11.3 Power Cost, Demand Charges, and the Cost of Poor Operation

11.3 Power Cost, Demand Charges, and the Cost of Poor Operation

The January 2023 CIRO Study Guide lists compute cost in $/hr and energy over time as a reason the operating screens exist. A 255 kW abnormal picture is not complete until you can say what it costs. This section uses $0.17/kWh as a typical industrial energy price — the same figure that appears in CIRO study-guide practice prompts. It is a teaching rate, not your plant's tariff and not an official RETA utility table. If a stem gives another rate, use that rate. The method does not change.

Two different utility lines get mixed on exam items and in engine-room arguments:

1. Energy (kWh) — what you used, priced in $/kWh.
2. Demand (kW) — how hard you peaked, priced in $/kW-month (or a similar interval).

Cleaning a condenser can cut both. Treating demand as if it were kWh, or kWh as if a one-time peak did not matter, is how people "prove" a wash is not worth the overtime.

Energy: hour, day, week, 5,000-hour year

Cost = kW × hours × ($/kWh)

Use input kW from √3 × V × I × PF / 1000, or from a power meter. Do not convert nameplate HP × 0.746 and call it the bill.

Worked — West screw at the 11.1 normal screen, 240.0 kW, $0.17/kWh.

Arithmetic you should be able to repeat without a spreadsheet:

• Hour: 240 × 0.17 = 40.80
• Day: 40.80 × 24 = 979.20 (or 240 × 24 × 0.17)
• Week: 979.20 × 7 = 6,854.40
• 5,000 h: 240 × 5,000 = 1,200,000 kWh; × 0.17 = 204,000

Per-kW rule of thumb at this rate: one wasted kilowatt running 5,000 hours costs 5,000 × 0.17 = $850/year. Sixteen wasted kilowatts (the 11.1 abnormal delta of about 15.8 kW) is on the order of $13,400/year in energy alone, before demand. You do not need a new compressor to find that money. You need a condenser that still matches wet bulb.

Second machine — HS-1 at 187.3 kW. Hourly energy = 187.3 × 0.17 = $31.84/hr. 5,000 h: 187.3 × 5,000 × 0.17 = $159,205. If icing drops tons and you start a second 128 kW machine to hold rooms, that second machine is $21.76/hr and $108,800 over 5,000 h at the same rate — often larger than the kW/ton drift on the first machine. Poor operation is not only "amps went up." It is also "we ran an extra frame because the first one stopped making cheap tons."

Demand charges are a separate penalty

A demand charge bills the highest electrical kW recorded in a defined interval — commonly 15 minutes — during the billing month, at $/kW-month. The rest of the month can be gentle. If the dirty-condenser afternoon set a 15-minute peak, that peak can set the demand line for the entire month.

Example teaching demand price: $12.50 per kW-month. This is an illustration, not a tariff.

If West screw's 15-minute average is 240 kW during the plant's peak interval, that machine contributes 240 kW to billed demand:

240 × 12.50 = $3,000/month of demand allocated to that kW, on top of the kWh in the table above.

If a fouled condenser and extra lift push that same interval to 265 kW:

Extra demand kW = 25 kW

Extra demand cost = 25 × 12.50 = $312.50 that month

If the plant peaks at that dirty condition every month, extra demand is 312.50 × 12 = $3,750/year, in addition to extra kWh.

Demand does not care that you washed the condenser at 18:00 if the peak was at 14:00. It does not care that nights are light. Operators who only quote $/kWh understate the cost of a high-head afternoon coinciding with the rest of the site's peak (production, freezers coming out of defrost, and every other motor).

Ratchet / lookback (awareness, not a number to invent). Some industrial tariffs include a demand ratchet: a fraction of a prior peak follows you for months. CIRO will not make you recite an unpublished tariff. It will expect you to keep demand conceptually separate from energy and to see why a short abnormal kW spike is still expensive.

Excess cost of dirty condensers and high head

Lift kills COP. High condensing temperature means more compressor work per remaining ton, higher discharge temperature, more oil-cooler load, and often fewer tons if the machine hits a limit. The screen translation is extra input kW and sometimes a second machine.

Worked — same West screw, badly fouled evaporative condenser (original values).

Normal (11.1): 480 V, 328 A, PF 0.88 → 240.0 kW. Condensing 85°F, wet bulb 67°F, approach 18°F.

Fouled teaching case: 480 V, 375 A, PF 0.85, same suction, condensing 95°F+, approach in the high 20s.

Input kW = 1.732 × 480 × 375 × 0.85 / 1000. 831.36 × 375 = 311,760; × 0.85 = 264,996; ÷ 1000 = 265.0 kW.

Combined self-inflicted bill if the 25 kW of dirt is both always-on energy and the demand adder all year:

$21,250 + $3,750 = $25,000/year for one screw, before counting lost tons or a second compressor.

If tons also slip from 210 to 190, kW/ton goes from 240/210 = 1.14 to 265/190 = 1.39. You paid more money for less refrigeration. That is the definition of poor operation on a monitoring item.

Noncondensables look like a dirty condenser on the screen (high head, wide approach to wet bulb, discharge temperature up) and cost the same extra kW until you purge. Failed fans or spray on an evaporative unit do the same. Floating head in the other direction is free money: if wet bulb is 50°F and you still hold 95°F condensing with a flooded pile of discharge pressure, you are buying lift nobody ordered. Chapter 13 covers floating-head strategy; here you only need to price the kW you did not have to draw.

How to attack a CIRO cost item under the clock

1. Compute input kW (include PF).
2. Multiply by the given $/kWh (or $0.17/kWh if that is the stem's example) for $/hr.
3. Multiply hours: 24, 168, 5,000 — read which period the question wants.
4. If the stem mentions demand, compute extra kW × $/kW-month separately. Do not add demand dollars into a kWh product without converting units.
5. Excess cost of poor operation is (kW_bad − kW_good) × rate × hours, plus demand if the peak moved.

Exam traps

• Using nameplate 300 HP × 0.746 as billed kW.
• Dropping PF, then pricing the kVA-shaped number as if it were kWh.
• Quoting $0.17 as if it were dollars per hour with no kW.
• Mixing demand ($/kW-month) into a kWh multiplication.
• Pricing only the hour and ignoring 5,000-hour energy, or pricing only energy and ignoring a peak that sets the month.
• Assuming a washed condenser erases a demand peak that already occurred.