Battery arbitrage and demand-charge stacking for an industrial consumer
An industrial consumer with 2.0 MVA contract demand is evaluating a 2 MW / 8 MWh lithium-iron-phosphate battery at its 11 kV bus. The plant load is flat at about 1.6 MW from 06:00 to 22:00 and 1.1 MW overnight. The tariff has three windows: ₹4.20/kWh from 22:00 to 06:00, ₹6.50/kWh from 06:00 to 18:00, and ₹9.80/kWh from 18:00 to 22:00. Demand charges are ₹400 per kVA per month, billed on the monthly maximum demand. The battery's AC-to-AC round-trip efficiency is 86% and the OEM permits 90% depth of discharge with 6,000 cycles to 80% state of health; calendar life is quoted as 12 years. Installed cost is ₹2.4 crore per MWh. The plant runs 330 days a year and the evening load is always well above the battery's discharge capability.
1. Compute the net daily and annual arbitrage revenue and the simple payback on the battery.
Read the worked answer
Usable energy = 8 MWh × 0.90 = 7.2 MWh discharged at the peak tariff.
Grid energy drawn off-peak = 7.2/0.86 = 8.37 MWh, costing 8.37 × 4.20 = ₹35,160. Peak-window value = 7.2 × 9.80 = ₹70,560. Net = 70,560 − 35,160 = ₹35,400 per day.
Annual (330 days) = ₹35,400 × 330 = ₹1.17 crore/yr.
Capex = 8 MWh × ₹2.4 crore/MWh = ₹19.2 crore. Simple payback = 19.2/1.17 = 16.4 years, which exceeds the 12-year calendar life. The spread captured is only about ₹4.9 per kWh discharged (5.6 minus efficiency losses), and the battery must recover roughly ₹24,000 per kWh of installed capacity over its life - it cannot do that on this spread alone.
2. The battery can also hold the monthly maximum demand at 1.70 MVA. Quantify that second revenue stream and the revised payback.
Read the worked answer
Demand reduction = 2,000 − 1,700 = 300 kVA. Annual demand-charge saving = 300 × ₹400 × 12 = ₹14.4 lakh/yr, i.e. ₹1.44 crore/yr.
Stacked revenue = 1.17 + 0.144 = ₹1.31 crore/yr, payback = 19.2/1.31 = 14.6 years.
Two engineering checks before accepting the stacking claim. First, the battery must be available at the time of the monthly maximum, which is often an unpredictable 20-minute excursion; a demand controller with a 5-minute response is required, not a fixed ToD schedule. Second, the same 7.2 MWh cannot simultaneously be sold into the evening peak and reserved for a demand spike - the controller must prioritise, and reserving 1.5 MWh of headroom cuts arbitrage revenue by roughly 20%.
On realistic numbers the stacked payback stays above the warranty life. The project is not financeable at ₹2.4 crore/MWh.
3. State what conditions would make the project viable and what you would recommend to the board.
Read the worked answer
Solve for the required annual revenue: 19.2 crore / 12 years = ₹1.60 crore/yr just to break even on capital, which is ₹43 lakh/yr more than the stacked case.
Sensitivity to what the developer controls:
- Capex: at ₹1.5 crore/MWh (₹12 crore) the stacked revenue gives a 9.2-year payback - viable.
- Tariff spread: the peak-to-off-peak gap must roughly double to about ₹9/kWh for arbitrage alone to work.
- Additional streams: frequency-regulation or ancillary-service payments, or a ToD widening by the regulator, would close the gap.
Recommendation: do not sanction the 8 MWh unit on arbitrage plus demand charges. Run a competitive tender to test whether ₹1.5 crore/MWh is available, and in parallel pursue the no-regret alternatives - stagger the 18:00 load ramp and reschedule batch processes - which capture part of the demand saving at near-zero cost. Cycle life (6,000 cycles at 330 cycles/yr = 18 years) is not the constraint; calendar life and capital cost are.