Cold-Storage Demand Charges: Efficiency First or Battery First?
A cold-storage and food-processing site has a connected load of 4.2 MW, annual electricity use of 28 GWh, and a load factor of 0.76. The tariff has an energy charge of $0.11/kWh and a monthly demand charge of $18/kW based on the facility's maximum 15-minute demand. The annual bill is about $4.0 million, of which demand charges are roughly $0.9 million. The facility manager proposes a 1 MW/2 MWh battery to shave peaks. An energy service company instead proposes a $180,000 LED retrofit saving 1.2 GWh/yr and 220 kW of coincident peak, and a $650,000 compressor VFD and controls package saving 2.4 GWh/yr and 380 kW of coincident peak. The site has 15-minute interval data for three years, but production throughput varies by 20% and summer cooling degree days vary by 18%. The CFO wants a ranking by return and a measurement and verification plan that will survive an audit.
1. Calculate simple payback and cost of saved energy for each measure, including demand savings.
Read the worked answer
LED: energy saving 1.2 GWh × $0.11 = $132,000; demand saving 220 kW × $18/kW-month × 12 = $47,520; total $179,520; simple payback $180,000/$179,520 ≈ 1.0 year. VFD/controls: 2.4 GWh × $0.11 = $264,000; demand 380 × 18 × 12 = $82,080; total $346,080; payback $650,000/$346,080 ≈ 1.9 years. Battery: 1 MW × 18 × 12 = $216,000 demand saving; arbitrage 2 MWh × 300 cycles × $0.06 ≈ $36,000; total ≈ $252,000; payback $1.1M/$252,000 ≈ 4.4 years before degradation and control losses. The efficiency measures win on both payback and cost of saved energy. The demand savings are only valid if the reductions are coincident with the monthly peak; the 15-minute data must be used to check coincidence rather than assuming full kW savings.
2. Should the site invest in the battery before the efficiency measures? Explain using the load factor, demand charge share, and interaction.
Read the worked answer
No. The demand charge is about 22% of the bill, while energy is the dominant cost. Efficiency first reduces both kWh and kW, is cheaper per saved unit, and shrinks the peak that the battery must shave. If the battery is installed first, it may be oversized: after LED and VFD the peak could fall from 4.2 MW to about 3.6 MW, so a 1 MW battery would have less value and fewer equivalent full cycles. Batteries also degrade and have limited cycles, while the efficiency measures persist for 10-20 years. The correct sequence is: implement LED and VFD/controls, verify savings with 15-minute data, update the load profile, then size any battery to the residual peak and arbitrage window. A battery may still be justified later for resilience or for a residual 200-400 kW peak, but not before the cheaper demand-side measures.
3. Design an M&V plan consistent with IPMVP Option C, including baseline adjustments for throughput and weather.
Read the worked answer
Use IPMVP Option C, whole-facility metering, because the measures interact and sub-metering every compressor is costly. Establish a 12-month baseline using 15-minute kWh and kW data, production tonnage, and cooling degree days. Fit separate regression models for daily kWh and for monthly peak kW: kWh = a + b·tonnage + c·CDD + error; peak kW = d + e·tonnage + f·CDD + error. Use routine adjustments for throughput and weather, and non-routine adjustments for schedule or equipment changes. For demand savings, identify the baseline peak day and estimate what the peak would have been under post-retrofit conditions using the model, not simply the difference in monthly maxima. Report avoided energy and demand separately, with uncertainty. Inspect installations and verify persistence at 6 and 12 months. Avoid double counting: the battery and efficiency measures must be evaluated against the same adjusted baseline, and any interactive effects should be attributed transparently.