Prepare for the CEA by mastering the auditor's chain of reasoning: classify the audit scope, analyze consumption and demand from billing data, build a defensible baseline, quantify energy conservation measures with honest load assumptions, and document findings with appropriate economic metrics. Practice each link with worked numbers and a self-check rubric rather than isolated facts.
Audit levels are scope decisions, not just vocabulary
Preliminary, walk-through, detailed, and investment-grade audits differ in depth, measurement effort, and the confidence of the savings estimates they can support. Matching scope to the client's decision is a core auditor competency.
Compare a walk-through audit with an investment-grade audit. A walk-through identifies obvious opportunities from observation and brief data review; its savings figures are order-of-magnitude. An investment-grade audit supports capital commitments, so it demands measured loads, verified operating hours, and documented assumptions. A finding that is acceptable at one level can be indefensible at another.
Apply this by asking, for any practice case: what decision will the report support, and does the evidence depth match it? If a case asks you to recommend a chiller replacement based only on a site visit and nameplate data, the correct response is usually to specify the measurement campaign needed, not to finalize savings numbers. Practicing that boundary judgment is more valuable than memorizing audit names.
Scenario: a case describes a two-hour facility visit where a candidate recommends a 25% heating-energy reduction from a controls upgrade. The better decision is to present the measure as a preliminary opportunity requiring logger data on boiler cycling and space temperatures before quantification. The mistake is treating a screening estimate as a bankable figure; the distinction matters because report credibility and client investment decisions both depend on it.
- Walk-through: identifies opportunities; relies on observation, brief bill review, and screening estimates.
- Detailed audit: adds measured end-use data, load profiles, and energy balance reasoning.
- Investment-grade: supports capital decisions; requires verified run hours, measured loads, and documented assumptions.
| Audit level | Typical evidence | Savings estimate character | Decision it supports |
|---|---|---|---|
| Preliminary / walk-through | Site observation, utility bill review | Order-of-magnitude, wide ranges | Whether a deeper audit is warranted |
| Detailed audit | End-use measurements, load profiles, operating schedules | Refined estimates with stated assumptions | Prioritizing measures and O&M actions |
| Investment-grade | Verified loads, run hours, vendor quotes, risk analysis | Tight estimates suitable for financing | Capital approval and contract terms |
Reading bills: separating energy, demand, and load factor
Auditors must distinguish kilowatt-hours (energy, what you used) from kilowatts (demand, how fast you used it), because commercial tariffs price both, and load factor summarizes the gap between average and peak use.
Load factor equals total kWh in a period divided by (peak kW multiplied by hours in the period). A high load factor means consumption is steady; a low one means short peaks drive demand charges. This single number tells an auditor whether demand-side measures (scheduling, peak clipping, load shifting) deserve attention before or alongside kWh-reduction measures.
In practice, start every bill analysis by building a small table: monthly kWh, peak kW, billed demand, load factor, and cost per kWh all-in. Billed demand can differ from metered peak under ratchet or minimum-demand provisions, which is exactly why the two columns are kept separate. Trace one year of data before proposing anything.
Worked example: a facility uses 180,000 kWh in a 720-hour month with a 500 kW peak. Load factor = 180,000 / (500 × 720) = 0.50. A plausible mistake is to propose a large kWh-saving measure to address this; the better first look is a load profile, because a 0.50 load factor with high demand charges suggests peaks clustered in a few hours, and scheduling shifts may cut cost faster. In a simplified tariff with a $12/kW-month demand charge, shaving the peak by 60 kW saves 60 × 12 = $720 per month, illustrating why demand and energy need separate treatment.
Demand analysis worked scenario: the ratchet assumption
Billing structures modify raw savings math. Minimum-demand or ratchet provisions can make measured peak reductions worth less than they appear, so auditors must check the tariff before claiming demand-charge savings.
Scenario: a plant's July peak is 900 kW; the fall peak drops to 400 kW after a production change, yet the bill still shows 700 kW of billed demand. A plausible mistake is to report the full demand-charge saving from 900 kW down to 400 kW. The better decision is to compare metered peak against billed demand each month and identify the tariff provision carrying the difference, then recompute savings against billed demand.
Why it matters: savings claims that ignore billing rules overstate value and damage the audit's credibility. Self-check exercise: take twelve months of bills from any commercial example (or a sample dataset you construct), compute load factor and cost per kWh each month, and write one sentence per quarter explaining which driver—energy, demand, or rate—changed the cost most. Expected observation: months with similar kWh but different peaks have noticeably different bills, confirming demand is an independent lever.
Extend the habit to rate structures generally: block rates make marginal kWh more or less valuable than average cost, and time-of-use pricing attaches value to when consumption occurs. An auditor who quotes only an average cost per kWh cannot correctly rank measures that shift load across hours. In your notes, keep three columns per tariff type you study: what it prices, what it rewards, and what measure class it favors.
Baselines and degree days: adjustment is not savings
A baseline is normalized consumption under defined conditions, commonly weather-adjusted using degree days. Savings are the difference between the adjusted baseline and actual use, not between two raw monthly totals.
Compare two comparisons: this January versus last January, and this January versus a degree-day-adjusted baseline. The first ignores weather; a mild winter can masquerade as savings. The second asks what this facility would have consumed given this month's weather under pre-retrofit operating conditions, which is the auditor's question. Named tools include heating and cooling degree days, per-day normalization, and simple regression of energy use on degree days.
