load power, verify schedules, and test the boiler. Conduct an ultrasonic leak survey while the system is pressurized, and use a portable flow meter on compressed air. Use a clamp-on BTU meter on the boiler and read run hours from the control system.
Where logging is impossible, estimate from nameplate, measured amps, and run hours, and label the estimate. Cross-check the sum of sub-metered loads against the main meter. Treat missing data with low, base, and high assumptions rather than a single number. Report data quality and uncertainty for each end use, and do not present savings to three significant figures when the input is an estimate.
3. How would you screen and prioritize measures for the owner's two lists? Include economic and non-economic criteria.
Read the worked answer
Create two lists. The quick-win list requires no production downtime, simple payback under one year, and no product-quality risk: schedule setbacks, oven idle, compressed-air leak repair, boiler reset, and lighting controls. The capital list includes projects with payback, life-cycle cost, downtime, and non-energy benefits.
Screen on annual savings, installed cost, simple payback, and life-cycle cost, but also score production risk, maintenance, safety, comfort, and implementation complexity. Revalidate prior audit measures. Present savings as ranges and state the measurement method. Do not rank by savings alone: a large project that requires a shutdown production will reject is not a quick win. Use a decision matrix and a clear M&V plan so the owner can compare bids.