Trace an example: regress monthly fuel use against heating degree days. The slope is weather-sensitive consumption (per degree day); the intercept approximates the non-weather base load. If a controls retrofit changes schedules rather than envelope performance, expect the intercept to move while the slope stays similar, and check that prediction against the post-retrofit data. This is how an auditor reasons about which measure class changed which model parameter.
Scenario: a report claims 18% heating savings by comparing fuel bills before and after a retrofit across different winters. The better decision is to rebuild the baseline with degree-day normalization and restate savings against the adjusted baseline, stating the uncertainty. The mistake matters because weather variation can exceed the measure's real effect, and a savings figure that cannot survive normalization will not survive review.
Quantifying ECMs: nameplate data versus measured reality
Energy conservation measure calculations live or die on load assumptions. Nameplate ratings, guessed run hours, and rule-of-thumb diversity factors routinely produce inflated savings; measured inputs produce defensible ones.
Worked example: a lighting retrofit replaces 100 fixtures at 100 W each with 60 W equivalents, quoted at 4,000 annual operating hours. Nameplate math gives 100 × 40 W × 4,000 h = 16,000 kWh saved. A plausible mistake is stopping there: nameplate wattage may reflect a lamp-and-ballast system different from the installed one, and operating hours come from a schedule assumption, not the space. The better decision is to measure a sample of fixture circuit wattages and log hours in representative spaces, then recompute.
Recompute with measurements: circuit measurement shows actual 85 W existing fixtures, and logging shows 3,200 hours. Savings = 100 × 25 W × 3,200 h = 8,000 kWh, half the original claim. Note also that lighting changes affect heating and cooling: reduced internal gains slightly raise winter heating and lower summer cooling loads. A defensible estimate names these interactions even if it bounds them rather than modeling them precisely.
Apply the same skepticism to motors, air compressors, and HVAC: nameplate horsepower is not absorbed power, and full-load hours are not run hours. For each measure you practice, list the two or three inputs that dominate the calculation and ask what evidence the case provides for each. If the evidence is absent, the correct exam-style answer is usually to specify the measurement, not to pick the largest number among the choices.
Field measurement quality: a self-check exercise
Auditors choose instruments and methods that answer the question without disrupting operations. Spot readings, short-term logging, and long-term monitoring differ in cost, effort, and the confidence they support.
Compare three approaches to one question, such as exhaust fan runtime: a spot current reading taken once, a one-week current logger, and an interview-based schedule. The spot reading cannot reveal cycling; the interview cannot verify actual practice; the week-long log shows both duty cycle and variation. Matching method to question is the skill. Practice by writing, for five common questions (motor loading, lighting hours, hot water draw, space temperature setpoints, compressor cycling), the minimum credible method for each.
Exercise with expected observations: take one motor in your study scenario. Record nameplate horsepower and efficiency, then reason through what a clamp-on current reading implies about load given motor part-load behavior. Self-check rubric, score each 1–4: (1) Can you state the difference between nameplate and absorbed power? (2) Can you explain why a motor near full load differs electrically from one lightly loaded? (3) Can you name a non-intrusive way to estimate runtime? (4) Can you state what logging interval you would choose and why? A total of 12 or higher suggests you can reason about measurement quality; lower scores point to which concept to re-study. These are learning milestones for your own tracking, not a prediction of any exam result.
- Spot measurement: cheap, instantaneous, blind to cycling and variation.
- Short-term logging: captures duty cycles and schedules over days.
- Long-term monitoring: supports verification and persistence tracking.
- Interviews and schedules: useful context, never sufficient alone for savings claims.
Economics, reporting, and a realistic preparation sequence
Rank measures with simple payback for screening but understand its limits, document assumptions explicitly, and prepare through a sequence that cycles content review with scenario practice and self-scoring.
Simple payback divides installed cost by annual savings; it is fast and intuitive but ignores measure life, timing of cash flows, and value beyond the payback point. Contrast it with life-cycle thinking, where a longer-lived measure with the same payback can be clearly superior. An auditor report should state which metric it uses and why, and flag assumptions (energy prices, escalation, hours) that drive the ranking.
Scenario: two measures cost the same and save the same energy; one lasts five years, the other fifteen. A plausible mistake is calling them equivalent because payback matches. The better decision is to compare total value over measure life and note the difference explicitly. Why it matters: recommendation quality, not just savings math, is what an audit deliverable is judged on.
Suggested preparation sequence, adaptable to your schedule: weeks one and two, build the bill-analysis toolkit (kWh, kW, load factor, cost per kWh) and normalize one year of sample data. Weeks three and four, work ECM calculations for lighting, motors, and HVAC with honest input assumptions. Weeks five and six, practice full mini-cases: classify the audit level, build the baseline, quantify two measures, and rank them. Each week, run the self-check rubric from the measurement section and re-study the weakest concept. This order builds skills in the sequence an actual audit uses them.
- Readiness check 1: compute load factor and all-in cost per kWh from any sample bill set without notes.
- Readiness check 2: normalize heating energy to degree days and state what slope and intercept mean.
- Readiness check 3: for three ECM types, list the two inputs that dominate the savings calculation and a credible source for each.
- Readiness check 4: given any case, state the audit level the evidence supports and the decision it can back.
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.